The Hidden Cracks in Systematic Strategies | Systematic Investor | Ep. 392
Summary
Energy & Commodities: Backwardation and supply shocks in the oil complex and sharp moves in natural gas dominated returns, with Middle East risks and weather-driven spikes stressing commodity carry trades.
Trend Following: Faster-speed trend models outperformed slower ones in March, while diversified multi-strategy portfolios captured oil trends and managed rising volatility via dynamic risk scaling.
Portfolio Construction: Man Group research highlights that a core, liquid universe maximizes crisis alpha, while broader universes improve long-term Sharpe via more idiosyncratic trends and diversification.
Options Microstructure: Zero-DTE options activity can induce intraday mean reversion or trend amplification via dealer hedging, affecting short-term CTAs and offering potential risk management tools.
Macro Outlook: Stagflation risk discussions intensified as rates rose and oil fed into inflation nowcasts, with uncertainty around the persistence of the energy shock.
AI in Research: Agentic LLM workflows can synthesize and iterate on trend systems, but require strict human oversight to avoid overfitting and ensure robust, simple signal design.
Quant Equities: A “quant renaissance” is aided by dynamic factor allocation and alternative data, improving regime resilience versus the prior quant winter.
Diversification Matters: Commodities’ low internal correlations and weighting choices materially affect CTA outcomes, with precious metals, oil, ags, and livestock adding distinct trend sources.
Transcript
Imagine [music] spending an hour with the world's greatest traders. Imagine learning from their experiences, their successes, and their [music] failures. Imagine no more. Welcome to Top Traders Unplugged, [music] the place where you can learn from the best hedge fund managers in the world, 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 and past performance does not guarantee or even infer [music] anything about future performance. Also understand that there's a significant risk of financial loss with all [music] investment strategies and you need to request and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran [music] hedge fund manager Neil's Krup Larson. Welcome or welcome back to another edition of the systematic investor series where Nick Bolters and I, Neils Castro Blaston, where each week we take the pulse of the global market through the lens of a rules-based investor. Let me just say a very warm welcome if today's first time you're joining us. And if someone who cares about you and your portfolio recommended that you tune into the podcast, I would like to also say thank you to the person for sharing the episode with your friends and colleagues. It really does mean a lot to us. Nick, it is um it's wonderful to be back with you uh this week. It's been a little while. That's how it feels. But um how have you been? How's 2026 treating you so far? >> Yeah, it's been it's been some time. Good to see you, Neils. Um 2026 has been quite a quite a year so far, right? I think you know we all started with this kind of booming uh expectation and here we stand debating how much of a of a contagion effect we have ahead or maybe not. Um but other than that, you know, it's been quite busy at work, you know, busy at home, you know, some travels here and there. Um yeah, so good, you know, so good so far, >> I would say. How about you? >> Yeah. No, no, absolutely. I can't complain. Um it's been it's been busy. I feel that the conversations are um highly interesting. Um I feel investors are um um yeah, leaning a little bit more into uh our space. Um, so so far, as you say, so far so good. Um, but you know what? By the way, before we even dive into some of the things we're going to talk about, I couldn't help noticing, maybe I should have put it on my radar uh in in the in the next section, but I couldn't help noticing that there was an article out only a couple of days ago. I think it was a Bloomberg article where they were already saying, you know, CTAs have their worst draw down since Liberation Day. Um, and they were quoting like the trend index being down 4%. And at the time when I looked it said first of all the number was wrong. The index was down like 2% for March at the time. Um, and I was just wondering oh wow is that you know is that all it takes you know a couple of weeks of correction after 8 months of positive returns. That's all it takes for the for these headlines to come out. um maybe not surprised but but but maybe I am actually that that's really what they they uh find interesting. Um but you and I are going to talk about things that we do find interesting at the moment. And um let's jump into um kind of what's been on our radar recently. Um I know you have an interesting uh observation about um the QIS space I think. >> Um yes, what's been in my radar? So I would say I can start professionally right and you know speak about the day you know the the days um in um in in the office and obviously discussing QIS and systematic strategies. So, and obviously I can I can easily give like a personal perspective as well. But I would say obviously something that is um that is top of mind is what is happening in in the Middle East and how that is impacting um performance and how that is impacting uh specific discussions we have with clients and how we respond to the current uh environment in terms of commentary in terms of performance analysis so on and so forth. Um and you know one of the most popular trades uh across the QIS space and maybe it's probably the most long-standing um QIS strategy across the industry is the you know the classic short commodity spread the curve carry >> um so a few days ago I think it was last week uh so Bloomberg came out with an article uh specifically quoting um the backordation dynamics in the oil market obviously as a as a consequence of the supply concerns um with the with the currencical tensions and how that has impacted quite aggressively uh and maybe kind of wrote off a few years of performance uh in a commodity curve trade. You know obviously implementation matters and we can have this conversation at any level of depth but this came pretty much weeks after another draw down in that particular strategy was caused by a completely different event and that was obviously the the cold weather in the states and and and in North Europe um that led gas prices kind of skyrocket temporarily still u but you know if you look into kind of a backtoback events uh of substantial backordation in the gas market and eventually in the oil complex X um you know for a strategy that is effectively going long the back end of the curve and short the front end of the curve that historically has been wellrewarded because that that is the risk premium that is shared between consumers and and and producers. It leads to um some challenging performance that you know we spend time kind of understanding right and you know we spend time um discuss with clients positioning we spend time discussing the the recovery dynamics by all means um it's a trade that historically had very strong min reverting properties because backation unless um the supply shock is permanent uh m reverts very quickly. So normalization of the curves would bring back if you like on a marktomarket basis the lost return. Um so yeah that that is a topic that you know we spend a lot of time you know that is a topic that you know it's on our radar for sure um and it's a consequence of the micro dynamics um that literally came on the back end of us kind of defending um you know the natkas exposure and how that would kind of turn around uh should things normalize on the I guess on the weather side. So yeah, there was this coverage on on Bloomberg, you know, obviously was circulated between the team. Obviously, it has been discussed with clients. Um, and I think it's a topic that, you know, you and I discussed um as an interesting one to kind of bring about. Uh, the article itself, you know, obviously quotes a number of um um a number of players in the space from the kind of a buy side, the sell side. Uh, you know, it quotes obviously how big the UI space has become, which is something that we have discussed. I think I discussed that before. It's back in November. Um it is probably now on the 6 to 700 billion mark. Uh again the convention >> the CTA size it's quite substantial. It's quite substantial but you know this is now a multiasset space delta 1 v retail institutional. So it kind of spans across um you know a number of return profiles. Um, and yeah, commodity curve. I mean, I mentioned that this is probably one of the oldest because it is perhaps one of the reasons why QIS products were built 15 years ago or 20 years ago. You just want to have commodity exposure. You don't want to hold physical commodities. You're going to go to the futures market. You're going to roll systematically. That's a systematic strategy. Then you realize that doing it on the front has negative ro yield. You do it on the back