AI: Extended discussion on how AI compresses prediction horizons and may amplify reflexivity, potentially benefiting trend-following over predictive strategies.
Crude Oil: Multiple references to oil price dynamics, sanctions-driven moves, and a detailed micro-to-macro example of how oil trends scale through feedback loops.
Electricity: Emphasis on electricity as the successor to oil for economic growth, with grids maxed out, base-load shortfalls, and data center demand underscoring a structural power constraint.
Precious Metals: Noted violent volatility with gold’s largest daily drop in a decade, broader metals retracements, and a still constructive trend backdrop per models.
GICS Sectors: Energy and Utilities highlighted via oil and electricity themes, while Materials captured the precious metals discussion.
GICS Sub-Industries: Focus on Precious Metals & Minerals and Electric Utilities as key structural beneficiaries/risks within the macro narratives.
Market Outlook: Speakers see a robust environment for trend following, aided by decoupling, episodic volatility, and structural trends across metals and energy.
Companies/Tickers: No specific public tickers were pitched; the conversation centered on commodities, macro drivers, and systematic risk management.
Transcript
Imagine spending an hour with [music] the world's greatest traders. Imagine learning from their experiences, their successes, and their [music] failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world, [music] so you can take your manager due diligence or investment career to the next level. Before we begin today's conversation, [music] remember to keep two things in mind. All the discussion we will have about investment performance is about the past [music] and past performance does not guarantee or even infer anything about future performance. Also understand [music] that there's a significant risk of financial loss with all investment strategies and you need to request [music] and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran hedge fund manager Neil's Krup [music] Larson. Welcome or welcome back to this week's edition of the systematic investor series with Richard Brennan and I Neils Castro Blast where each week we take the polls of the global markets through the lens of a rules-based investor. And I also want to say a warm welcome if today's your first time you're joining us. And if someone who cares about you and your portfolio recommended that you tune in to the podcast, I want to say a big thank you as well for sharing this episode with your friends and colleagues. It really means a lot to us. Rich, it is wonderful, really wonderful to be back with you this week. How are you doing? What how are things down under? >> It's great to be here, Neils. And it's hot down under. So, um, the Australian interior is boiling and it's blowing across the the coast. So, um, I'm up in Queensland, of course, and I'm sitting in weather that's about 35° Celsius at about 7:30 at night. So, it's boiling. >> You know what? I I actually did notice that. Um, I saw a headline that Melbourne, because I know I think you went to Melbourne earlier this week. Um, that they were up to 44° um this week, which is just um out of this world really. And probably not usual, I imagine. >> No, it's it's it's certainly not usual. I think uh we're all suffering a bit over here in Aussie land. >> Yes. Well, I wish I could give you a little bit of the uh European cold weather um to share. [laughter] Anyways, it's great to uh it's great to be back with you. We've got a as usual wonderful um lineup of topics. Um and a little bit of a debate between you and I, I think might come up uh towards the end. Um and um before we get into all of that, of course, I would love to hear what's uh been on your radar the last few weeks since we last uh spoke. >> Okay, Neil. So look, over the past few weeks, I've been thinking a lot about um adaption feedback and how the human mind fits into modern markets. And this I've been doing quite a few publications on my blog posts uh dealing with uh fractals and um and the non-stationerity of markets and all of these aspects. But uh I've sort of been delving into um the likes of uh people like Ian McGillchrist and there Seth and others who they've explored explored what's called the divided brain and the nature of consciousness and they often describe how the the right hemisphere of the brain you've got to think of this. So the right hemisphere of the brain is taken from my perspective my right side of the brain. Um that that's the right hand side of the brain. um that uh perceives the world as relational, embodied and full of meaning. Whereas the left hand side of the brain dissects things into rules, categories and models. So if you can think of it, the the left side of the brain is your data center. It's where the the crunching occurs. The abstract models are created. It's using all of the the sensory inputs from um inside your body to basically create a model of what it thinks is out there in abstract terms. And then the right hand hand hemisphere is what contextualizes it. It's where you get this sense of self through feedback which says I am me and there's this thing outside me called an external environment. And what it's doing is it's mapping those models created by the left hemisphere and saying how good are those models. So the brain is this prediction engineed and this is how both hemispheres of the the brain with feedback between each other model and they adapt. They allow for adaption. This is different to AI models. AI models are when we look at it are very left hemispheric. They're all about crunching data, simulating, abstracting, creating models. But where they lack this ability to define self is they don't contextualize that within the greater environment. And this is where both hemispheres of our brain seem to be from what these these writers are telling us, these neuroscientists and philosophers etc. is that the sense of self is created through the feedback between mapping the internal model that's being generated by your brain versus what is out there and how well that model works with what is out there and that's where you get this sense of self and for a brain to work and Neil Seth and Ian McGrul argue it's not just needing a brain it's needing embodiment a body with senses measurements etc all of these things is necessary to create conscious human beings and that they they argue that AI has got a long way to go uh before they get to that level of selfidentification, self-recognition. They're marvelous at mimicking, simulating left hemisphere work alone, but very bad at that they don't have a sense of self. And the thing is when you get this sense of self with the right hemisphere and left hemisphere, it tells you how aligned you are with the external environment. And if you are not aligned with the external environment, it it creates symptoms such as pain, fear, all of these symptoms uh which your your models then recalibrate to remove that and your model adjusts. So it's this continual feedback between right and left hemisphere. And what I was thinking was that well that's how our brains have evolved in these environments and it's been very successful in these environments but the markets financial markets is not something we've evolved in. And this is where I think actually we want the left hemisphere only. We want to remove the the right hemisphere's involvement in emotion, feelings, all of these things. And we wanted our rules-based systematic processes to be purely left hemispheric. And this is where I think AI probably does work very well in the financial markets. In fact, AI, if you think about it, has evolved in these financial markets. Humans haven't evolved in these financial markets. And so a discretionary trader comes across all of these emotions which comes from a a natural environment it evolved in worked in that environment was necessary for survival. But within this context they're the things that are really constraining its ability to follow rules to not fear feel fear when you get increased volatility. all of these things which is necessary in these financial markets because as we'll get to in the later topics we'll talk about things such as feedback reflexivity all of these issues that show that the markets are not these stationary entities that these alive markets compress and expand etc. And within that context, you don't want the the right hemisphere to interfere with that context. You want rules, systematic processes. And you want data to be your guide, not your emotions to be your guide. So I just thought it was interesting and seeing well what what's the difference between AI in this context and the human brain in this context. And that's the sort of conclusions I was reaching. Now without going into too much detail right now because as you say we're going to talk about uh a lot of stuff today uh in a few minutes but when you say that AI has evolved in the financial markets how how do you see that because I saw some comments and I if from some of the big well-known hedge fund managers I don't know if it was Ken Griffin or Paul Tuda Jones it seemed like some of those guys who came out and saying well you know AI is far away from finding anything that and you know produce alpha whatever there might be some efficiencies in programming etc but it's not really something that's going to revolutionize um our production of alpha so to speak um so what did you mean by it when you said >> so what I mean is that it's evolved in the context of pattern recognition >> so um early forms of of of processes that were looking looking at pattern recognition has now sort of got to the state of AI artificial intelligence where where it's always backward looking. It's always assessing the data. It's always looking at patterns and what it's attempting to do is predicting forward. >> That that's what AI is trying to do. But as we'll see in the discussions later, prediction in these markets is like prediction in the weather. It's ephemeral. It's got a very short range. And this is because of the reflexivity in the markets. And we'll talk about how when a trader interacts with this this market, they change the nature of the market. Now, because of this reflexivity, what I think's going to happen with AI is that the prediction horizon that is currently there with things like convergence strategies is actually going to much shorten. Even though it is short, it's going to compress to be much shorter, which is going to mean that the the influence of AI is going to create much shorter prediction horizons. And I think this is wonderful for trend following because what it's going to do is it's going to reduce the impact of prediction and create much more structural alignments with things such as um um consolidated concerted directional behavior. All of these things which I think is very good for trends which as we'll discuss later on in the the episode are more uh from patterns to structure. It's more structural features of the fractal nature of these markets. It's not about prediction. Prediction is something now that I think is even going to compress further the window of prediction and that's because of this influx of AI. I think that's that's where it's going. So I think they're it's going to def defeat its own purpose in these financial markets if you know what I mean. >> Uh yes. No absolutely and as people can already here it's going to be a super educational uh conversation once again with you uh Rich and um can't wait to dig into it before we do. So on my radar I mean I couldn't help noticing the um sudden evolution not evolution explosion of volatility in the precious metals sector we've seen this week. Um the gold if people don't follow this um has had it [clears throat] had had its biggest price drop on Tuesday I think it was in over a decade. Uh it lost more than 6% in a day. Uh silver and platinum losing even more on the day and palladium has given up 16% of its price in just 2 days. So lots of volatility showing up in this sector. um which will lead into a conversation we're going to also have later on about position sizing which will be uh fun no doubt. So that's interesting. Um the other thing I just noticed was a note that um uh I think the former chief strategist at Saxopstein Yakabson uh sent out and um you know it was kind of interesting when you think about the role of oil and how a lot of people have sort of been focusing on on oil um in terms of its its role its importance um and of course uh we're recording on Thursday and we saw yesterday that the US uh introduced further sanctions against a couple of oil companies in Russia and the oil prices is up uh quite sharply uh today 5% right now uh crude oil up as we speak but he writes the following he says the world is short of electricity data centers need juice cooling systems need water grids are maxed out base load is insufficient where economist once watched oil prices to gauge cyclical swings electricity is now the core or input to growth, how to get it, at what price, and with what utility to society. These are new macro questions. Oil was the lifeblood of the post Second World War economy. Electricity is the successor. Let that sink in. A society short of infrastructure, energy and delivery mechanisms mechanisms cannot grow. Inflation will rise. Productivity will stall. I mean, kind of interesting observations. Um and I'm not so sure that um that the world is really looking at electricity the way uh stain Yakosman is describing it in relation to oil. I think still people are more biased to looking at the oil price but it hasn't been doing anything for a long time despite all the challenges we've had in um in the Middle East. Um, and I think he also describes the current um, uh, or at least this is my interpretation of it. He also describes the current debacle in private uh, private credit um, and uh, as as something that maybe it's a new acronym that we have to get used to. D A DT for markets. Don't ask, don't think. So, so we'll see. We'll see. Anyways, in terms of trend following update, the environment I would say continues to be pretty constructive. Albeit uh there is more volatility in the uh returns from managers um at least when I see the ones I can see on a daily basis uh including our own uh for that matter. Uh precious metals although they come a little bit under pressure it's still pretty constructive uh this month. Um but you know big swings um which may illustrate some short-term top correction. Who knows? Equities continue to be fairly well uh behaved from a trend following perspective. But the new kid on the block, not really new, but certainly this month seems to be also helping out is the soft sectors. Um we have uh markets like sugar supporting coffee, cotton uh in terms of being a source of returns for trend followers as far as I can tell. And then we have as I mentioned energies and currencies pretty much stuck in a range not doing much uh for the moment. What is your take on on on sort of the the state of the trend following uh environment at the moment, Rich? What are you seeing? >> I'm pretty bullish about it at the moment, Neil. So, of of course, we had the the the metals um retracement um to be expected considering their their rises they've been having all all of them, you know, gold, silver, palladium, platinum, >> you name it. Um but um I I think it's a very robust environment for trend following and um you know I'm keeping my fingers crossed. I think we're going to probably come out of this year I'm keeping my fingers crossed strong um which is a welcome relief from the the first 6 months of this you know that that was when you know of course we never know um what's going to happen in the future and I had all of these bad thoughts in the first 6 months I was thinking Christ our models aren't working anymore. No, of course you do. And uh but you just got to stick to the process. But it's it's a robust environment. Now, I think um I think we're in for some good things. A lot of the in my sort of explanation of the fundamental reasons why these things happen. um the more I'm getting into it, the more I'm I'm very comfortable uh with our methodology and process to the extent that I I think that um you know the events we saw post GFC up to 2018 I think were the anomaly. Um I think now we're in an environment where um is trend heaven basically uh for a period of time. Of course, what stops that is these um um you know, everyday uh uh change in policy sort of that um creates this whipsaw environment and that that's what we experienced in the first 6 months which made me think if that continues and we continue to get those whipsaws it'll be death by a thousand cuts. Um but I think we're out of that now and I think that the markets are reasserting themselves. Uh we're getting some solid trends and it's great to see the soft start picking up though quiet for a bit. Um but um you know the energies they're turning. They're going bullish. Um the the metals I think you know it might be a bit of a