end of the curve. You realize you have some alpha. You sort the front. Here's your curve. talk to me about one thing. I'm definitely not an expert uh in commodity spreads and certainly not in in the details of um some of these uh energy markets that uh are being affected. But I did hear a conversation uh very very recently um that talked about something that I had never really thought about too much and that is uh except I've noticed obviously the price differences and the um and the volatility differences between the gases and the oils. And was quite fascinating about this conversation was really that the person explained it's very different dynamics that drives those two because um whilst the oil transportation cost is really not a big part of the setting of the price but in the gases it plays a huge role uh is my understanding in terms of the pricing of of gas and with this week's um I think bombing of uh perhaps the largest gas facility in the world in in Kata, I think it was. Um, and and and I think that's that's why it prompted this guest to be invited on a podcast to talk about it. Um, but it's quite interesting. Something I've never really thought about that the what makes up the price is is quite different between those two quote unquote energies that we just take for granted that, you know, they're either in the pump or they're, you know, being used for heating or whatever we use it for. >> Yeah. Yeah, it's quite interesting and also from a portfolio construction risk management perspective you know you think of them as like no energy markets but in reality you start looking into even statistical correlations you know there's >> pretty much zero like there's very little correlation between the gases and the oil complex you know the oil complex in itself is much more homogeneous and that obviously has significant implications when you build a portfolio so then you have to start thinking obviously nap gas is the most volatile probably of them all um specifically on the I guess on the on the right tail so it's it spikes quite aggressively Um you know for oil we can have a conversation on both sides of it because you know you can argue that you know there is a supply shock as it currently is the case but also there's a macro shock that can drive the oil price you know further down because of reduced productivity right so it you know can be more kind of birectional with fat tails I think that gas and maybe I'm just kind of front running myself here without having looked at the data recently but I think it's more of a right skewed um but that level of volatility specifically because it's very um very obviously tough to to to foresee um and it's also driven by >> you know by by obviously the you know the uh the winter um more than anything >> um you know it has implications of portfolio design right so you build a trend following strategy how do you think of those markets how do you think of clustering them how do you think of volatility scaling them you look into the big complex that has specific weights that are driven primarily by by um if you like production and liquidity it doesn't account for volatility So then you end up being exposed in risk terms more than what your nominal exposure suggests. So all these are important questions and all these are important considerations when we built a portfolio. Uh and and and I should say that the oil complex and moves that have happened recently maybe bringing that kind of closer to home for us. They were very well captured by trend following which in itself given the rise in the volatility it is reducing risk exposure precisely because yes it might be spiking but it's spiking so aggressively so that the volatility itself suggests look there is a trend but maybe you should be bit more you know conscious of of the amount of risk you deploy. So I guess in the broader context um you know the benefits of diversification and multistrat portfolio when it comes to commodity systematic strategies it keeps on playing out quite well like know the skew or value dynamics now we have discussed here um you know have been benefiting recently because they are benefiting from the petroleum sector. So in a way strategy level details and nuances um are always upon question the benefits of you know the diversification uh they know they have played out quite nicely uh yet to date at least from a from a QIS or systematic standpoint. >> Yeah. >> But you're very right on the on the volatility dynamics. You're very right on this one. >> Sure. Yeah. No, super interesting. I mean I think uh I think Bloomberg will be mentioned quite a few times because on my radar was not something I spent a lot of time looking at. It was just an interesting uh headline and that was that they had an article out um today or yesterday about I mean we we talk a lot or at least we refer a lot to these pot shops and the success they've had and all the assets they've gathered. Um but they actually had an article out uh talking about how traders are now ditching these uh pot shops and and much prefer to set up their own uh go their own way instead. Um and they have um they have you know numbers for 23 24 and 25 and in 23 according to their uh numbers uh or the numbers quoted in the article uh 9% uh of people uh would set up their own would leave and set up their own. In 24 it's 12% and in 25 it was 17%. Uh so um kind of an interesting little take on that and also uh they they I think they have some numbers actually from Goldman Sachs. I can see the sources Goldman Sachs. It must be really really good reliable data. Uh Nick, I'm sure you would agree um that um fewer investors are um are actually looking to invest uh in the space compared to before you know and and a few more expecting to actually decrease the allocation to some of these uh type of strategies. So just a little fun uh observation. The other thing that that hit my um um sort of inbox was just uh obviously the latest FOMC meeting uh and and the comments. It was interesting that Fed Chair Powell uh kind of does not really think we're we're in a stackflationary environment just yet. Um, I think he made the uh distinction that when the term was invented or or or coined back in the 1970s, the difference was really that we also had very high unemployment and and I think for him maybe that plays a role that we're not quite quite in the same scenario. Um, but um at least at least they're talking about the word stackflation even from central banks point of view. Um so um and and you know on that note I mean interest rates have been moving higher uh in the last few days in the last few weeks um and [snorts] um maybe some of the things we've talked about on the podcast not least um people like Jim and and um and and others in terms of perhaps the wishes of the White House in terms of lower interest rate is not what we're going to see. we're going to end up seeing the opposite. Um that that may well, you know, play out that way. >> I mean, I would not disagree. Um some of the discussions we're having at the moment and and I personally had to have with um uh you know, with our colleagues uh in the research as well as with um uh with investors. Um you know, the year as we said 10 minutes ago when we started um was very um I guess bullish in in in the in the sentiment. um and it's now kind of turning around. It's not I don't think it's yet in a place whereby we feel that you know um kind of a down market or an aggressive down market is coming. But you know to your point I think memories of 2022 um are coming back. Um I think the discussion around trend following is here. The discussion you just said it um about volatility uh in the interest rate market you know inflation volatility is back. Um the question as to how central banks will act or react to the current situation is here. Obviously some of that is quite mechanical with the oil moves that feed into I guess now cast of inflation. >> I guess the bigger question certainly I don't have an answer for is um uh is how much of an escalation we see uh and how much of that oil shock um is permanent or about to be subdued. Uh you know historically those events you know took some time before they they die out. eventually they do. Um but I think that's the biggest kind of question mark at the moment. >> But to conclude, I think the the stockflation um maybe the stockflation risk is is is probably elevated visa v kind of two months ago, right? >> You know what that was also and and obviously I'm not trying to make any forecast or political statement or any other kind of statement but just an obser both becoming very neutral here. Um, but at least as a Swiss person, I am generally quite neutral as people would know. But anyways, uh, joking aside, you know, there's one thing I thought about. We we talk about, oh yeah, but they can just, um, you know, the the war can stop any moment and and and and things will go back to normal. But I was thinking about this people who if you think about the Gulf States a lot of them is you know oil is so important but in the last two or