breather for them. Um but I don't think they're necessarily over yet according to our models. They're certainly um we're still sort of active in those those um things. So it's it's a pretty good environment to be a trend follower in Neils at the moment. >> Yeah. you you mentioned briefly and then it wasn't really part of our plan but but I'm always curious you know as time passes um and we get wiser hopefully uh or at least we see things differently um you mentioned this thing about post GFC um that period [clears throat] that difficult period although I will say it was only a few years that really were difficult there were also plenty of good years uh as well but there were a couple of years that were difficult um do you see those years now any different or do you have a better understanding as to why they were more challenging? I have my own views but I don't necessarily need to share those but but do you have have you become any more certain about why specifically a couple of those years and I think mostly of 20 frankly 2018 through 2020 or thereabouts uh I think both early in that in that decade uh there were some good years like 2014 was a pretty strong year um but how do you do you you see >> different I view that as um it was period of what I call compression. In other words, it wasn't a period of equilibrium at all. So, and this is what I'll get into a bit later in my topics where I talk about um >> these markets are way away from equilibrium. So what was perceived as calm was actually compression and you know that that old saying that um energy can um you know the conservation of energy it it might not have a visual appearance that it's there in a in a compressed state but it is there and it's how these markets compress and expand and I think through quantitative easing and central bank coordination we had this period of of trend suppression and compression but I I think the surplus of that has been felt from 2020 onward where the the markets had exploded out of that compressionary state. The risk never went away even though it appeared calm over that period but it was like an avalanche. was slowly building snow, snow, snow, snow. And then in 2020, it flipped out of out of this compression and boom, we had this amazing explosion in trend. And now I see this environment with the decoupling um that I'm seeing now and the the lack of central coordinated >> um actions. I see this decoupling as an environment where trends are just very favorable. um uh we we things are less correlated than they were when when things were in compression in convergence. Uh we we get um buying the dip syndrome I think was something of the past. I don't think buying the dip is going to continue going forward but that was this sort of um um this phenomenon this behavioral phenomenon associated with this compression. Um >> yeah no I agree with that. I'm actually I think I completely agree with that. I think it had a lot to do with the uh with the success of the central banks actually and I think now that they are doing their own thing and they're not really coordinating because they have different challenges, different problems. >> Uh I completely success of I think what they've done is they've kicked the the can down low with compression. >> Yeah. No, what I meant by it was they succeeded in keeping uh inflation low and stable. Yeah. and and and I think that's what they succeeded whether they kind of completely controlled it or not but that was the result and I think I think people and myself included I think we've underestimated historically the importance of not necessarily just the level of inflation because that dictates the level of interest rates etc etc but it's actually um the stability of inflation how much that really impacts uh the markets we trade and therefore the opportunities we we we um we see. So, so that's what I meant. But anyways, we'll we'll get to into all of that good stuff. Let me just mention that the trend barometer yesterday finished at 52. So, that is a strong reading and uh >> it's coming up and uh supports uh what's going on because uh the data um as of Tuesday evening um and I think by the way yesterday was a slight negative maybe mix day for for for the space. But anyways, as of Tuesday evening, the Btop 50 index is up 2.12% for the month, up now 2.58% for the year. Uh so CT index up 2.35% uh for the month, down still 40 basis points for the year. Trend index up 2.85% in October up now 53% uh sorry not 53% up 0.53% for the year and the stock short-term traders index up about 1% but still down 4%. Uh so far this year, MSEI World up 40 basis points in October as of last night, up 18.3% for the year, very strong. S&P US aggregate bond index down 44 basis points in October, but still up very strong 9.8% for the year. And the S&P 500 total return up 22 basis points uh as of last night, up [music] 15.08% so far this year. >> [music] >> All right, enough said about uh all of these things. We definitely need to get into um your agenda today, Rich. You put it together um as a true professor. And um I'm going to basically pass it over to you. I'm going to try as as a good student to keep up um and and maybe have a question or two along the way. All right, Neils, we've got four topics today, but um I I thought I'd take this today's conversation in a slightly different direction where um rather than focusing about a determined universe and things like that that we get into in past conversations, we'll stick to the markets this time. Um but I'd like to uh dig into the foundations of how markets actually build themselves. >> Mhm. So we'll start at the smallest possible scale with a single trade and trace how structure then grows upward through feedback and interaction from impact to fractal geometry which I'd like to get into. So um we'll then move into how that scaling creates trends and how patterns differ from structure and finally what all of this means for risk survival and how we position size using closed equity and then uh that'll give you the opportunity to jump on board. I know you want to say something about that. >> So we'll begin at the beginning with this concept called impact. So every price move starts with a decision. So a trader can buy, they can sell or they can do nothing. That single action no matter how small exerts a force on the market and that force is impact. It's the fundamental impulse of price movement. So most people think price moves because information is revealed but markets move because impact is applied. So that's the difference in this interpretation. So each trade order changes the balance of supply and demand and therefore reshapes the conditions for every trade that follows. Can you see that slight reflexivity there? >> Yeah. So at the micro level, individual impacts interact through the order book and liquidity network. Some cancel, some impacts cancel and some impacts reinforce. When reinforcement occurs, feedback begins. Now prices therefore no longer just reflect information. They therefore create it through this feedback. So traders they observe price they therefore once they observe the price they then update their models and then act again. So they'll make initial trade they'll see what the impact on the market is then they'll update their models then they'll act again and this loop continues and that's reflexivity in action. It's where a self-modifying system learns from its own output. So here we've got an input, a trader making a decision. We've then got an output response, what the market responds to, and then an adjustment with the next input that comes in from that trader. There's just reflexivity involved. So that's reflexivity and action. Now, markets don't require rational or informed participants. They simply require actors whose behavior feeds back. Each irrational trade changes the landscape and become part of the evolving structure. And that's why markets are seen to be self-organizing. They evolve through interaction, not through equilibrium. But here's where it gets really interesting. Market impact is not linear. A trade 10 times larger doesn't move the market 10 times as much. It's not a linear relationship. M so Jean Philipe Bashard's research shows impact grows with a square root of trade size and what that means is if you double your order you only increase expected impact by around 40%. Not the same amount it's a reduced amount and that's because liquidity is adaptive. It's not a static pool. It's seen to be like an elastic surface that moves and reforms with order flow. So every trade that we do consumes liquidity and signals potential direction and other traders react to that quotes adjust and the whole system bends. This is where we turn away from linear to nonlinear. This is what gives u markets their nonlinear fractal character. small trades disappear into the noise. That's because it's a sublinear rel relationship, not a superlinear relationship, a sublinear um relationship. Small trades disappear into the noise. But when many small trades align, feedback compounds, volatility clusters, and directional bias emerges. And that's how trends, crashes, and factiles are born. not from randomness which the efficient market hypothesis will tell you but from interaction. So if these events were independent like a gorian model would would say they are independent random events returns would form a bell curve but in markets events are all conditional. Each one changes the probability of the next because of this reflexive nature and this is what destroys erodicity in these markets and replaces it with path dependence. Um so I know you don't like that term erodicity and >> well I want you to just remind people what it means. So um an an erotic system is one where an ensemble of um results are the same as the um an individual across time. In other words um it's saying the ensemble of statistical results at a particular point in time is equivalent to a single um outcome statistically over the course of time. The both are equivalent. But in a non-gotic system, there's an asymmetry there. And we find that wealth paths in financial markets because of compounding this geometry, they're not um that they you can't apply erotic statistics. In other words, the the statistics break down. A good example, the expectancy equation is what we call an erotic statistic. It says that um if we remember the um expectancy equation, it says your percentage win multiplied by the dollar win less the percentage loss multiplied by the dollar loss gives you an expectancy equation. Now this therefore says to the trader ah if I have positive expectancy I will be profitable. However, when you look at that equation, what you don't see is the sequence of events. And this is critical. And this is what Ole Peters found when he found that markets are non-gotic. And this is because there's a conditional um reaction that occurs with a finite sum of money which has a lower bound of zero. Um so it's an asymmetric system. It's not an open-ended system where egotic systems are. The the reason why we get the equivalence of ensemble averages and time averages in an in an agotic system is because they are open-ended on either side. No lower bound, no upper bound. And there's basically stochastic movement available across all dimensions. But in a non-rootic wealth path, there's an asymmetry in the system. There's a lower finite bound and there is an open-ended bound. And this is what we find in fractal systems. And we also find that is why they are non-aggotic. Which means that in those systems no longer is the um the statistical account of an individual over the course of time the same as the statistical ensemble at a point in time which means that the path um of compounding uh is different to what the statistics are going to represent. Now this is the problem with the Gaussian theory. The Gaussian theory, this independent um independent um random events um produced a bell curve and it assumed that there was no asymmetry in that. It was open-ended upper bound lower bound and we got all of the associated toolkit with that which were the statistics that made that model work which was things you see in modern portfolio theory. Marowitz's efficient frontier, sharp ratio, standard deviation, um, expectancy equation, risk of ruin equation. All of these assume an erotic system. But as reality tells us, it's actually a path dependent system and all of those toolkits break down. And that's because we get in these asymmetric systems that are fractal in nature, we get power laws, we get non-stationary environments, we get a a decrease in the predictive power as we extend into the near future. Um we get um no equilibrium away from equilibrium. All of these things are saying that Gaussian model is is incorrect. And then we got um this this instance where I think in the 1960s Benoir Mandelro came out and said look we can demonstrate that these financial markets are non-Gorian because all of these rules are not being obeyed by the financial markets and he came to the conclusion that it was a fractal system. That's the conclusion also. >> Yeah, >> I've sort of come to realize over the course of time. >> Yeah. So no, >> if these events were independent, returns would form a bell curve. But markets events are conditional. Each one changes the probability of the next with this reflexivity. This is what destroys eroticity and replaces it with path dependence because the path matters. What happens previously, there's memory in the market. Each of these events because of this reflexive nature of the market are not independent. What happens before shapes what happens next through this reflexivity. trader impact. This therefore means a sequence that over time there's a memory in the market and it's captured by things like the hertz exponent. uh we've got these addition new tools we can use that says if there is market memory um using Bord's analysis we've got a different set of tools in our toolkit which aren't gausian but Hurst exponent tail properties tail decay all of these things in a power law driven system these are better tools to use that are more representative of our market says that the old tools could never account for all of the calamities that occurred over the course of time by relying on a model that had these assumptions that did not reflect reality. And hence anyone who relied on those tools, portfolio managers, industry, all of those things would always get blindsided regularly much more than what a Gaussian um distribution would imply. And hence, you know, um a Gaussian distribution says that a five standard sigma event should occur every 170,000 years. But when we look at I looked at the S&P 500 or the ES uh futures market for instance and in the last 30 years there's been 32 five sigma events and this isn't just associated with a single market. Every single liquid market I examine has these tail properties fat tails. This is saying they're non Gausian. They're fractal. um the and and the look the verdict's out whether they are exactly fractal but certainly a fractal model has a better account and can understand why these situations occur. So I'm not saying it's a definitive answer. I'm just saying at the moment fractal models and the fractal market hypothesis probably is the closest we get to reality because it explains so much more. So, let me ask you one thing while you uh get a sip of water for your for your throat. Um, >> and I don't know if this is um correctly understood, but but maybe there's a question in there as well. And by the way, this all this fractual stuff um I think you and I touched up uh on this before we pressed record. It does remind me of my conversation with Bill Rice. I think he actually back then >> um in the 70s when he started um built his pattern recognition model based on fractal um you know fractals. So anyways, but if volatility is an output of reflexivity, not just exogenous news, so to speak, are we underestimating the role of our own participation in creating the risk we seek to avoid? The risk we seek to avoid. Okay. The bottom line is we can't predict these things. So we might um we might have tools that say um that volatility equals risk. And this is exactly the problem, the dilemma I find in that I don't associate volatility with risk. I view markets as always having this potential risk event around the corner. >> What I view is um low volatility regimes I actually view as high latent risk events, warehouse risk. In other words, the risk hasn't gone away. It's there, but it's now compressed. It's like a sponge that you compress and it's going to explode at some point in time, waiting for a tipping point. Um, when we see this massive expansion, this transition. Um, so I'm seeing the market breathe, contract, and expand, contract, and expand. I don't think we can measure volatility using the standard statistical tools we've we've done. I can understand what volatility. mean we can't measure risk >> no >> or volatility because we can measure volatility right no >> I we can measure volatility but then there's um directionless