three decades what has also become super important to them is tourism and I just wonder I mean are you likely to say to suggest another holiday in in in some of those countries even if the war stops um I don't know so I just think that these after effects in terms of of um you know activity, growth, whatever could be much more um impactful. Uh and that these because we know memories are long, right? If you just if you just been, you know, in an area where there's actually real conflict and bombs are falling and whatever, you're not likely to suggest that as your next destination for your family holiday, I would have thought. So, u it'll be very interesting to see what the real fallout is. uh in my opinion other than just higher oil prices at the moment. Anyways, let's uh jump over to the the trend following uh update that we uh we always do. Uh so I mean so far March uh not surprising has been a month of uh some corrections not nothing too dramatic I think. Um but obviously things like equities where markets have been selling off bonds market has been selling off precious metals quite severely actually. um in some of the markets where maybe it's uh an easy way for people to raise cash and therefore they're selling those uh markets in particular. I don't know. Um but yeah, so we have some corrections. This is of course completely normal especially after a long run of eight consecutive month uh for the industry to deliver positive returns. Uh so this may come to an end this month but it may not. I mean who knows the month is still pretty long. But is there anything that has stood out to you in the first few months uh from from um changes in positioning or um or performance-wise that's that sort of stands out to you? Uh so I think what is very interesting uh from my perspective is and that was more of a March effect rather than necessarily a year to date but I guess you make the extrapolation it can be like a year-to- date realization. um you know what we're discussing back in December and pretty much I would say over the last um couple of years that um kind of slower speeds uh would typically allow to navigate through Vshapes. Uh what we're seeing at the moment is almost like the flip side and perhaps because we're in the middle of the situation. Um but faster speeds at the moment um are substantially outperforming slower speeds. Um like for instance if I look into some of um some of our performance in March it's close to being flat flag frankly. So it's not that it been like an a negative performance and obviously commodities have been the big driver of performance here. Um and to your point it's um um it's the oil complex. So literally from heating oil gas oil brand you know this is really the spectrum of of performance. Uh but you also get some um you also gain some positive return from some of the some of the complex specifically you know some sort of short positions um that have delivered good um good performance month to date. Obviously, equities have been the ones that have been hit. Uh but net net um so far the month is not just within statistical ranges. It's almost like um I guess a non-event frankly from from a trend following standpoint, right? >> Um so I mean to to your earlier point the 4% fine I mean you know we can we've seen 4% up and 4% down. I mean I would not >> I would not get any concerns in this regard. Um so that that is maybe the one point I would kind of flag here that um I've seen a dispersion in the speed complex >> that is almost mirroring but antithetically what we saw last So let me ask you just to put some context on that when you say um it's been the the commodities in a typical QIS portfolio um what would you say the split is between commodities and financial markets or would typically be what you would expect because obviously if you have like we do at Don we have a higher weight to commodities um which we which we like but not everybody likes to have that uh of course so what would you say is your sweet spot in terms terms of um allocation between those two types of markets. >> If I if I ask one clarification, you say you have a higher allocation is that it is just market number of markets in the portfolio because the risk allocation is obviously dynamic, right? So we we obviously don't know uh what that's going to be. I'm just thinking if you put together a portfolio of of say 60 markets or however I I don't even know how many markets would be typical in a QIS. How many of those markets would be commodities uh as a percentage? Uh I I would say from the from implementations with a core universe that can have maybe 10 commodities out of a universe of like 2025 to a much broader you can have up to maybe even 40 markets and commodities. Um but yeah I mean if I were to make a guess and that's just a guess about the industry I would imagine something like a big universe maybe tilted more towards the liquid market so I don't know 15 to 20 is a reasonable number but risk-wise my uh and again that's more my hunch would be that u people allocate a quarter of the risk you know visa v equities and rates and and and currencies um there is an argument to be had that commodities are much more diversifying as an ecosystem them specifically if you have like a broader universe and therefore there is a reason to have maybe even a higher risk budget uh to reflect that that the tagging of that universe being commodities is not as similar as the tagging of equities being equities because in equities the you know the commonality and the homogeneity is significantly higher. So the the clustering we do in commodities is more of a tagging exercise and I know that some of the CPAs would think of commodities as um kind of energy and metals and a one of them being in you know an asset class in isolation. So in other words, instead of having four asset classes, you have seven, but three of them span the universe of commodities, whereas the correlation between those classes remains as low as it is between let's say equities and currencies or you know bonds and a so on so forth. So I think it's a significant philosophical divergence between models that is not heavily typically discussed. I think we speak about you know the speed and the risk management and kind of correlation structure. I don't think we speak too often as to how many or in what context commodities are part of a trend following portfolio and it can have a significant um effect and specifically over the last you know eight months to your point it was it was um a very strong precious metals rally and now oil rally so it's been a big >> and actually I would say probably ever since 2023 when coco started sort of making noises uh in the portfolio and then came coffee and live cattle and all of that stuff. So, um it's been an interesting um period for sure. We may touch on this a little bit because we will review uh a paper that uh talks about sort of the importance of market universe for sure. So, so we'll we'll probably touch on that. So, um yeah, no, so from my perspective, my trend barometer finished at 45 uh last night. So, that's kind of neutral. So, obviously uh not really showing any signs of of stress uh even though performance uh is probably down a little bit um so far this month. Um nothing major in terms of changes as I can tell exposure-wise uh just yet. Um so no real changes of sector exposure perhaps maybe fixed income being uh one where we may see some some changes soon or have seen some markets um you know going from from long to short uh or some more markets for sure. But anyways, performance-wise, as of Tuesday this week, uh yesterday, I should say, was probably a negative day for sure for the industry. I'm pretty sure of that. But before that, uh on Tuesday, the Btop 50 index, um uh sorry, the S SG uh CTA index was down 1.38% for the month, up 6.82% uh so far this year. Uh stockchain trend down 1.77 for the month, up 6.87 for the year. And the short-term traders index as you suggested uh down just just a little bit quarter of a percent um not quite keeping up for the year up 3.55% but still doing better in March. Msei World on the other hand having a tougher time down 4.64% 64% so far this month down now for the year 3 uh 1.34%. The US aggregate bond index also no uh diversification benefit there down 1 a.5% uh in March uh up 17 basis points though for the year and the S&P 500 total return down 3.61% 61% as of last night. Up almost 3%, sorry, down almost 3% [music] I should say so far this year. [music] Before we jump into the uh topics, uh we have uh a couple of questions from Tim that [music] we're going to deal with. But before I do that, I just want to mention to everyone that if they haven't been on the website recently, uh there is a new version of what we call the top traders ultimate guide. Uh and that's essentially a guide to about 600 books now. Um and uh hopefully there will be some of those that will be uh inspirational. Um and uh we put it all together in one uh resource. So you can go to toptraersunplug.com/ultimate and then you can get the guide uh for