volatility there's directional volatility um one might come with correlation might one might come without correlation um it's it's a signature of a fractal system that's alive compressing and expanding but because of the the undefined nature of predictability in fractal systems And in other words, what I'm saying there is that whilst there, you know, when you look at the weather, which is a fractal system, there is a limited prediction horizon within weather. You might get a fairly accurate assessment within >> 10 days or 14 days. That doesn't preclude the the ability for these large events to interrupt that prediction, but it does in certain regimes give you this prediction horizon. But over the long term, they are they are unpredictable. that they are deterministic systems that are undefined and this is where they follow these strange attractors and in the financial markets which I view very much similar to weather systems where instead of molecules of water um in in the the clouds etc. I'm viewing agents traders all as collectives that there's no central governor there. all working according to their own mandate but they are interacting with each other and these impacts and feedback loops that occur between each other make it calculate impossible to calculate or impossible to predict um so I I refer to them as deterministic um unpredictable systems which a lot of people say well that's chaos and I say well no um we get periods of very rational order um you know it's this reflexive nature of the market they adapt they respond to what the participants are doing in it. But so I don't have I prefer what I call engineered outcomes rather than statistical um statistical tools to define how to protect myself in this market. So when we talk about volatility, I'm not using volatility measures to protect me or statistical measures because I think they're they're unsound. The the assumptions of these statistical measures and the toolkits of statistics have come from a gauian world. I'm saying you got to think more like an engineer. So in fractals, unfortunately the thing with fractals is these Gaussian models are just too simple to describe this very complex thing we call the markets. You've got to be an engineer. A bit like an engineer designing a bridge in a weather system. You you don't blame the weather for blowing down the bridge. You blame the bridge being too brittle. So, it's the engineer's fault if they break down because the weather, as we know, can be very unpredictable. You can get typhoons. If the bridge falls down, that's that's because it was optimized, overoptimized, brittle, could not handle the environment. The same way as I think our systems need to be engineered to not be brittle, not be overoptimized, not be not be overfit. They've got to be robust, resilient, able to stand up against anything that thrown against them. That's that's how I view robustness. >> Yeah. No, that's that's a great way of looking at it. All right. Um, was that topic one? >> So, let's go to topic two. So, we talked about impact, >> right? >> Uh, and now what I want to talk about is how fractals create structure. So, let's move up the scale and see how these micro impulses of of impact assemble into the structures we call trends. So when you zoom right in on a market chart and you zoom right down into the detail, it does look chaotic, jagged, noisy, directionless. But as you zoom out, hourly, daily, weekly, the chaos starts to organize visually. What looks like randomness when you're zoomed in actually becomes rhythm as you zoom out. And that transformation is a signature of fractality. This is different to a gauian world where as you drill in the structure dissipates. It disappears until you're left with linear results. That's just this linear independent gausian world. But in fractal systems, you can never get rid of the structure. It's always there's structure in there, but it looks chaotic. You're going down into zooms. The the structure never dis disappears. It's always there. But see at this scale the different um views of scale perceptions of this at different levels of of perception what looks random at one might have structure in the in the other patterns in the other. It's a bit like looking at a TV screen uh when you're looking at the pixels and you get move away and you're looking at a different resolution you start seeing the images um that are are linked together from the pixels. This is how we've got to understand fractality. So, u fractality isn't about prediction. It's about relationships across scales. Each time frame represents a layer of feedback. And we'll find that people are responding depending on the time frame they're interacting on the market with the patterns that are observable at their scale. And so, you get this this scale difference occurring. high frequency traders at the lowest end of the scale see markets a different way. They're seeing this sort almost chaotic frenzy of movement. As you step out, you start seeing more and more trend followers start participating in the market. Why is that? And it's because as you scale out, you you start seeing these these impacts we're talking about with the fractal nature of markets. Some are cancelling, some are reinforcing. the reinforcing elements of these structures is what's creating the trend as we're zooming out rather than a cancellation. So mean reversion is is a is a a cancellation >> um environment. It's where we get opposing forces. So we get a force of of um reverting back to an equilibrium and then we get a force of going away from an equilibrium. This alternating zigzag up down up down up down. They're almost linear in in nature. They cancel each other out as we go out in resolution. But directional impulses that aren't canceled out start aggregating together and compounding. They're compounding structure into the market. So this is why convergence systems are linear results. And this is why when you look at the P&L of a convergent system, you get a linear profit, a linear profit, a linear profit, a linear profit, a linear profit until they come headon with a negative skew event where you get a a nonlinear loss. But when we look at trend following, it's the reverse. We get a a small linear loss, a small linear loss, a small linear loss, a small linear loss. When we come up against an outlier, a structural a structured directional trend we get a nonlinear gain. This is the fundamental difference between it and this is because of this cancellation and reinforcement that occurs in the market at the fractal level level. So when short-term reinforcement persists and aligns through time through scale you get coherence and the coherence is what we call trend. It's also called serial correlation or bias. So positive feedback drives reinforcement. Buying therefore attracts more buying. That's how it works. This reflexivity. When people see a trend, people start jumping onto the ride. Jumping onto the ride. Buying begets more buying. Negative feedback, however, drives regulation. Selling that restores balance. One is moving back to equilibrium. one is moving away from equilibrium trend directional positive reinforcement buying begets more buying or selling begets more selling moves away from um this equilibrium zone and negative feedback is the reverse a restoration trying to get back to uh this is what we call regulation um selling that restores balance markets oscillate between these two drivers and when positive feedback dominates energy compounds and structure emerges fractally. This dance between expansion and contraction creates a fractal rhythm of markets. You know, you get quiet periods punctuated by bursts of volatility, compression followed by release and feedback is the pulse of adaption. So what do I mean? People are reacting. They are they are impacting as they see a trend and they're buying. They are accelerating that trend. When people see a trend and they want to revert against that trend, they're they're applying negative feedback and trying to restore balance. But you see how this works. So when feedback cascades through scales, small interactions become large outcomes. So [clears throat] cancellation gets rid of structure. Positive reinforcement creates structure and small inter when it cascades across the scales, small interactions create large outcomes. A local burst of buying becomes a cluster. Clusters form rhythm. Rhythm becomes flow. That's how feedback builds structure through time. It also explains Neils why you and I are medium to long-term trend followers because the frenetic activity that occurs in the shorter time scales are more mean reverting in nature. there's more cancellation going on relative. But as we get out to the the higher sales, we'll find that the feedback tends to operate on the structure, not on the mean reversion. When I say feedback, most participant interaction, institutions, etc. out to the medium to the long end are positively reinforcing trends. They're not going against the trends. Mean reversion is something that occurs in the the finer time time frames. So it's the same in nature. ripples becoming currents. Currents forming rivers. This is positive feedback in natural systems. In markets, local feedback becomes directional flow. So take crude oil for example. At the micro level, traders buy ahead of a report for instance. Algorithms detect it and they join. But on the hourly chart, those bursts at the micro level appear as clusters. Zoom out to the daily and a narrative forms. Oil is recovering. Come the headlines. Oil is recovering. Starts turning into narrative. At the monthly level, that narrative becomes systemic. Producers, investors, and policy makers all reinforce that movement. Feedback has scaled from the tick to the macro. The market has literally organized itself around its own success. And that's what it means from frames to becoming trends. So eventually though every feedback loop reaches its limit. The same alignment that built the trend becomes its constraint. Positive feedback will flip to negative feedback at some point in time. Rising prices ultimately exhaust buyers. Valuation stretch. Risk becomes concentrated. one event or simply fatigue breaks as a symmetry. Then selling accelerates, stops trigger, the system unwinds. Maybe even margin calls come into play, the system unwinds. Reversals aren't random. They are feedback inverting. And because energy stored through the long positive reinforcement, you'll find that during these reversals, it's released quickly. Reversals are fast and violent. They're not like the the typical buildup that occurs on the long side. These reverses are fast and violent. Exactly what we saw with gold and the metals earlier this week. Now, so this feedback flip isn't failure, it's renewal, system renewal. It resets the system so the process can begin again. And it's this continuous process, this continual cycle, reinforcement, alignment, reversal. It's the living rhythm of a fractal market. So that's topic two. >> Yes. I'm going to save any questions because uh again we have a quite a bit to get through. Um and so I'm going to give you as much time as possible to to move on in your own narrative so to speak. >> So now I want to talk about the difference between what I'm referring to as patterns structure and the fractal nature of outliers. So, >> okay, >> now that we've seen how structure builds through feedback, we'll explore the difference between pattern and structure and why the outliers matter most. So, of course, I'll say that cuz I'm an outlier hunter as you know. >> You >> okay? Markets are full of patterns. We all know that. We see flags, we see triangles, we see breakouts, we see moving average crosses. These are all surface forms, transient, visible, easy to name. When I'm referring to structure, I'm referring to what lies beneath. It's the geometry that govern how markets behave, not how they look. Now, I remember in a previous podcast with you, I've talked about an outlier hunter um doesn't have a prescriptive definition of what it refers to as trend. It's looking for the structure, the things that create the bias. And that can create come in many different visual forms. But those people that are looking treating trends as patterns are probably too prescriptive because these trends, we're looking for the structure. What drives these trends? Because as you know, Neils, trends can actually be a a random result from from um no bias in the price series. We can get a random trend very easily. We can get a a a trend that actually is a a segment of a mean reverting cycle. Um or we can get these structural trends with serial correlation in them that gives persistence into the future. That's what I'm calling outliers. These are structural things. They're not patterns. It's created by causitive drivers um that actually create this vast array of different directional patterns which I'm calling uh outliers. So it's the geometry that governs how a market behaves, not how they look. It's built from the relationships, the incentives, the feedback loops that determine how energy flows through that system. So patterns describe what price does. Structure explains why it does it. So patterns are what we observe. Structure is what connects those observations through time. Notice that structure is what connects those observations through time. It's talking about a memory. What happens before happens later. This is a serial correlation. This it's not independent that there is a a time comp over time. This feedback occurs over time in these fractal systems. So trends are patterns out outliers I regard as structure revealed. An outlier is not an accident. It's a phase transition. It's a moment when feedback alignment pushes a system far from equilibrium and forces it to reorganize. It's not just a mere pattern. It's something structural that's really changing the system. Compression. Now compression, you know, I talk about compression and expansion. Compression hides structure. Expansion reveals structure. Outliers are those expansions. The moment when the market changes its own geometry. So in fractal systems, small fluctuations typically cancel, but large ones dominate. The tails contain the power. That's where this adaption happens and where returns are made. That's why I call myself an outlier hunter. I'm not chasing a price pattern. I'm aligning with structural change. That's how I view it. So I design system ensembles which are families of reactive systems that listen for different ways structure might express itself. I don't trade single trend following systems. I put in these ensembles. They are what I call I'm not looking for correlations here. I'm looking for behavioral orthogonality. That's a big word. So in other words, >> orthogonal organic. >> Yeah, I'm going to let you pronounce that word for sure, but but so you're essentially saying diversifying by behavior. >> Yes, a systems behavior. That's why I will consider a breakout. I will consider a mean reverting into a trend. I'll consider these these are what I call behaviorally orthogonal. They will never all act in concert together. They are structurally looking at at the different manifestations of how trends can form. And it's not saying there is one prescriptive form. It's saying I've got to diversify across as many because this structural outlier can come in a array of different varieties. So I'm not looking for correlations here because these are very fickle things. I'm looking for structural behavioral differences between things. That's that's my choice of >> you know the risk you know the risk of this Rich is that you are maybe less classic than than what [laughter] what other what other trend followers might perceive as being classic but I like the idea. >> Yeah. Okay. I look I'm I prefer to call myself an outlier hunter. >> Yes. Yes. No I know. I know. >> So these different systems okay one for instance might detect volatility expansion. Another might detect smooth persistence. That's that's looking for a breakout from a congestion Darvis box breakout. That's >> another might focus on breakouts. Each hears a different voice of the same feedback process, but it's relating to structure. Together, >> they form a coherent adaptive framework, not predictive, just ready to strike and activate when the signals erupt. That readiness is everything to me. And because we can't know which move will reveal structure, but we can ensure through this system ensemble, we're alive to capture it when it does. So what I'm now getting to, there's a shift in the topic here. What I'm getting to is the need to survive until the outliers arrive. Okay? And this is where uh so to summarize topic three I'll say patterns describe form structure defines cause outliers are structure made visible through feedback. That's why in a fractal world we don't diversify by correlation. We diversify by behavior by how our systems respond to changing structure. Okay. So now I'll get into topic four. The last topic and this is where we come to our a possible debate coming up. Neils. So this is I