free. Anyways, um Tim, who has been following the podcast for a long time and is very kind to uh um share and like um our content uh on social media, send a couple of questions in. Um and I think it's quite an interesting question. So he writes, "The options markets have experienced incredible growth in recent years, especially postcoid and the widespread of use of uh zeroday options. uh trend followers most mostly only use futures price data as their inputs to their models. Have we reached a point or will we reach a point in the near future where trend followers should be looking more closely at the option markets and the information within those? And then he says, "As a follow-up question, do you think that by ignoring options data, the performance of existing trend following models could deteriorate or that incorporating options information would produce improvements on existing models? So who better to ask than u than you, Nick? [laughter] >> Okay, I would need gem on this one, but um you know, I'll give it a go. I'll give it a go. I'll give it a go." I mean the observation that um that options markets have become um um to a good extent dominated by the zero DTS is um is not just something that that we did in the news. You know you look into the hard data um you know in anticipation obviously of the discussion I kind of pulled up some numbers. So um currently the volume in S&P options um 60% of it is zero DTS and 40 is anything from daily up to you know longerterm tennos and in that zero DT complex more than half is dital participation. So we can make the argument that it's a place that you know almost like know democratized access to to to kind of intraday leverage um and equally um access to those um I guess cheap bets that you can you know put on the day uh out of the money put or out of the money call depending on the on the view um for a significant premium to be had um or significant payout to be to to be had at a very small premium. uh I guess the question on how this I guess market evolution in the option space can to a certain extent impact and followers. I would look at it in two ways. The first one is how we calculate signals, how we think about establishing market trends that we kind of capture them and then we end up deciding to go long or short. But secondly, I would also look at it into the use that those um options and derivatives can have in risk managing trend following portfolios. So if I go to the first one, right? If I go to the first one, it's it's it's all about the marginal or not impact that dealers would have when hedging their gamma, which is a consequence of investors selling or buying options. So if an option you know if if investors are selling out of the money options you know to generate some carry for instance then dealers would typically be like no long for example and therefore in this activity they would be buying when the market drops and they would be selling when the market goes up. So they have a tendency to min revert to to to force a min reversion behavior. Conversely, if you have the opposite activity, you can be in a negative gamma situation and leverage ETFs is perhaps one reason why this this dynamic can play out uh that you end up buying as the price goes up and sell as the price falls and that is exacerbating trends. So you being a trend follower, you might be in a situation whereby either volatility is suppressed and min reversion is I guess a the artifact of of this auxiliary activity or that prices accelerate at a higher volatility dynamic. So then the question becomes in the absence of those dynamics how would trend in itself perform and you know would be established for us to be able to capture it and I think there is a question here to be had you know should we somehow take that into account not too clear to me how but there is certainly a good question to be had specifically for shortterm managers I think the longerterm managers you know if your signal is very close to zero anyway your risk exposure would be small to be honest with you so does it change the Um if you like the the um the measuring of uh of the signal probably not. Um the other point which perhaps is equally interesting is that you know on specific days that those options specifically uh now not just the zero DS but more broadly um you know have an expiration you have this kind of so-called pin risk. Uh so there's a bit of gravitational power towards the you know the strike price that um that the open interest is large. Um, again, it can have an impact as to how the signals themselves are documented for a trend follower. For a medium-term trend follower, is that impactful? Not very clear to me. Just to be just to be open. Um, with that, you know, at the same time, should be ignored. Certainly not. But frankly, I think more value at least in the present market environment of utilizing those tools to maybe take intraday exposures when the market goes against you, right? So suppose you're holding equities and you know equity starts selling off. Uh so your long position in your long-term model would either take some time to revert or during the day it's going to suffer. Maybe there is a premium to be spent on a zod option to just cover that exposure. And that is more how we can use those more shorter term to risk manage the portfolio without cutting the exposure but rather adding to it at a cost a kind of short-term protection. Um so think know it remains to be seen what the outcome can be but these are some of the some of the quick thoughts. Maybe the last one I would have I think there was a paper recently but you know I hope I don't misrepresent it now. It just came up to my mind actually. Um I think it's by Greg Vilov and some colleagues. Um I think it's a project perhaps sponsored by the CME if I'm not mistaken or CBOE. >> Um doesn't really matter. The point I think they're making they're looking into zero volumes. Um and they find that dealers are typically more long gamma than short gamma. And therefore there is an argument to be made that short-term reversions have become more prevalent than you know than than intraday trends. um which again in itself um can have an impact as to how markets trade intraday. I think we've seen a good amount of reversals uh recently and and over the last few years in daily is that caused by ZD is there is certainly a a a force into it. Um so that's how I look at it. Point number one how the models can become aware if that is required. Number two, can we actually make use of them for risk management? I think the latter is more of a direct use case. a former. >> Okay. >> Maybe. >> Okay. Maybe. >> So, so I would say I tend to agree with you on this one. Um, and but I think about things in in much simpler uh non-quanty terms. Um so in in one in one way I would I just make sort of a simple observation saying yeah the last period of time it's been definitely more challenging for say short-term strategies and maybe that is exactly because there is a force out there that is much more convergent uh and has this sort of mean reverting um interest and that could well be from uh all the people around you know surrounding the the the zero DTA options markets and and how they uh manage risk and so on and so forth. That would be one observation could also be the pot shops. Uh but they're probably also part of maybe the zerodt options market. I have no idea. But the other thing um spotted specifically to Tim's question about you know should we use that data? And I'm thinking what would would we use it for? We're trying to find trends that last for 3 to 6 to 12 to 24 months and a zero DT option volume. And again, not I'm not a quant here, so but I'm thinking, okay, that gives you a sense of people making a bet for a single day. That's not really going to inform me what the price is going to do over the next 3 6 9 12 24 months. And but even for risk management purposes, why would we start trading options when we have plenty of liquidity in the futures markets? And I think one of the trend following I mean one should never say never but one of the trend following mantras has always been you know you you test what you trade and you trade what you test right and and and we are we are using futures data to build all our models and so we should trade futures um and by the way we adjust positions on a daily basis it's not like we're making massive bets any any one day unless something really crazy uh is taking place. So I'm I'm less certain about the use case of option datas for classical trend following models. Um but as I said I'm not a quant so I could be completely wrong here. I mean I can see the hypothesis that uh periods of min reverting behavior can create excess turnover without necessarily being any particular reason for it. I think most of our models do some sort of short-term smoothing anyway. So no maybe we can look at the symptoms of the whip sowing dynamics ultimately if averaging helps a bit moderate that as it has historically done with let's say asynchronicity possibly it's captured indirectly but I can see the