call the fractal reckoning part dependence survival and closed equity. Okay. So I'm going to connect all of this um to the most practical question of all. How do we size positions, manage risk, and stay alive in a world that's part dependent and fractal, non-stationary, um unpredictable, uncertain? How do we survive in that world? In an ergotic world, a Gaussian based world, um, like a casino, a casino is a good example of an ogotic system, the average outcome across many plays equals the average outcome through time. You know, I talked about the average outcome across an ensemble is the same as the average outcome over time. That's in a casino. We've got that situation. However, markets are non-gotic. We live only one path. We can't live in these parallel universes that statistics say we can. We have one path. And once ruin occurs, the game ends for us. Once ruin occurs, game over. That's why survival, not expectancy, is the true measure of success. This is why I harp on what's the best method of determining the best managers track record. Survival. Survival. This this comes down to this conclusion. You can't measure this statistically because the statistics is the wrong toolkit because the statistics come from the Gaussian model. So Kelly criterion bad statistic expectancy bad statistic I don't want to say this Neils but I'll say variance at risk is a bad statistic however I know your opinion on that but all of these are coming from this Gaussian model which isn't the reality as we've observed. So in a non-aggotic world, the path is everything and the order of wins and losses determines whether you compound or whether you collapse. So think of two traders with identical expectancy. One experiences losses early and they run out of capital and they never reach the recovery phase. The other survives long enough for the outlier to arrive. Same expectancy but completely different destinies. They're part dependent. Wealth doesn't grow additively. It grows multiplicatively. Every trade we make changes the base from which the next trade grows. That's this reflexivity and action. Okay? It's not independent reflexivity. Every trade we make changes how the next trade grows. Losses shrink the base and that drag compounds. That's why minus 50% followed by plus 50% doesn't equal zero for your wealth. It equals a permanent hole in your wealth. So if we have $100 and we lose 50% we're down to $50. But if we gain 50% we're only up to $75 or whatever. We've got a permanent hole in place. This is this sequence risk which isn't addressed by the Gaussian world in oootic systems. Part dependencies everything. So this is where closed balance equity I believe becomes crucial. We use closed balance not floating balance or or equity to size new positions because to us it represents realized capital not illusory equity. What I mean there is closed balance is the only equity you actually possess after the path of returns has spoken. So what I mean is it already incorporates the non-aggotic journey every draw down every recovery that is what it represents. By sizing from close balance we ensure that our risk per trade remains constant relative to survivable capital. It automatically adjusts exposure downward during draw downs when the system is under stress and scales up only when the process has rebuilt through trades that have been confirmed. This is a structural response as far as I'm concerned to a non-aggotic environment. Now I [clears throat] can for instance demonstrate to you how a system with uh 5% positive expectancy but have um large size positions um and at even under that arrangement we get this massive compounding drag and we never get a wealth return out of it. It goes to zero. Reduce the size of those bets significantly reduce the size of the bets. I can show you a system where we've got a 30% win rate, a 70% loss rate, a three times um win win amount to a a one times loss amount, but with very small positions, the expectancy is only 2%. But compounded geometrically, you get a very very good return. >> Now, this is this path dependence sequence. So, this is a structural response, I believe, to a non-aggotic environment. It keeps us in the game long enough for the next outlier to appear. The closed balance principle is a way I believe of embedding sort of humility into the system. It acknowledges that we don't know what the next path will bring. So we size from what is real what has survived. We're not trying to forecast variance or optimize volatility. We're simply aligning our capital to the geometry of survival. In a world that compounds through feedback, the key is not maximizing return. It's avoiding ruin. And that comes from this fractal mindset that I have. Prediction fails because it assumes a world stands still. Process succeeds because it accepts that it will never stand still. Um, so that's how I view this issue with closed balance. So if we can imagine when uh because I've got positive skew in my trend following models my open equity or my realized equity always sits above except at the beginning when I start it always sits above my closed balance equity which means I am much more conservative in applying position sighting because I'm using a lower limit to compound. I'm not using um the higher limit that equity provides which is unre a large component of that is unrealized at this point in time and I don't know the way that's going to go. If I for instance assume that um um it can support this higher position sizing. I'm starting to leverage up and it's starting to get towards that model I was talking about 5% expectancy but very large position sizes is starting to get me an unfavorable compounded path. That's how I see it. But over to you over to me. Okay. So a lot of the stuff that you said I completely agree with. Um I think my um comments um maybe more in terms of how uh the difference between how most managers um size positions and uh what you and and a handful others refer to as the classic trend followers. I think it's more than the the way it's being described that um concerns me a little bit. Um the way I hear it when you and some of our friends talk about uh and maybe not so much you but talk about this difference is that one is right and one is wrong and that closed using closed end equity is right. And I'm I'm I'm I'm thinking well there must be a reason why the majority at least of managers are not using that methodology. Uh so maybe you can't really say one is right and one is wrong. Um you can say they're different. Now I hear the criticism of the camp that I find myself in that dynamic position sizing is only being done to reduce volatility. It's also sometimes referred to as volatility, you know, managing volatility or volatility targeting, which it's not necessarily the same. um that that you know that you're trying to manage volatility and you're trying to make this the the ride a bit smoother um as if there's something wrong with that. But then when I hear uh this closed end equity position sizing being described, I'm thinking well that is also reducing volatility because you're simply sizing your trades on a smaller equity. So that's reducing the volatility as well of that strategy and and and and thereby creating a smoother ride. So what is the difference? What is why is why is one wrong and the other one is right. So that that's the kind of the first thing that where I just think that the the way it's being described is is you know is not necessarily how I would describe it. And the other thing is you said something very interesting. You said that the open equity always sits above or the total equity always sits above the closed end equity only exception. >> Yes. Well, that is very important. >> Yeah. Yeah. >> Because in my view that kind of means that that method works only and I know I'm going to get some some some feedback on this. Um only works if you're trading your own capital because you were there at the inception. But if you're trading an a a a fund with external clients that comes in every month, then they are not necessarily at inception of that open equity um calculation. And that to me opens up and a risk that is kind of undefined because they're you're taking on a riskmanagement methodology that is not specifically to uh you know geared towards them but is geared towards quote unquote the original investor. So just putting it out there just putting it out there. the the other thing that I've heard you you describe the open equity as some kind of illusorary equity and and I'm thinking for me at least because we trade the most liquid markets in the world yeah >> we can all even the biggest managers in the world we can all liquidate our portfolios in 24 hours so it doesn't seem that illusory uh to me it's pretty deal. Um, so again that takes me towards this idea of actually you should consider your full equity as the basis of your your of your your trade sizing. And the final point I just want to um throw out there is just that having these rules are fine, but I hear from um from from this from your camp, not you specifically, but from your camp that on top of this you throw in something called the cutback rule, which is a completely discretionary rule um about when you want to cut down your uh position size. um you know when you when you feel the pain more than you're willing to feel the pain and you keep the positions cut until you hit a new all-time high. Well, at the same time, I hear that people who do dynamic position sizing, they're somehow handicapping their own lifting power in coming out of a draw down. And I'm thinking, well, hang on. If you're cutting your position size at a certain level of a draw down and you don't touch that until you get to a new high, that is cutting your lifting power as well. And if you're doing dynamic position sizing, even if you're in a draw down, and I do admit that that is not necessarily something that reduces your risk, but if certain things occur, you could actually see dynamic positions being increased whilst you're in a draw down. And that could be good uh in in terms of of lifting your recovery. It could also be bad if you're not doing it uh in a proper way. So anyways, my my point is that when I look at these track records, I don't see, you know, the long-term track records, people who've been around for 20, 30, 40 years plus, right? It's not that I see a massive difference in the returns. I hear a difference in the narrative and and and one being right, one being wrong. All I'm just all I just want to open up to and I'm not I I don't I I don't want to say one is right and one is wrong. I just want to open the debate about saying yeah there are different ways you can skin a cat and um at the end of the day we need to do what we believe in and what we think is right. So there's nothing wrong in what you describe. I don't think there's anything wrong in in in what other people do. Uh it's just a matter of preference. But I am a little bit quote unquote allergic to this. Oh, you're doing this wrong and I'm the only one doing it right. I'm a little bit allergic to that >> argument. And I'm not suggesting necessarily that >> we need to dust up. We need a dust up with boxing gloves. >> And what we need to do is get Jerry on with me and you get to get on with your compatriots and we have a a brutal debate about it. It would be fun. It would be fun. I think >> it would be fun. But more importantly, I think this shows and let's spin it positively. I think what it shows is that as long as you follow the golden rules. Uh I think there is a little bit of leeway in terms of how you become a trend follower and and I will admit as well that the way you describe it where you are focused on a trade by trade basis is of of course a lot easier or at least I think of it. I'm not a quad, so it's easy for me to say, but I think of it as an easier way to build systems, to understand systems, maybe um maybe with less moving parts, but also a few discretionary parts, which I don't personally think um helps, but but be that as it may, >> yeah, >> um I think the dynamic uh approach is more complex. It's not that I'm a fan of more comp complexity per se. Um, but I also think that after 25, 30 years where that's probably been around by now, it's proven itself. >> We've all survived, Neils. >> Exactly. That's what I mean. We're all here. Um, so again, I don't want this to turn out to be, oh, I don't like what this guy is saying or what what what this girl is saying. All I'm just saying is I I think we should I think we should just think about the narrative we put out there. Uh, of course, some of of the narrative will give better headlines, uh, for sure and it'll be, you know, fun and pro prerogative. So, I understand there's a little bit of that as well. >> Well, see, Neil, this is why I'm classifying myself as an outlander. I'm distinguishing myself from a trend followers. And the way I'm doing that is because I do have a particular set way in my mind of how to deal with these markets. So, >> I prefer to therefore say if I'm an outlier hunter, I've got specific objectives >> to achieve that outcome. I don't necessarily think we all need the same objectives trend, you know, there are lots of different forms of trend follower, but I'm trying to set myself apart to say, all right, I've got a distinctive philosophy. It is different to a lot of others. These are the reasons I like it. Um, and that's up to them. Gates of gray I would call because remember a few months ago you did a very good um description of >> four times four and you said I added five um descriptions of trend follower and I actually think that's a much more important topic than a lot of people realize um because it can be difficult to see the difference from the outside but even more so >> investors really need to think about why would you want one over the other because it comes down to what is the objective of making an investment with a trend follower in the first place. Yeah. >> Um I think that's argument. >> They view us as a nebulous not a lot but uh we're experts in our field and we we've obviously got our idiosyncrasies. >> Sure. And I think that that actually that conversation um is winning a little bit of traction uh among investors that it's we're not just one and the same. um and um and one should think about uh the type of trend follower they you really want and need in in your portfolio. Now the funny thing is of course when you say I consider myself as an outlier hunter in a sense even though we use pos dynamic position sizing I do the same because we both want the long longer trend than anyone could even imagine and we're all long gold or we're all long stocks and so this is the fun part. It's just the path and the journey. And do we keep the same, you know, exposure at all times during this? And of course, we often think about, well, isn't it better just to have the same position size if you're into a big gold trade or into the cocoa trade a few years ago? Of course, sure. But we do forget that there's all the other positions that can be troublesome. So if you have the same position in those, they can detract from the benefits you get from being in the same position in Coco and so on and so forth. So there's all these pros and cons and I just want to invite um you know a little bit of um um not so much black and white if you know what I mean but um it's doesn't take anything away from all your observations of course and the fact that it has worked very well for people using this approach for many years at the same time using dynamic position sizing has also worked very well for those people doing that. So, which is the beautiful part of of this um conversation. Anything you want to push back on, Rich? And feel free to be brutally honest. >> You you get the last word. It's your podcast. That's the way it should be. That's great. >> It's a joint it's a joint effort as you as you well know, Rich. Anyways, this was wonderful. I'm sure people will have enjoyed it as they always do when you are on. And um to those listening, I would only encourage you to go and leave a a raving uh review and rating of Rich's Conversation on your favorite uh podcast platform. It really does help more people to uh discover the show and and uh help us spread the words of the wonderful weird and wonderful world of trend following. Anyways, um we're going to wrap up. So, next week a few changes. Next week, Moritz will be sitting in for me while I visit my daughter in Montreal and he'll be joined by Nick Bolters. So, that's going to be a fun uh conversation. And the following week, Alan will sit in for me and he'll be joined by Yoo. That's also going to be a super fun conversation while I scoot over to our headquarters in Florida uh and visit some of my colleagues uh or all of them uh for that matter. But [snorts] if you do have any questions to any of these uh esteemed trend followers, by all means send them to me as usual, info@ top traders unpluglo.com and I will forward them to um Morates and Allen from Rich and me. Thanks ever so much for listening. We look forward to being back with you next week. And until next time, take care of yourself and take care of each other. >> [music] >> Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the best way to stay updated is to go on over to iTunes and subscribe [music] to the show so that you'll be sure to get all the new episodes as they're released. We have some amazing guests lined up for you. [music] And to ensure our show continues to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the [music] best way to show us you love the podcast. We'll see you next time on Top Traders Unplugged. >> [music]
Markets Aren’t Random: The Drivers Behind Trend Formation | Systematic Investor | Ep.371
Summary
Transcript
Imagine spending an hour with [music] the world's greatest traders. Imagine learning from their experiences, their successes, and their [music] failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world, [music] so you can take your manager due diligence or investment career to the next level. Before we begin today's conversation, [music] remember to keep two things in mind. All the discussion we will have about investment performance is about the past [music] and past performance does not guarantee or even infer anything about future performance. Also understand [music] that there's a significant risk of financial loss with all investment strategies and you need to request [music] and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran hedge fund manager Neil's Krup [music] Larson. Welcome or welcome back to this week's edition of the systematic investor series with Richard Brennan and I Neils Castro Blast where each week we take the polls of the global markets through the lens of a rules-based investor. And I also want to say a warm welcome if today's your first time you're joining us. And if someone who cares about you and your portfolio recommended that you tune in to the podcast, I want to say a big thank you as well for sharing this episode with your friends and colleagues. It really means a lot to us. Rich, it is wonderful, really wonderful to be back with you this week. How are you doing? What how are things down under? >> It's great to be here, Neils. And it's hot down under. So, um, the Australian interior is boiling and it's blowing across the the coast. So, um, I'm up in Queensland, of course, and I'm sitting in weather that's about 35° Celsius at about 7:30 at night. So, it's boiling. >> You know what? I I actually did notice that. Um, I saw a headline that Melbourne, because I know I think you went to Melbourne earlier this week. Um, that they were up to 44° um this week, which is just um out of this world really. And probably not usual, I imagine. >> No, it's it's it's certainly not usual. I think uh we're all suffering a bit over here in Aussie land. >> Yes. Well, I wish I could give you a little bit of the uh European cold weather um to share. [laughter] Anyways, it's great to uh it's great to be back with you. We've got a as usual wonderful um lineup of topics. Um and a little bit of a debate between you and I, I think might come up uh towards the end. Um and um before we get into all of that, of course, I would love to hear what's uh been on your radar the last few weeks since we last uh spoke. >> Okay, Neil. So look, over the past few weeks, I've been thinking a lot about um adaption feedback and how the human mind fits into modern markets. And this I've been doing quite a few publications on my blog posts uh dealing with uh fractals and um and the non-stationerity of markets and all of these aspects. But uh I've sort of been delving into um the likes of uh people like Ian McGillchrist and there Seth and others who they've explored explored what's called the divided brain and the nature of consciousness and they often describe how the the right hemisphere of the brain you've got to think of this. So the right hemisphere of the brain is taken from my perspective my right side of the brain. Um that that's the right hand side of the brain. um that uh perceives the world as relational, embodied and full of meaning. Whereas the left hand side of the brain dissects things into rules, categories and models. So if you can think of it, the the left side of the brain is your data center. It's where the the crunching occurs. The abstract models are created. It's using all of the the sensory inputs from um inside your body to basically create a model of what it thinks is out there in abstract terms. And then the right hand hand hemisphere is what contextualizes it. It's where you get this sense of self through feedback which says I am me and there's this thing outside me called an external environment. And what it's doing is it's mapping those models created by the left hemisphere and saying how good are those models. So the brain is this prediction engineed and this is how both hemispheres of the the brain with feedback between each other model and they adapt. They allow for adaption. This is different to AI models. AI models are when we look at it are very left hemispheric. They're all about crunching data, simulating, abstracting, creating models. But where they lack this ability to define self is they don't contextualize that within the greater environment. And this is where both hemispheres of our brain seem to be from what these these writers are telling us, these neuroscientists and philosophers etc. is that the sense of self is created through the feedback between mapping the internal model that's being generated by your brain versus what is out there and how well that model works with what is out there and that's where you get this sense of self and for a brain to work and Neil Seth and Ian McGrul argue it's not just needing a brain it's needing embodiment a body with senses measurements etc all of these things is necessary to create conscious human beings and that they they argue that AI has got a long way to go uh before they get to that level of selfidentification, self-recognition. They're marvelous at mimicking, simulating left hemisphere work alone, but very bad at that they don't have a sense of self. And the thing is when you get this sense of self with the right hemisphere and left hemisphere, it tells you how aligned you are with the external environment. And if you are not aligned with the external environment, it it creates symptoms such as pain, fear, all of these symptoms uh which your your models then recalibrate to remove that and your model adjusts. So it's this continual feedback between right and left hemisphere. And what I was thinking was that well that's how our brains have evolved in these environments and it's been very successful in these environments but the markets financial markets is not something we've evolved in. And this is where I think actually we want the left hemisphere only. We want to remove the the right hemisphere's involvement in emotion, feelings, all of these things. And we wanted our rules-based systematic processes to be purely left hemispheric. And this is where I think AI probably does work very well in the financial markets. In fact, AI, if you think about it, has evolved in these financial markets. Humans haven't evolved in these financial markets. And so a discretionary trader comes across all of these emotions which comes from a a natural environment it evolved in worked in that environment was necessary for survival. But within this context they're the things that are really constraining its ability to follow rules to not fear feel fear when you get increased volatility. all of these things which is necessary in these financial markets because as we'll get to in the later topics we'll talk about things such as feedback reflexivity all of these issues that show that the markets are not these stationary entities that these alive markets compress and expand etc. And within that context, you don't want the the right hemisphere to interfere with that context. You want rules, systematic processes. And you want data to be your guide, not your emotions to be your guide. So I just thought it was interesting and seeing well what what's the difference between AI in this context and the human brain in this context. And that's the sort of conclusions I was reaching. Now without going into too much detail right now because as you say we're going to talk about uh a lot of stuff today uh in a few minutes but when you say that AI has evolved in the financial markets how how do you see that because I saw some comments and I if from some of the big well-known hedge fund managers I don't know if it was Ken Griffin or Paul Tuda Jones it seemed like some of those guys who came out and saying well you know AI is far away from finding anything that and you know produce alpha whatever there might be some efficiencies in programming etc but it's not really something that's going to revolutionize um our production of alpha so to speak um so what did you mean by it when you said >> so what I mean is that it's evolved in the context of pattern recognition >> so um early forms of of of processes that were looking looking at pattern recognition has now sort of got to the state of AI artificial intelligence where where it's always backward looking. It's always assessing the data. It's always looking at patterns and what it's attempting to do is predicting forward. >> That that's what AI is trying to do. But as we'll see in the discussions later, prediction in these markets is like prediction in the weather. It's ephemeral. It's got a very short range. And this is because of the reflexivity in the markets. And we'll talk about how when a trader interacts with this this market, they change the nature of the market. Now, because of this reflexivity, what I think's going to happen with AI is that the prediction horizon that is currently there with things like convergence strategies is actually going to much shorten. Even though it is short, it's going to compress to be much shorter, which is going to mean that the the influence of AI is going to create much shorter prediction horizons. And I think this is wonderful for trend following because what it's going to do is it's going to reduce the impact of prediction and create much more structural alignments with things such as um um consolidated concerted directional behavior. All of these things which I think is very good for trends which as we'll discuss later on in the the episode are more uh from patterns to structure. It's more structural features of the fractal nature of these markets. It's not about prediction. Prediction is something now that I think is even going to compress further the window of prediction and that's because of this influx of AI. I think that's that's where it's going. So I think they're it's going to def defeat its own purpose in these financial markets if you know what I mean. >> Uh yes. No absolutely and as people can already here it's going to be a super educational uh conversation once again with you uh Rich and um can't wait to dig into it before we do. So on my radar I mean I couldn't help noticing the um sudden evolution not evolution explosion of volatility in the precious metals sector we've seen this week. Um the gold if people don't follow this um has had it [clears throat] had had its biggest price drop on Tuesday I think it was in over a decade. Uh it lost more than 6% in a day. Uh silver and platinum losing even more on the day and palladium has given up 16% of its price in just 2 days. So lots of volatility showing up in this sector. um which will lead into a conversation we're going to also have later on about position sizing which will be uh fun no doubt. So that's interesting. Um the other thing I just noticed was a note that um uh I think the former chief strategist at Saxopstein Yakabson uh sent out and um you know it was kind of interesting when you think about the role of oil and how a lot of people have sort of been focusing on on oil um in terms of its its role its importance um and of course uh we're recording on Thursday and we saw yesterday that the US uh introduced further sanctions against a couple of oil companies in Russia and the oil prices is up uh quite sharply uh today 5% right now uh crude oil up as we speak but he writes the following he says the world is short of electricity data centers need juice cooling systems need water grids are maxed out base load is insufficient where economist once watched oil prices to gauge cyclical swings electricity is now the core or input to growth, how to get it, at what price, and with what utility to society. These are new macro questions. Oil was the lifeblood of the post Second World War economy. Electricity is the successor. Let that sink in. A society short of infrastructure, energy and delivery mechanisms mechanisms cannot grow. Inflation will rise. Productivity will stall. I mean, kind of interesting observations. Um and I'm not so sure that um that the world is really looking at electricity the way uh stain Yakosman is describing it in relation to oil. I think still people are more biased to looking at the oil price but it hasn't been doing anything for a long time despite all the challenges we've had in um in the Middle East. Um, and I think he also describes the current um, uh, or at least this is my interpretation of it. He also describes the current debacle in private uh, private credit um, and uh, as as something that maybe it's a new acronym that we have to get used to. D A DT for markets. Don't ask, don't think. So, so we'll see. We'll see. Anyways, in terms of trend following update, the environment I would say continues to be pretty constructive. Albeit uh there is more volatility in the uh returns from managers um at least when I see the ones I can see on a daily basis uh including our own uh for that matter. Uh precious metals although they come a little bit under pressure it's still pretty constructive uh this month. Um but you know big swings um which may illustrate some short-term top correction. Who knows? Equities continue to be fairly well uh behaved from a trend following perspective. But the new kid on the block, not really new, but certainly this month seems to be also helping out is the soft sectors. Um we have uh markets like sugar supporting coffee, cotton uh in terms of being a source of returns for trend followers as far as I can tell. And then we have as I mentioned energies and currencies pretty much stuck in a range not doing much uh for the moment. What is your take on on on sort of the the state of the trend following uh environment at the moment, Rich? What are you seeing? >> I'm pretty bullish about it at the moment, Neil. So, of of course, we had the the the metals um retracement um to be expected considering their their rises they've been having all all of them, you know, gold, silver, palladium, platinum, >> you name it. Um but um I I think it's a very robust environment for trend following and um you know I'm keeping my fingers crossed. I think we're going to probably come out of this year I'm keeping my fingers crossed strong um which is a welcome relief from the the first 6 months of this you know that that was when you know of course we never know um what's going to happen in the future and I had all of these bad thoughts in the first 6 months I was thinking Christ our models aren't working anymore. No, of course you do. And uh but you just got to stick to the process. But it's it's a robust environment. Now, I think um I think we're in for some good things. A lot of the in my sort of explanation of the fundamental reasons why these things happen. um the more I'm getting into it, the more I'm I'm very comfortable uh with our methodology and process to the extent that I I think that um you know the events we saw post GFC up to 2018 I think were the anomaly. Um I think now we're in an environment where um is trend heaven basically uh for a period of time. Of course, what stops that is these um um you know, everyday uh uh change in policy sort of that um creates this whipsaw environment and that that's what we experienced in the first 6 months which made me think if that continues and we continue to get those whipsaws it'll be death by a thousand cuts. Um but I think we're out of that now and I think that the markets are reasserting themselves. Uh we're getting some solid trends and it's great to see the soft start picking up though quiet for a bit. Um but um you know the energies they're turning. They're going bullish. Um the the metals I think you know it might be a bit of a breather for them. Um but I don't think they're necessarily over yet according to our models. They're certainly um we're still sort of active in those those um things. So it's it's a pretty good environment to be a trend follower in Neils at the moment. >> Yeah. you you mentioned briefly and then it wasn't really part of our plan but but I'm always curious you know as time passes um and we get wiser hopefully uh or at least we see things differently um you mentioned this thing about post GFC um that period [clears throat] that difficult period although I will say it was only a few years