hypothesis there I can see the hypothesis of maybe some sort of option implied information like no skew um providing some sort of sentiment indicator that I don't know that in itself might be like a a penalty to the specific direction of the market or maybe multiple um so you know there could be nuances >> anyway full stop full stop there's no comment in my sentence I think I I think I think these are like some of the hypothes forward as a consequence of team's questions but as long to get those questions >> allows us to think about stuff that we may not um think about on a dayby-day basis >> always do all right well let's jump to the papers now we had kind of several choices I think we've narrowed it down to two or three that we liked in particular. Now the first one and by the way courtesy uh mostly by our friends over at man Group um because they are they have been very busy um recently um and they have put out some really interesting one although the first one is a probably a few weeks old um and I think Al and I quickly touched on it that's why I was delighted when I saw you wanted to say a few words about it and even more delighted to say that I think I might get one of the authors authors of the paper on the podcast along with Katie in a couple of weeks. So, we could also maybe leave some questions um for him. Um you never know. So, as I said, I may have touched on it already a little bit. It's a paper uh called a trend following deep dive, the optimal market mix for a trend follower. Now, of course, the topic itself is uh relevant because we talk a lot about it um over the years. what is which market should we trade? Uh do we trade uh too many too few? Uh and and all of that and we all have our different views and and we favor different uh solutions but it's not always we've been very good at uh eloquently describe the benefits of doing you know one choice versus another choice. And I this is what I love about the paper. It's it's a very visual, easy walk through um in terms of what people should be likely to expect from choosing managers trading different market universes. I think they they made that in in a very eloquent way. Um [snorts] of course, if you want to read the paper, you should go to the um man.com and the insights. That's where you will find the paper. Um so um anyways, I very much look forward to hearing your thoughts about it. uh Nick and then we'll we'll we'll take it from there. >> It was a very good read, right? And I think so. Why did that resonate quite well with me? Because we always discuss how defensive your trend follower is, what's a universe, you know, a core market, a broad market, alternative market, whatever market. I think the reason why I like this particular one is because it connects a use case to the universe that you use. So we typically say you want to be defensive, be faster, you want to be longer term performing, be slower, but you know bear in mind you're going to get a lot of beta risk in your portfolio like you know we always have this conversation but I don't think we have ever touched upon the point of you want to be defensive this is the universe to do it with and I think that's why the paper is interesting because it says or maybe taking a step back when we speak about managed futures um there is this kind of duality of objectives I think cliffas wrote about it a couple of years back uh and I keep on using now this kind of duality term. Um it's one of the very few systematic strategies that deliver long-term positive returns and kind of downside protection some sort of a crisis alpha to use Katis term some sort of a reactivity when the markets are falling but not obviously sharply but you know in a kind of a medium term. So this duality of objectives frankly I don't think anything but trend following commodity curve to basically say the the one that we started from. So it's typically defensive as well because kind of shorting in front of commodities. Um so in a noninflationary recession it actually performs well. Uh and then I guess interest rate volatility. There's nothing else at least in my mind that uh doesn't showcase a trade-off between kind of reactivity or defensiveness and and some sort of a kind of cost associated with it. So if we look into this duality which is performance and defensiveness, we can stretch it out and say, well, how can I maximize my defensiveness and how can I maximize my kind of long-term sharp ratio? And surely there's going to be like a trade-off now between the, you know, I guess the choice I have available while still obviously maintaining the duality. So it's not about making, I don't know, negative long-term return or making non-reactive kind of a profile. So I think there is a commonality here in the solution that being that the duality is preserved but it's more about kind of the major and the minor in the objective and they make the point that look if you really want to be defensive and defensive is more about generating positive return when the market goes through a a stress situation then you should have a market universe that is more kind of core and more kind of standard and more mainstream and perhaps more liquid >> because no surprise is when you have a massive correction, it is not that assets fall, it is principal components that are falling. So equities as a risk materializes um perhaps I don't know in the duration space you have central bank intervention and and you just want to be long bonds or maybe some of the cyclical commodities except maybe precious metals are suffering from a drop. So ultimately in a period whereby asset prices get squeezed and maybe correlations um are shifting to the extremes you just want to trade the principal component. So if you stick to a core market universe even for the same specification of speed and correlation so on and so forth you'll end up maximizing your kind of crisis alpha or defensiveness. Um and I I think Andre would actually be quite um uh quite opinionated on this one because I think that's part of his uh of his pitch, right? That you know the universe actually quite quite quite short but very much representative of the macro moves. Conversely, all the alternative markets you know maybe the alternative commodities or you know we can go into kind of EM equities or even equity factors that they also utilize in the paper. You know this expansion of the universe by design brings idiosyncratic risk that with the assumption of trendiness would provide longerterm diversification and therefore return and therefore maximum long-term sharp ratio is achieved with a broader universe with more independent bets. Now this is not as defensive as the tide universe would be. It is still defensive to my earlier point but longerterm has a better sharp. So now this poses the question you know you [clears throat] being asked allocator what do you want to solve for if you want to solve for defensiveness how should you look into it is it like a speed discussion is it a universe discussion is the risk management discussion but there has to be a discussion if instead you're looking into it more as an alpha seeking overlay then maybe a broader universe can be more associated with your objective and and and you know you can meet that objective more more successfully so with the recognition that you might not have the best defensiveness you know in in in D. So I think I'm posing here for for you to reflect but this is how I think about Well, I mean I just wanted to ask your opinion as well. Um, and that is would you agree? It's a leading question you can hear. Would you agree that actually that's exactly what I think this discussion brings? Because previously I've always felt that the crisis element protection element was really always a discussion that related back to speed. >> Yeah, >> 100%. And I plead guilty that you know I was all always going back to the speed discussion or maybe the you know the amount of controlling of your equity risk and you can do it with data or maybe some force constraint exposures all that >> somehow it I don't think it was very natural to think of the markets as directly at this as this one says I mean yeah we could control equity risk maybe we kind of thought about it but not to that extent um so I think it's interesting to to your point I think it kind of shifts a bit the discussion >> complete keep keep going. >> Um, look, no, I mean that that's pretty much it. I think they have another um, you know, another interesting kind of um, uh, I guess path that they follow which is um, >> you know, in addition to the maximum sharp price and the and the maximum crisis alpha portfolio. >> Um, looking into also cash efficiency and obviously the fact that we trade futures, you know, futures are, you know, instruments that trade on margin. So you don't have to spend, you know, $100 for a $100 nostal exposure. you just pay a fraction of it which is determined by the exchange. So the question then becomes obviously how is that margin determined and you know that margin should be financed from actual dollars. So the lower