that really were difficult there were also plenty of good years uh as well but there were a couple of years that were difficult um do you see those years now any different or do you have a better understanding as to why they were more challenging? I have my own views but I don't necessarily need to share those but but do you have have you become any more certain about why specifically a couple of those years and I think mostly of 20 frankly 2018 through 2020 or thereabouts uh I think both early in that in that decade uh there were some good years like 2014 was a pretty strong year um but how do you do you you see >> different I view that as um it was period of what I call compression. In other words, it wasn't a period of equilibrium at all. So, and this is what I'll get into a bit later in my topics where I talk about um >> these markets are way away from equilibrium. So what was perceived as calm was actually compression and you know that that old saying that um energy can um you know the conservation of energy it it might not have a visual appearance that it's there in a in a compressed state but it is there and it's how these markets compress and expand and I think through quantitative easing and central bank coordination we had this period of of trend suppression and compression but I I think the surplus of that has been felt from 2020 onward where the the markets had exploded out of that compressionary state. The risk never went away even though it appeared calm over that period but it was like an avalanche. was slowly building snow, snow, snow, snow. And then in 2020, it flipped out of out of this compression and boom, we had this amazing explosion in trend. And now I see this environment with the decoupling um that I'm seeing now and the the lack of central coordinated >> um actions. I see this decoupling as an environment where trends are just very favorable. um uh we we things are less correlated than they were when when things were in compression in convergence. Uh we we get um buying the dip syndrome I think was something of the past. I don't think buying the dip is going to continue going forward but that was this sort of um um this phenomenon this behavioral phenomenon associated with this compression. Um >> yeah no I agree with that. I'm actually I think I completely agree with that. I think it had a lot to do with the uh with the success of the central banks actually and I think now that they are doing their own thing and they're not really coordinating because they have different challenges, different problems. >> Uh I completely success of I think what they've done is they've kicked the the can down low with compression. >> Yeah. No, what I meant by it was they succeeded in keeping uh inflation low and stable. Yeah. and and and I think that's what they succeeded whether they kind of completely controlled it or not but that was the result and I think I think people and myself included I think we've underestimated historically the importance of not necessarily just the level of inflation because that dictates the level of interest rates etc etc but it's actually um the stability of inflation how much that really impacts uh the markets we trade and therefore the opportunities we we we um we see. So, so that's what I meant. But anyways, we'll we'll get to into all of that good stuff. Let me just mention that the trend barometer yesterday finished at 52. So, that is a strong reading and uh >> it's coming up and uh supports uh what's going on because uh the data um as of Tuesday evening um and I think by the way yesterday was a slight negative maybe mix day for for for the space. But anyways, as of Tuesday evening, the Btop 50 index is up 2.12% for the month, up now 2.58% for the year. Uh so CT index up 2.35% uh for the month, down still 40 basis points for the year. Trend index up 2.85% in October up now 53% uh sorry not 53% up 0.53% for the year and the stock short-term traders index up about 1% but still down 4%. Uh so far this year, MSEI World up 40 basis points in October as of last night, up 18.3% for the year, very strong. S&P US aggregate bond index down 44 basis points in October, but still up very strong 9.8% for the year. And the S&P 500 total return up 22 basis points uh as of last night, up [music] 15.08% so far this year. >> [music] >> All right, enough said about uh all of these things. We definitely need to get into um your agenda today, Rich. You put it together um as a true professor. And um I'm going to basically pass it over to you. I'm going to try as as a good student to keep up um and and maybe have a question or two along the way. All right, Neils, we've got four topics today, but um I I thought I'd take this today's conversation in a slightly different direction where um rather than focusing about a determined universe and things like that that we get into in past conversations, we'll stick to the markets this time. Um but I'd like to uh dig into the foundations of how markets actually build themselves. >> Mhm. So we'll start at the smallest possible scale with a single trade and trace how structure then grows upward through feedback and interaction from impact to fractal geometry which I'd like to get into. So um we'll then move into how that scaling creates trends and how patterns differ from structure and finally what all of this means for risk survival and how we position size using closed equity and then uh that'll give you the opportunity to jump on board. I know you want to say something about that. >> So we'll begin at the beginning with this concept called impact. So every price move starts with a decision. So a trader can buy, they can sell or they can do nothing. That single action no matter how small exerts a force on the market and that force is impact. It's the fundamental impulse of price movement. So most people think price moves because information is revealed but markets move because impact is applied. So that's the difference in this interpretation. So each trade order changes the balance of supply and demand and therefore reshapes the conditions for every trade that follows. Can you see that slight reflexivity there? >> Yeah. So at the micro level, individual impacts interact through the order book and liquidity network. Some cancel, some impacts cancel and some impacts reinforce. When reinforcement occurs, feedback begins. Now prices therefore no longer just reflect information. They therefore create it through this feedback. So traders they observe price they therefore once they observe the price they then update their models and then act again. So they'll make initial trade they'll see what the impact on the market is then they'll update their models then they'll act again and this loop continues and that's reflexivity in action. It's where a self-modifying system learns from its own output. So here we've got an input, a trader making a decision. We've then got an output response, what the market responds to, and then an adjustment with the next input that comes in from that trader. There's just reflexivity involved. So that's reflexivity and action. Now, markets don't require rational or informed participants. They simply require actors whose behavior feeds back. Each irrational trade changes the landscape and become part of the evolving structure. And that's why markets are seen to be self-organizing. They evolve through interaction, not through equilibrium. But here's where it gets really interesting. Market impact is not linear. A trade 10 times larger doesn't move the market 10 times as much. It's not a linear relationship. M so Jean Philipe Bashard's research shows impact grows with a square root of trade size and what that means is if you double your order you only increase expected impact by around 40%. Not the same amount it's a reduced amount and that's because liquidity is adaptive. It's not a static pool. It's seen to be like an elastic surface that moves and reforms with order flow. So every trade that we do consumes liquidity and signals potential direction and other traders react to that quotes adjust and the whole system bends. This is where we turn away from linear to nonlinear. This is what gives u markets their nonlinear fractal character. small trades disappear into the noise. That's because it's a sublinear rel relationship, not a superlinear relationship, a sublinear um relationship. Small trades disappear into the noise. But when many small trades align, feedback compounds, volatility clusters, and directional bias emerges. And that's how trends, crashes, and factiles are born. not from randomness which the efficient market hypothesis will tell you but from interaction. So if these events were independent like a gorian model would would say they are independent random events returns would form a bell curve but in markets events are all conditional. Each one changes the probability of the next because of this reflexive nature and this is what destroys erodicity in these markets and replaces it with path dependence. Um so I know you don't like that term erodicity and >> well I want you to just remind people what it means. So um an an erotic system is one where an ensemble of um results are the same as the um an individual across time. In other words um it's saying the ensemble of statistical results at a particular point in time is equivalent to a single um outcome statistically over the course of time. The both are equivalent. But in a non-gotic system, there's an asymmetry there. And we find that wealth paths in financial markets because of compounding this geometry, they're not um that they you can't apply erotic statistics. In other words, the the statistics break down. A good example, the expectancy equation is what we call an erotic statistic. It says that um if we remember the um expectancy equation, it says your percentage win multiplied by the dollar win less the percentage loss multiplied by the dollar loss gives you an expectancy equation. Now this therefore says to the trader ah if I have positive expectancy I will be profitable. However, when you look at that equation, what you don't see is the sequence of events. And this is critical. And this is what Ole Peters found when he found that markets are non-gotic. And this is because there's a conditional um reaction that occurs with a finite sum of money which has a lower bound of zero. Um so it's an asymmetric system. It's not an open-ended system where egotic systems are. The the reason why we get the equivalence of ensemble averages and time averages in an in an agotic system is because they are open-ended on either side. No lower bound, no upper bound. And there's basically stochastic movement available across all dimensions. But in a non-rootic wealth path, there's an asymmetry in the system. There's a lower finite bound and there is an open-ended bound. And this is what we find in fractal systems. And we also find that is why they are non-aggotic. Which means that in those systems no longer is the um the statistical account of an individual over the course of time the same as the statistical ensemble at a point in time which means that the path um of compounding uh is different to what the statistics are going to represent. Now this is the problem with the Gaussian theory. The Gaussian theory, this independent um independent um random events um produced a bell curve and it assumed that there was no asymmetry in that. It was open-ended upper bound lower bound and we got all of the associated toolkit with that which were the statistics that made that model work which was things you see in modern portfolio theory. Marowitz's efficient frontier, sharp ratio, standard deviation, um, expectancy equation, risk of ruin equation. All of these assume an erotic system. But as reality tells us, it's actually a path dependent system and all of those toolkits break down. And that's because we get in these asymmetric systems that are fractal in nature, we get power laws, we get non-stationary environments, we get a a decrease in the predictive power as we extend into the near future. Um we get um no equilibrium away from equilibrium. All of these things are saying that Gaussian model is is incorrect. And then we got um this this instance where I think in the 1960s Benoir Mandelro came out and said look we can demonstrate that these financial markets are non-Gorian because all of these rules are not being obeyed by the financial markets and he came to the conclusion that it was a fractal system. That's the conclusion also. >> Yeah, >> I've sort of come to realize over the course of time. >> Yeah. So no, >> if these events were independent, returns would form a bell curve. But markets events are conditional. Each one changes the probability of the next with this reflexivity. This is what destroys eroticity and replaces it with path dependence because the path matters. What happens previously, there's memory in the market. Each of these events because of this reflexive nature of the market are not independent. What happens before shapes what happens next through this reflexivity. trader impact. This therefore means a sequence that over time there's a memory in the market and it's captured by things like the hertz exponent. uh we've got these addition new tools we can use that says if there is market memory um using Bord's analysis we've got a different set of tools in our toolkit which aren't gausian but Hurst exponent tail properties tail decay all of these things in a power law driven system these are better tools to use that are more representative of our market says that the old tools could never account for all of the calamities that occurred over the course of time by relying on a model that had these assumptions that did not reflect reality. And hence anyone who relied on those tools, portfolio managers, industry, all of those things would always get blindsided regularly much more than what a Gaussian um distribution would imply. And hence, you know, um a Gaussian distribution says that a five standard sigma event should occur every 170,000 years. But when we look at I looked at the S&P 500 or the ES uh futures market for instance and in the last 30 years there's been 32 five sigma events and this isn't just associated with a single market. Every single liquid market I examine has these tail properties fat tails. This is saying they're non Gausian. They're fractal. um the and and the look the verdict's out whether they are exactly fractal but certainly a fractal model has a better account and can understand why these situations occur. So I'm not saying it's a definitive answer. I'm just saying at the moment fractal models and the fractal market hypothesis probably is the closest we get to reality because it explains so much more. So, let me ask you one thing while you uh get a sip of water for your for your throat. Um, >> and I don't know if this is um correctly understood, but but maybe there's a question in there as well. And by the way, this all this fractual stuff um I think you and I touched up uh on this before we pressed record. It does remind me of my conversation with Bill Rice. I think he actually back then >> um in the 70s when he started um built his pattern recognition model based on fractal um you know fractals. So anyways, but if volatility is an output of reflexivity, not just exogenous news, so to speak, are we underestimating the role of our own participation in creating the risk we seek to avoid? The risk we seek to avoid. Okay. The bottom line is we can't predict these things. So we might um we might have tools that say um that volatility equals risk. And this is exactly the problem, the dilemma I find in that I don't associate volatility with risk. I view markets as always having this potential risk event around the corner. >> What I view is um low volatility regimes I actually view as high latent risk events, warehouse risk. In other words, the risk hasn't gone away. It's there, but it's now compressed. It's like a sponge that you compress and it's going to explode at some point in time, waiting for a tipping point. Um, when we see this massive expansion, this transition. Um, so I'm seeing the market breathe, contract, and expand, contract, and expand. I don't think we can measure volatility using the standard statistical tools we've we've done. I can understand what volatility. mean we can't measure risk >> no >> or volatility because we can measure volatility right no >> I we can measure volatility but then there's um directionless volatility there's directional volatility um one might come with correlation might one might come without correlation um it's it's a signature of a fractal system that's alive compressing and expanding but because of the the undefined nature of predictability in fractal systems And in other words, what I'm saying there is that whilst there, you know, when you look at the weather, which is a fractal system, there is a limited prediction horizon within weather. You might get a fairly accurate assessment within >> 10 days or 14 days. That doesn't preclude the the ability for these large events to interrupt that prediction, but it does in certain regimes give you this prediction horizon. But over the long term, they are they are unpredictable. that they are deterministic systems that are undefined and this is where they follow these strange attractors and in the financial markets which I view very much similar to weather systems where instead of molecules of water um in in the the clouds etc. I'm viewing agents traders all as collectives that there's no central governor there. all working according to their own mandate but they are interacting with each other and these impacts and feedback loops that occur between each other make it calculate impossible to calculate or impossible to predict um so I I refer to them as deterministic um unpredictable systems which a lot of people say well that's chaos and I say well no um we get periods of very rational order um you know it's this reflexive nature of the market they adapt they respond to what the participants are doing in it. But so I don't have I prefer what I call engineered outcomes rather than statistical um statistical tools to define how to protect myself in this market. So when we talk about volatility, I'm not using volatility measures to protect me or statistical measures because I think they're they're unsound. The the assumptions of these statistical measures and the toolkits of statistics have come from a gauian world. I'm saying you got to think more like an engineer. So in fractals, unfortunately the thing with fractals is these Gaussian models are just too simple to describe this very complex thing we call the markets. You've got to be an engineer. A bit like an engineer designing a bridge in a weather system. You you don't blame the weather for blowing down the bridge. You blame the bridge being too brittle. So, it's the engineer's fault if they break down because the weather, as we know, can be very unpredictable. You can get typhoons. If the bridge falls down, that's that's because it was optimized, overoptimized, brittle, could not handle the environment. The same way as I think our systems need to be engineered to not be brittle, not be overoptimized, not be not be overfit. They've got to be robust, resilient, able to stand up against anything that thrown against them. That's that's how I view robustness. >> Yeah. No, that's that's a great way of looking at it. All right. Um, was that topic one? >> So, let's go to topic two. So, we talked about impact, >> right? >> Uh, and now what I want to talk about is how fractals create structure. So, let's move up the scale and see how these micro impulses of of impact assemble into the structures we call trends. So when you zoom right in on a market chart and you zoom right down into the detail, it does look chaotic, jagged, noisy, directionless. But as you zoom out, hourly, daily, weekly, the chaos starts to organize visually. What looks like randomness when you're zoomed in actually becomes rhythm as you zoom out. And that transformation is a signature of fractality. This is different to a gauian world where as you drill in the structure dissipates. It disappears until you're left with linear results. That's just this linear independent gausian world. But in fractal systems, you can never get rid of the structure. It's always there's structure in there, but it looks chaotic. You're going down into zooms. The the structure never dis disappears. It's always there. But see at this scale the different um views of scale perceptions of this at different levels of of perception what looks random at one might have structure in the in the other patterns in the other. It's a bit like looking at a TV screen uh when you're looking at the pixels and you get move away and you're looking at a different resolution you start seeing the images um that are are linked together from the pixels. This is how we've got to understand fractality. So, u fractality isn't about prediction. It's about relationships across scales. Each time frame represents a layer of feedback. And we'll find that people are responding depending on the time frame they're interacting on the market with the patterns that are observable at their scale. And so, you get this this scale difference occurring. high frequency traders at the lowest end of the scale see markets a different way. They're seeing this sort almost chaotic frenzy of movement. As you step out, you start seeing more and more trend followers start participating in the market. Why is that? And it's because as you scale out, you you start seeing these these impacts we're talking about with the fractal nature of markets. Some are cancelling, some are reinforcing. the reinforcing elements of these structures is what's creating the trend as we're zooming out rather than a cancellation. So mean reversion is is a is a a cancellation >> um environment. It's where we get opposing forces. So we get a force of of um reverting back to an equilibrium and then we get a force of going away from an equilibrium. This alternating zigzag up down up down up down. They're almost linear in in nature. They cancel each other out as we go out in resolution. But directional impulses that aren't canceled out start aggregating together and compounding. They're compounding structure into the market. So this is why convergence systems are linear results. And this is why when you look at the P&L of a convergent system, you get a linear profit, a linear profit, a linear profit, a linear profit, a linear profit until they come headon with a negative skew event where you get a a nonlinear loss. But when we look at trend following, it's the reverse. We get a a small linear loss, a small linear loss, a small linear loss, a small linear loss. When we come up against an outlier, a structural a structured directional trend we get a nonlinear gain. This is the fundamental difference between it and this is because of this cancellation and reinforcement that occurs in the market at the fractal level level. So when short-term reinforcement persists and aligns through time through scale you get coherence and the coherence is what we call trend. It's also called serial correlation or bias. So positive feedback drives reinforcement. Buying therefore attracts more buying. That's how it works. This reflexivity. When people see a trend, people start jumping onto the ride. Jumping onto the ride. Buying begets more buying. Negative feedback, however, drives regulation. Selling that restores balance. One is moving back to equilibrium. one is moving away from equilibrium trend directional positive reinforcement buying begets more buying or selling begets more selling moves away from um this equilibrium zone and negative feedback is the reverse a restoration trying to get back to uh this is what we call regulation um selling that restores balance markets oscillate between these two drivers and when positive feedback dominates energy compounds and structure emerges fractally. This dance between expansion and contraction creates a fractal rhythm of markets. You know, you get quiet periods punctuated by bursts of volatility, compression followed by release and feedback is the pulse of adaption. So what do I mean? People are reacting. They are they are impacting as they see a trend and they're buying. They are accelerating that trend. When people see a trend and they want to revert against that trend, they're they're applying negative feedback and trying to restore balance. But you see how this works. So when feedback cascades through scales, small interactions become large outcomes. So [clears throat] cancellation gets rid of structure. Positive reinforcement creates structure and small inter when it cascades across the scales, small interactions create large outcomes. A local burst of buying becomes a cluster. Clusters form rhythm. Rhythm becomes flow. That's how feedback builds structure through time. It also explains Neils why you and I are medium to long-term trend followers because the frenetic activity that occurs in the shorter time scales are more mean reverting in nature. there's more cancellation going on relative. But as we get out to the the higher sales, we'll find that the feedback tends to operate on the structure, not on the mean reversion. When I say feedback, most participant interaction, institutions, etc. out to the medium to the long end are positively reinforcing trends. They're not going against the trends. Mean reversion is something that occurs in the the finer time time frames. So it's the same in nature. ripples becoming currents. Currents forming rivers. This is positive feedback in natural systems. In markets, local feedback becomes directional flow. So take crude oil for example. At the micro level, traders buy ahead of a report for instance. Algorithms detect it and they join. But on the hourly chart, those bursts at the micro level appear as clusters. Zoom out to the daily and a narrative forms. Oil is recovering. Come the headlines. Oil is recovering. Starts turning into narrative. At the monthly level, that narrative becomes systemic. Producers, investors, and policy makers all reinforce that movement. Feedback has scaled from the tick to the macro. The market has literally organized itself around its own success. And that's what it means from frames to becoming trends. So eventually though every feedback loop reaches its limit. The same alignment that built the trend becomes its constraint. Positive feedback will flip to negative feedback at some point in time. Rising prices ultimately exhaust buyers. Valuation stretch. Risk becomes concentrated. one event or simply fatigue breaks as a symmetry. Then selling accelerates, stops trigger, the system unwinds. Maybe even margin calls come into play, the system unwinds. Reversals aren't random. They are feedback inverting. And because energy stored through the long positive reinforcement, you'll find that during these reversals, it's released quickly. Reversals are fast and violent. They're not like the the typical buildup that occurs on the long side. These reverses are fast and violent. Exactly what we saw with gold and the metals earlier this week. Now, so this feedback flip isn't failure, it's renewal, system renewal. It resets the system so the process can begin again. And it's this continuous process, this continual cycle, reinforcement, alignment, reversal. It's the living rhythm of a fractal market. So that's topic two. >> Yes. I'm going to save any questions because uh again we have a quite a bit to get through. Um and so I'm going to give you as much time as possible to to move on in your own narrative so to speak. >> So now I want to talk about the difference between what I'm referring to as patterns structure and the fractal nature of outliers. So, >> okay, >> now that we've seen how structure builds through feedback, we'll explore the difference between pattern and structure and why the outliers matter most. So, of course, I'll say that cuz I'm an outlier hunter as you know. >> You >> okay? Markets are full of patterns. We all know that. We see flags, we see triangles, we see breakouts, we see moving average crosses. These are all surface forms, transient, visible, easy to name. When I'm referring to structure, I'm referring to what lies beneath. It's the geometry that govern how markets behave, not how they look. Now, I remember in a previous podcast with you, I've talked about an outlier hunter um doesn't have a prescriptive definition of what it refers to as trend. It's looking for the structure, the things that create the bias. And that can create come in many different visual forms. But those people that are looking treating trends as patterns are probably too prescriptive because these trends, we're looking for the structure. What drives these trends? Because as you know, Neils, trends can actually be a a random result from from um no bias in the price series. We can get a random trend very easily. We can get a a a trend that actually is a a segment of a mean reverting cycle. Um or we can get these structural trends with serial correlation in them that gives persistence into the future. That's what I'm calling outliers. These are structural things. They're not patterns. It's created by causitive drivers um that actually create this vast array of different directional patterns which I'm calling uh outliers. So it's the geometry that governs how a market behaves, not how they look. It's built from the relationships, the incentives, the feedback loops that determine how energy flows through that system. So patterns describe what price does. Structure explains why it does it. So patterns are what we observe. Structure is what connects those observations through time. Notice that structure is what connects those observations through time. It's talking about a memory. What happens before happens later. This is a serial correlation. This it's not independent that there is a a time comp over time. This feedback occurs over time in these fractal systems. So trends are patterns out outliers I regard as structure revealed. An outlier is not an accident. It's a phase transition. It's a moment when feedback alignment pushes a system far from equilibrium and forces it to reorganize. It's not just a mere pattern. It's something structural that's really changing the system. Compression. Now compression, you know, I talk about compression and expansion. Compression hides structure. Expansion reveals structure. Outliers are those expansions. The moment when the market changes its own geometry. So in fractal systems, small fluctuations typically cancel, but large ones dominate. The tails contain the power. That's where this adaption happens and where returns are made. That's why I call myself an outlier hunter. I'm not chasing a price pattern. I'm aligning with structural change. That's how I view it. So I design system ensembles which are families of reactive systems that listen for different ways structure might express itself. I don't trade single trend following systems. I put in these ensembles. They are what I call I'm not looking for correlations here. I'm looking for behavioral orthogonality. That's a big word. So in other words, >> orthogonal organic. >> Yeah, I'm going to let you pronounce that word for sure, but but so you're essentially saying diversifying by behavior. >> Yes, a systems behavior. That's why I will consider a breakout. I will consider a mean reverting into a trend. I'll consider these these are what I call behaviorally orthogonal. They will never all act in concert together. They are structurally looking at at the different manifestations of how trends can form. And it's not saying there is one prescriptive form. It's saying I've got to diversify across as many because this structural outlier can come in a array of different varieties. So I'm not looking for correlations here because these are very fickle things. I'm looking for structural behavioral differences between things. That's that's my choice of >> you know the risk you know the risk of this Rich is that you are maybe less classic than than what [laughter] what other what other trend followers might perceive as being classic but I like the idea. >> Yeah. Okay. I look I'm I prefer to call myself an outlier hunter. >> Yes. Yes. No I know. I know. >> So these different systems okay one for instance might detect volatility expansion. Another might detect smooth persistence. That's that's looking for a breakout from a congestion Darvis box breakout. That's >> another might focus on breakouts. Each hears a different voice of the same feedback process, but it's relating to structure. Together, >> they form a coherent adaptive framework, not predictive, just ready to strike and activate when the signals erupt. That readiness is everything to me. And because we can't know which move will reveal structure, but we can ensure through this system ensemble, we're alive to capture it when it does. So what I'm now getting to, there's a shift in the topic here. What I'm getting to is the need to survive until the outliers arrive. Okay? And this is where uh so to summarize topic three I'll say patterns describe form structure defines cause outliers are structure made visible through feedback. That's why in a fractal world we don't diversify by correlation. We diversify by behavior by how our systems respond to changing structure. Okay. So now I'll get into topic four. The last topic and this is where we come to our a possible debate coming up. Neils. So this is I call the fractal reckoning part dependence survival and closed equity. Okay. So I'm going to connect all of this um to the most practical question of all. How do we size positions, manage risk, and stay alive in a world that's part dependent and fractal, non-stationary, um unpredictable, uncertain? How do we survive in that world? In an ergotic world, a Gaussian based world, um, like a casino, a casino is a good example of an ogotic system, the average outcome across many plays equals the average outcome through time. You know, I talked about the average outcome across an ensemble is