that margin is the more capital efficient the portfolio can be. So they make you know a a supposition here that a portfolio that is more cash efficient and therefore you can get higher level of volatility for the same dollar value of national uh sorry of of margin um is a portfolio that most likely would have liquid markets because that's no that's a component of um of of of the margin determination. Now that in itself brings you to a universe that is more of a core universe. So it's not a surprise that the cast efficient portfolio from that perspective resembles the maximum crisis alpha portfolio. Why? because them two for a different reasons want to capture the micro moves, want to capture the more liquid assets ends up, you know, allocating to the same ecosystem of of um if you like of assets. So like anyway, for me that's more of a byproduct of the discussion, you know, very nice to see it. Uh it almost says if I were to kind of reverse the argument, if you were to have the maximum crisis shar portfolio, it is more likely that your margin requirement would be lower. But yeah, to me to me the biggest uh the biggest point is is um is really the the association of the objective to to the universe. That's that that that's really >> another stat from the paper that I thought was kind of interesting uh was they they managed to to say that there are approximately 900 different markets we could trade. That that that surprised me that they could find so many. Um >> this is very true. I mean the the one the one thing I would point out you remind me I had forgotten about this one. If you look into equity styles, so they use equity factors and and they have 60 mark 60 markets or 60 factors. I mean something I'm sure they would they would attest to it and I guess you can ask one of the co-authors when when when he's here. There are no 60 factors that can be seen as to a good extent diversifying in the equity space. Um you know we can talk about earnings to price and book to price as valuation ratios. We can think of those as two factors but technically the amount of cross-sectional correlation is very high and they are value descriptors. So in a way 60 is almost like at least that's my understanding maybe I'm wrong but 60 here is almost like a signal by signal characterization of how many degrees we have to rank single stocks by and build single stock equity factors. But that is slightly at at least in my mind dissimilar to saying I'm gonna have 60 commodities >> because I think in the equities world maybe five maybe six factors >> probably explain the cross-section of returns. So I think 60 commodities versus 60 descriptors of equity returns would give you much less of a diversified universe in the equity space rather than in the in in the commodity space. So uh but you know setting that aside 900 is 900. So outside of that, I cannot really pose significant um questioning on on on on >> that's fine. We we'll look forward to having uh >> but yeah >> having one of the authors on the show in a couple of weeks and we'll we'll probably uh touch on this uh again no doubt but we're going to stay with man group as I said they've been really busy writing and you identified another paper um which is a little bit different I think uh in terms of the topic it's called um alpha trend and agentic research workflow so not necessarily specifically about building models but maybe processes um that can help do it easier or perceived to help to do it easier. Um what what were your takeaways from from the uh the the questions they raised? So I picked this up to discuss not so much because it talks about trend following obviously it came to my I guess to my inbox um as um as the output of of of the research at man. Um I think to me this poses like a bigger question as to how we see those new technologies and AI and Gen AI and LMS uh becoming not just tools that we use uh but eventually become core components of the research process. Um and and you know for the sake of um I guess of of those that are listening the paper talks about kind of designing an agent that can build a trend system. So to a certain extent no running a trend following strategy or any systematic strategy is a very deterministic process. You have your data you clean your data you build a signal you estimate risk you throw that into an optimizer maybe a ranking mechan methodology whatever that is you get target weights you have some liquidity control you have some volatility target and here you are the quantities to trade on a daily basis or whatever basis. So that process is very deterministic and certainly you can think of a world whereby single agents do that job for you and there's a kind of a supervisor that allows you to command this whole ecosystem. Um and they make the point that you know the current um kind of chat chat bots that we have are probably kind of more shallow in the amount of depth but very broad in the topics we can discuss. You know this model of conducting research is much more targeted. So very narrow in terms of scope but very deep in terms of um uh in terms of analysis. Um so in that context you know again to to to their paper you know they kind of give it out they give it like um a breakout signal um and they deliberately remove some good feature out of it and then they obviously ask the model to find what this feature could be u by describing it quite um uh I guess describing it with words and yes the model does pick it up and does outperform but you know then there's another feature that you know should not be um value accredititive and then the model indeed finds that it's not as useful as you'd expect it to be. Um and and broadly speaking they make the point that you know the model can actually go and operate as as a human being good but equally I think that's even more important that you know human judgment remains extremely extremely important not just from how you frame the questions and how you guide the process and how you interpret the results but also kind of guarding safeguarding against multiple testing overfitting like things that we've been discussing for years and how the culture of a research team is more important than the research team itself. Um I think they make the same point but now the research team happens to be like um kind of an agentic model rather than an individual or like a team. Um so I'm kind of bringing that up maybe the last point. Um they they kind of use different um different models like you know cloud and and GPT and so on so forth uh with the same prompts and and they found that the outcome was actually quite different uh but not necessarily worse or better. Uh so one for example was more of a kind of a single direction-minded you know with with with very correlated outcomes. The other one was a bit more dispersed in the way that it kind of treated the data. But broadly speaking the one thing that at least remains at least my personal view is that we still require this kind of critical thinking. Um I think the synthesis that those models can do is is insane. like I'm personally surprised every single day uh you know by using the tools but I think having some sort of a disciplined evaluation of the outcome and having critical thinking of the outcome is extremely extremely important. I'll give you this example unrelated to the paper. Um you know I I went to one of those engines um recently and on purpose I asked the following question. How has the first two months of the year been for trend following of all things? because I know the answer and I know that you know January and February were probably you probably one of the best two-month periods to start a year for trend followers right and you know the first sentence that comes out and I don't even care about what followed was like the start of the year has been mixed and I'm like you're so wrong [laughter] um so to to to that point I think some level of supervision is is more than important here it's actually essential specifically when you end up kind of managing money on behalf of of policy holders and and pensioners and and and is insurers. Um >> and on top of that actually I think a lot of what um >> companies >> I think that the value of what uh our esteemed researchers do nowadays is actually trying to keeps things simpler. Uh meaning uh I have a feeling without being an AI expert that they love to expand on things. That's kind of what they do. They find sentences and words and all of that stuff. But actually when you build a trend following system, yeah, you have a lot of options, but um but actually the skill and and the experience um goes towards actually stripping things down so you get the the cleanest uh signal, less noise. Um and uh and I'm not so sure that AI is really, you know, the DNA of AI may not actually be very compatible with that. >> We shall see. I I think the um the amount of um evolution we've seen in the last few months is is extraordinary. >> If I were to speak about like my personal experience and how I use it, it's just um for the same task. Uh I I I used to get rubbish. >> I'm getting high quality outcome now. >> Um so it's it's it's quite impressive. We shall see where time where time goes. But um I I found very interesting that you know they actually ended up kind of writing actually the second report on how they use AI for research purposes and obviously kudos to them. >> Um I guess opening up and and making that a topic of discussion rather than saying oh we use this model here's the line you know go trade with it. Let's make it three for three. Uh Nick, because you uh identified a third man uh paper um the quant renaissance >> exactly three for three. Talk to us about uh this paper and why it caught your your attention. >> No, this is now shifting gears away from trend following. So this is about uh contequity. So that's from the numeric crowd at uh at man. Um so I'm spending a good amount of my time to you know with uh with our equities. um kind of single equities uh strategies. It's a very interesting space because it's maybe one of the places that systematic invested started from. Obviously, trend following is probably the longest living from the 70s, but if I were to pick what the second one is, probably equity factors is the second one to come. And it could have equally been a an 80s or a 90s gig. um you know when Pam and friends came about um kind of the whole factor uh and then obviously uh who was a civ rose with a so this space of of equity factors driving returns and therefore being rewarded for a specific risk exposure um could have been a an investment mantra back in the '90s but it only became much more popularized when you know when we had you know high performance computers to turn all this data and the cross-section of stocks and you know did corporate actions and so on and so forth. Um so that space obviously became very very popular and then coming 2018 to 2020 it suffered from this kind of quant winter uh that had significant consequences also for the for the for the buy side managing those u those strategies. Um you know lo and behold postcoid uh significant resurrection uh so the multistrand the quant equity you know the long short space um has had very strong performance um and and what this report tries to bring into I guess into the discussion is how much of a risk we have of a repeat of a winter but also maybe how more resilient the space from a research product uh and and and and uh investment management has evolved over the years. Um so why did we have the winter back in the days? Frankly, there are two reasons. Um they bring about I would possibly just agree without having seen them. One was that the macro regime was not very accommodating for some of the factors obviously with the interest rate markets operating as they did back in the late late part of 2010s. You know, value was certainly impacted quite significantly. Then you had some reversions and the momentum did not play out well and the overall complex you know was just not accommodative of the micro environment. The second one is probably possibly some factor crowding. So the the theme became too quickly too popular to having not just perhaps alpha decay consequences but also having forced unwinding um events when some sort of a macro shock or funding liquidity driving uh driving some of the um driving some of the performance and and and then you have this circus of of of unwinding and obviously one is causing the other and so on and so forth. So some of the far cell dynamic and and you know if I were to quote some of the some of the stats in the paper they say that during stress periods you know classical macro factors can explain more than 50% of factor return variation which is quite substantial if you were to think that you know normal times is more like know 20%. So that was what happened back then and now the question is obviously if we bring this world to today how things have changed. They use an internal model they have for regime identification. So no surprise you typically have kind of recession early expansion late expansion and overheating. So the four typical regimes that the equity market or broadly the macro market goes through they utilize that to understand better the quan winter uh and associate if you like different premier different different environments. But then they make the argument that with us understanding better how the micro cycle works at least through the lens of factor investing. Their argument is that now we as investment practitioners you know as managers um are utilizing more dynamic allocation in the factor space and less static and I think that's something we also see and you know we have seen the QIS space doing and I think um it is now much more of an expectation that you have a dynamic allocation mindset rather than a static one. Um so there is some value that uh dynamic allocation can bring with I guess the caveat of uh concentration. Um there is obviously new data. So what historically used to be your quality score and your momentum score and your value score can now be either enhanced or expanded in the factor space by kind of geocation data and credit card data and you know you name it and patents and uh kind of sentiment from the buy side from the sell side you know using ML models. So there are new data to be utilized uh which inherently provides some diversification. So they put some quotes that for example you know alpha models back in the days um you know would have um correlations of like know u 70 to 80% these days more like know 40 to 50. Uh so there is some diversification at the signal level. Um and lastly um they make the argument that you know going through the more dynamic allocation going through new data going through more modern techniques uh designing portfolios or or crafting alpha scores eventually provides some sort of a macro regime resilience. So what was back in the day a macro regime dependence can become a macro regime resilience with the use of all those technologies and data sources in the quantity space. I think you know equities because of their dimensionality uh they have been the most researcher and they will continue being a space that return can be harvested. Um so I found this one quite um quite an interesting one both for myself as well as I guess for for for this discussion to be to be brought forward. So that that that's the whole story right is the quant winter something we can see again and if that is the case you know what drove it back then how the ch how the how the changes that we've seen in the space have maybe reduces probability um if the micro regime is not very accommodative so that's >> in a sense we could link that to trend following right the people talk about a trend following winter a few years ago as well right I mean have we >> I thought I thought you said in early March >> no I was thinking about the 2010s where people were complaining. I know. I know. And and where people were sort of um disappointed with returns. But when you think about the macroeconomic environment, low inflation, stable inflation, uh very little movement on GDP and so on and so forth, of course, um that's not necessarily the best environment. The question is, of course, with all the research we do, should that happen again, would we cope much better with it? That's kind of the same um argument or question that uh that only time will tell I guess. I mean I think I think we've discussed that maybe that was the that was the reason why we first met in the very very first place right you know this whole discussion I remember back in the days having about fundamental certainty and uncertainty and how the Fed put brings a lot of fundamental certainty I would not claim that currently we're sitting in a fundamental certain kind of environment >> doesn't look like it so >> doesn't look like >> certainty brings reversions uncertainty allows price trends to continue because a price trend in a certain environment provides information about where the fundamental value should sit whereas certainty brings you back to your prior which is very informative of valuation. So that I think this is the dynamic at least in my mind that the fedwood was was bringing to the trend space and therefore the challenges that came about. No, absolutely. This was great. Uh Nick, thank you so much for spending time preparing for all of this, finding these papers um and discussing them, of course. And uh for all of the for all of those of you listening um you know feel free and and please uh do uh head over to your favorite podcast platform um or YouTube and leave a uh rating and review to support uh the channel but also as a thank you to Nick and all your other co-hosts who do a tremendous job every week in uh preparing for these conversations. Um, before I wrap up, uh, let me just say that, uh, if you have questions for next week, which is where I will be joined by Yav, um, then send them to info toptraders.com, uh, just like Tim did, uh, this week, and I'll do my best to bring it up in our conversation. Um, with that said, from Nick and me, thank you ever so much for listening. We look forward to being back with you next week. And until next time, as usual, take care of yourself and take care [music] of each other. Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the [music] best way to stay updated is to go on over to iTunes and subscribe to the show so that you'll be sure to get all the new episodes [music] as they're released. We have some amazing guests lined up for you. And to ensure our show continues [music] to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the best way to [music] show us you love the podcast. We'll see you next time on Top Traders Unplugged. >> [music]