the same as the average outcome over time. That's in a casino. We've got that situation. However, markets are non-gotic. We live only one path. We can't live in these parallel universes that statistics say we can. We have one path. And once ruin occurs, the game ends for us. Once ruin occurs, game over. That's why survival, not expectancy, is the true measure of success. This is why I harp on what's the best method of determining the best managers track record. Survival. Survival. This this comes down to this conclusion. You can't measure this statistically because the statistics is the wrong toolkit because the statistics come from the Gaussian model. So Kelly criterion bad statistic expectancy bad statistic I don't want to say this Neils but I'll say variance at risk is a bad statistic however I know your opinion on that but all of these are coming from this Gaussian model which isn't the reality as we've observed. So in a non-aggotic world, the path is everything and the order of wins and losses determines whether you compound or whether you collapse. So think of two traders with identical expectancy. One experiences losses early and they run out of capital and they never reach the recovery phase. The other survives long enough for the outlier to arrive. Same expectancy but completely different destinies. They're part dependent. Wealth doesn't grow additively. It grows multiplicatively. Every trade we make changes the base from which the next trade grows. That's this reflexivity and action. Okay? It's not independent reflexivity. Every trade we make changes how the next trade grows. Losses shrink the base and that drag compounds. That's why minus 50% followed by plus 50% doesn't equal zero for your wealth. It equals a permanent hole in your wealth. So if we have $100 and we lose 50% we're down to $50. But if we gain 50% we're only up to $75 or whatever. We've got a permanent hole in place. This is this sequence risk which isn't addressed by the Gaussian world in oootic systems. Part dependencies everything. So this is where closed balance equity I believe becomes crucial. We use closed balance not floating balance or or equity to size new positions because to us it represents realized capital not illusory equity. What I mean there is closed balance is the only equity you actually possess after the path of returns has spoken. So what I mean is it already incorporates the non-aggotic journey every draw down every recovery that is what it represents. By sizing from close balance we ensure that our risk per trade remains constant relative to survivable capital. It automatically adjusts exposure downward during draw downs when the system is under stress and scales up only when the process has rebuilt through trades that have been confirmed. This is a structural response as far as I'm concerned to a non-aggotic environment. Now I [clears throat] can for instance demonstrate to you how a system with uh 5% positive expectancy but have um large size positions um and at even under that arrangement we get this massive compounding drag and we never get a wealth return out of it. It goes to zero. Reduce the size of those bets significantly reduce the size of the bets. I can show you a system where we've got a 30% win rate, a 70% loss rate, a three times um win win amount to a a one times loss amount, but with very small positions, the expectancy is only 2%. But compounded geometrically, you get a very very good return. >> Now, this is this path dependence sequence. So, this is a structural response, I believe, to a non-aggotic environment. It keeps us in the game long enough for the next outlier to appear. The closed balance principle is a way I believe of embedding sort of humility into the system. It acknowledges that we don't know what the next path will bring. So we size from what is real what has survived. We're not trying to forecast variance or optimize volatility. We're simply aligning our capital to the geometry of survival. In a world that compounds through feedback, the key is not maximizing return. It's avoiding ruin. And that comes from this fractal mindset that I have. Prediction fails because it assumes a world stands still. Process succeeds because it accepts that it will never stand still. Um, so that's how I view this issue with closed balance. So if we can imagine when uh because I've got positive skew in my trend following models my open equity or my realized equity always sits above except at the beginning when I start it always sits above my closed balance equity which means I am much more conservative in applying position sighting because I'm using a lower limit to compound. I'm not using um the higher limit that equity provides which is unre a large component of that is unrealized at this point in time and I don't know the way that's going to go. If I for instance assume that um um it can support this higher position sizing. I'm starting to leverage up and it's starting to get towards that model I was talking about 5% expectancy but very large position sizes is starting to get me an unfavorable compounded path. That's how I see it. But over to you over to me. Okay. So a lot of the stuff that you said I completely agree with. Um I think my um comments um maybe more in terms of how uh the difference between how most managers um size positions and uh what you and and a handful others refer to as the classic trend followers. I think it's more than the the way it's being described that um concerns me a little bit. Um the way I hear it when you and some of our friends talk about uh and maybe not so much you but talk about this difference is that one is right and one is wrong and that closed using closed end equity is right. And I'm I'm I'm I'm thinking well there must be a reason why the majority at least of managers are not using that methodology. Uh so maybe you can't really say one is right and one is wrong. Um you can say they're different. Now I hear the criticism of the camp that I find myself in that dynamic position sizing is only being done to reduce volatility. It's also sometimes referred to as volatility, you know, managing volatility or volatility targeting, which it's not necessarily the same. um that that you know that you're trying to manage volatility and you're trying to make this the the ride a bit smoother um as if there's something wrong with that. But then when I hear uh this closed end equity position sizing being described, I'm thinking well that is also reducing volatility because you're simply sizing your trades on a smaller equity. So that's reducing the volatility as well of that strategy and and and and thereby creating a smoother ride. So what is the difference? What is why is why is one wrong and the other one is right. So that that's the kind of the first thing that where I just think that the the way it's being described is is you know is not necessarily how I would describe it. And the other thing is you said something very interesting. You said that the open equity always sits above or the total equity always sits above the closed end equity only exception. >> Yes. Well, that is very important. >> Yeah. Yeah. >> Because in my view that kind of means that that method works only and I know I'm going to get some some some feedback on this. Um only works if you're trading your own capital because you were there at the inception. But if you're trading an a a a fund with external clients that comes in every month, then they are not necessarily at inception of that open equity um calculation. And that to me opens up and a risk that is kind of undefined because they're you're taking on a riskmanagement methodology that is not specifically to uh you know geared towards them but is geared towards quote unquote the original investor. So just putting it out there just putting it out there. the the other thing that I've heard you you describe the open equity as some kind of illusorary equity and and I'm thinking for me at least because we trade the most liquid markets in the world yeah >> we can all even the biggest managers in the world we can all liquidate our portfolios in 24 hours so it doesn't seem that illusory uh to me it's pretty deal. Um, so again that takes me towards this idea of actually you should consider your full equity as the basis of your your of your your trade sizing. And the final point I just want to um throw out there is just that having these rules are fine, but I hear from um from from this from your camp, not you specifically, but from your camp that on top of this you throw in something called the cutback rule, which is a completely discretionary rule um about when you want to cut down your uh position size. um you know when you when you feel the pain more than you're willing to feel the pain and you keep the positions cut until you hit a new all-time high. Well, at the same time, I hear that people who do dynamic position sizing, they're somehow handicapping their own lifting power in coming out of a draw down. And I'm thinking, well, hang on. If you're cutting your position size at a certain level of a draw down and you don't touch that until you get to a new high, that is cutting your lifting power as well. And if you're doing dynamic position sizing, even if you're in a draw down, and I do admit that that is not necessarily something that reduces your risk, but if certain things occur, you could actually see dynamic positions being increased whilst you're in a draw down. And that could be good uh in in terms of of lifting your recovery. It could also be bad if you're not doing it uh in a proper way. So anyways, my my point is that when I look at these track records, I don't see, you know, the long-term track records, people who've been around for 20, 30, 40 years plus, right? It's not that I see a massive difference in the returns. I hear a difference in the narrative and and and one being right, one being wrong. All I'm just all I just want to open up to and I'm not I I don't I I don't want to say one is right and one is wrong. I just want to open the debate about saying yeah there are different ways you can skin a cat and um at the end of the day we need to do what we believe in and what we think is right. So there's nothing wrong in what you describe. I don't think there's anything wrong in in in what other people do. Uh it's just a matter of preference. But I am a little bit quote unquote allergic to this. Oh, you're doing this wrong and I'm the only one doing it right. I'm a little bit allergic to that >> argument. And I'm not suggesting necessarily that >> we need to dust up. We need a dust up with boxing gloves. >> And what we need to do is get Jerry on with me and you get to get on with your compatriots and we have a a brutal debate about it. It would be fun. It would be fun. I think >> it would be fun. But more importantly, I think this shows and let's spin it positively. I think what it shows is that as long as you follow the golden rules. Uh I think there is a little bit of leeway in terms of how you become a trend follower and and I will admit as well that the way you describe it where you are focused on a trade by trade basis is of of course a lot easier or at least I think of it. I'm not a quad, so it's easy for me to say, but I think of it as an easier way to build systems, to understand systems, maybe um maybe with less moving parts, but also a few discretionary parts, which I don't personally think um helps, but but be that as it may, >> yeah, >> um I think the dynamic uh approach is more complex. It's not that I'm a fan of more comp complexity per se. Um, but I also think that after 25, 30 years where that's probably been around by now, it's proven itself. >> We've all survived, Neils. >> Exactly. That's what I mean. We're all here. Um, so again, I don't want this to turn out to be, oh, I don't like what this guy is saying or what what what this girl is saying. All I'm just saying is I I think we should I think we should just think about the narrative we put out there. Uh, of course, some of of the narrative will give better headlines, uh, for sure and it'll be, you know, fun and pro prerogative. So, I understand there's a little bit of that as well. >> Well, see, Neil, this is why I'm classifying myself as an outlander. I'm distinguishing myself from a trend followers. And the way I'm doing that is because I do have a particular set way in my mind of how to deal with these markets. So, >> I prefer to therefore say if I'm an outlier hunter, I've got specific objectives >> to achieve that outcome. I don't necessarily think we all need the same objectives trend, you know, there are lots of different forms of trend follower, but I'm trying to set myself apart to say, all right, I've got a distinctive philosophy. It is different to a lot of others. These are the reasons I like it. Um, and that's up to them. Gates of gray I would call because remember a few months ago you did a very good um description of >> four times four and you said I added five um descriptions of trend follower and I actually think that's a much more important topic than a lot of people realize um because it can be difficult to see the difference from the outside but even more so >> investors really need to think about why would you want one over the other because it comes down to what is the objective of making an investment with a trend follower in the first place. Yeah. >> Um I think that's argument. >> They view us as a nebulous not a lot but uh we're experts in our field and we we've obviously got our idiosyncrasies. >> Sure. And I think that that actually that conversation um is winning a little bit of traction uh among investors that it's we're not just one and the same. um and um and one should think about uh the type of trend follower they you really want and need in in your portfolio. Now the funny thing is of course when you say I consider myself as an outlier hunter in a sense even though we use pos dynamic position sizing I do the same because we both want the long longer trend than anyone could even imagine and we're all long gold or we're all long stocks and so this is the fun part. It's just the path and the journey. And do we keep the same, you know, exposure at all times during this? And of course, we often think about, well, isn't it better just to have the same position size if you're into a big gold trade or into the cocoa trade a few years ago? Of course, sure. But we do forget that there's all the other positions that can be troublesome. So if you have the same position in those, they can detract from the benefits you get from being in the same position in Coco and so on and so forth. So there's all these pros and cons and I just want to invite um you know a little bit of um um not so much black and white if you know what I mean but um it's doesn't take anything away from all your observations of course and the fact that it has worked very well for people using this approach for many years at the same time using dynamic position sizing has also worked very well for those people doing that. So, which is the beautiful part of of this um conversation. Anything you want to push back on, Rich? And feel free to be brutally honest. >> You you get the last word. It's your podcast. That's the way it should be. That's great. >> It's a joint it's a joint effort as you as you well know, Rich. Anyways, this was wonderful. I'm sure people will have enjoyed it as they always do when you are on. And um to those listening, I would only encourage you to go and leave a a raving uh review and rating of Rich's Conversation on your favorite uh podcast platform. It really does help more people to uh discover the show and and uh help us spread the words of the wonderful weird and wonderful world of trend following. Anyways, um we're going to wrap up. So, next week a few changes. Next week, Moritz will be sitting in for me while I visit my daughter in Montreal and he'll be joined by Nick Bolters. So, that's going to be a fun uh conversation. And the following week, Alan will sit in for me and he'll be joined by Yoo. That's also going to be a super fun conversation while I scoot over to our headquarters in Florida uh and visit some of my colleagues uh or all of them uh for that matter. But [snorts] if you do have any questions to any of these uh esteemed trend followers, by all means send them to me as usual, info@ top traders unpluglo.com and I will forward them to um Morates and Allen from Rich and me. Thanks ever so much for listening. We look forward to being back with you next week. And until next time, take care of yourself and take care of each other. >> [music] >> Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the best way to stay updated is to go on over to iTunes and subscribe [music] to the show so that you'll be sure to get all the new episodes as they're released. We have some amazing guests lined up for you. [music] And to ensure our show continues to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the [music] best way to show us you love the podcast. We'll see you next time on Top Traders Unplugged. >> [music]