The Hidden Cracks in Systematic Strategies | Systematic Investor | Ep. 392
Summary
Transcript
Imagine [music] spending an hour with the world's greatest traders. Imagine learning from their experiences, their successes, and their [music] failures. Imagine no more. Welcome to Top Traders Unplugged, [music] the place where you can learn from the best hedge fund managers in the world, 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 and past performance does not guarantee or even infer [music] anything about future performance. Also understand that there's a significant risk of financial loss with all [music] investment strategies and you need to request and understand the specific risks from the investment manager about their product before you make investment decisions. Here's your host, veteran [music] hedge fund manager Neil's Krup Larson. Welcome or welcome back to another edition of the systematic investor series where Nick Bolters and I, Neils Castro Blaston, where each week we take the pulse of the global market through the lens of a rules-based investor. Let me just say a very warm welcome if today's first time you're joining us. And if someone who cares about you and your portfolio recommended that you tune into the podcast, I would like to also say thank you to the person for sharing the episode with your friends and colleagues. It really does mean a lot to us. Nick, it is um it's wonderful to be back with you uh this week. It's been a little while. That's how it feels. But um how have you been? How's 2026 treating you so far? >> Yeah, it's been it's been some time. Good to see you, Neils. Um 2026 has been quite a quite a year so far, right? I think you know we all started with this kind of booming uh expectation and here we stand debating how much of a of a contagion effect we have ahead or maybe not. Um but other than that, you know, it's been quite busy at work, you know, busy at home, you know, some travels here and there. Um yeah, so good, you know, so good so far, >> I would say. How about you? >> Yeah. No, no, absolutely. I can't complain. Um it's been it's been busy. I feel that the conversations are um highly interesting. Um I feel investors are um um yeah, leaning a little bit more into uh our space. Um, so so far, as you say, so far so good. Um, but you know what? By the way, before we even dive into some of the things we're going to talk about, I couldn't help noticing, maybe I should have put it on my radar uh in in the in the next section, but I couldn't help noticing that there was an article out only a couple of days ago. I think it was a Bloomberg article where they were already saying, you know, CTAs have their worst draw down since Liberation Day. Um, and they were quoting like the trend index being down 4%. And at the time when I looked it said first of all the number was wrong. The index was down like 2% for March at the time. Um, and I was just wondering oh wow is that you know is that all it takes you know a couple of weeks of correction after 8 months of positive returns. That's all it takes for the for these headlines to come out. um maybe not surprised but but but maybe I am actually that that's really what they they uh find interesting. Um but you and I are going to talk about things that we do find interesting at the moment. And um let's jump into um kind of what's been on our radar recently. Um I know you have an interesting uh observation about um the QIS space I think. >> Um yes, what's been in my radar? So I would say I can start professionally right and you know speak about the day you know the the days um in um in in the office and obviously discussing QIS and systematic strategies. So, and obviously I can I can easily give like a personal perspective as well. But I would say obviously something that is um that is top of mind is what is happening in in the Middle East and how that is impacting um performance and how that is impacting uh specific discussions we have with clients and how we respond to the current uh environment in terms of commentary in terms of performance analysis so on and so forth. Um and you know one of the most popular trades uh across the QIS space and maybe it's probably the most long-standing um QIS strategy across the industry is the you know the classic short commodity spread the curve carry >> um so a few days ago I think it was last week uh so Bloomberg came out with an article uh specifically quoting um the backordation dynamics in the oil market obviously as a as a consequence of the supply concerns um with the with the currencical tensions and how that has impacted quite aggressively uh and maybe kind of wrote off a few years of performance uh in a commodity curve trade. You know obviously implementation matters and we can have this conversation at any level of depth but this came pretty much weeks after another draw down in that particular strategy was caused by a completely different event and that was obviously the the cold weather in the states and and and in North Europe um that led gas prices kind of skyrocket temporarily still u but you know if you look into kind of a backtoback events uh of substantial backordation in the gas market and eventually in the oil complex X um you know for a strategy that is effectively going long the back end of the curve and short the front end of the curve that historically has been wellrewarded because that that is the risk premium that is shared between consumers and and and producers. It leads to um some challenging performance that you know we spend time kind of understanding right and you know we spend time um discuss with clients positioning we spend time discussing the the recovery dynamics by all means um it's a trade that historically had very strong min reverting properties because backation unless um the supply shock is permanent uh m reverts very quickly. So normalization of the curves would bring back if you like on a marktomarket basis the lost return. Um so yeah that that is a topic that you know we spend a lot of time you know that is a topic that you know it's on our radar for sure um and it's a consequence of the micro dynamics um that literally came on the back end of us kind of defending um you know the natkas exposure and how that would kind of turn around uh should things normalize on the I guess on the weather side. So yeah, there was this coverage on on Bloomberg, you know, obviously was circulated between the team. Obviously, it has been discussed with clients. Um, and I think it's a topic that, you know, you and I discussed um as an interesting one to kind of bring about. Uh, the article itself, you know, obviously quotes a number of um um a number of players in the space from the kind of a buy side, the sell side. Uh, you know, it quotes obviously how big the UI space has become, which is something that we have discussed. I think I discussed that before. It's back in November. Um it is probably now on the 6 to 700 billion mark. Uh again the convention >> the CTA size it's quite substantial. It's quite substantial but you know this is now a multiasset space delta 1 v retail institutional. So it kind of spans across um you know a number of return profiles. Um, and yeah, commodity curve. I mean, I mentioned that this is probably one of the oldest because it is perhaps one of the reasons why QIS products were built 15 years ago or 20 years ago. You just want to have commodity exposure. You don't want to hold physical commodities. You're going to go to the futures market. You're going to roll systematically. That's a systematic strategy. Then you realize that doing it on the front has negative ro yield. You do it on the back end of the curve. You realize you have some alpha. You sort the front. Here's your curve. talk to me about one thing. I'm definitely not an expert uh in commodity spreads and certainly not in in the details of um some of these uh energy markets that uh are being affected. But I did hear a conversation uh very very recently um that talked about something that I had never really thought about too much and that is uh except I've noticed obviously the price differences and the um and the volatility differences between the gases and the oils. And was quite fascinating about this conversation was really that the person explained it's very different dynamics that drives those two because um whilst the oil transportation cost is really not a big part of the setting of the price but in the gases it plays a huge role uh is my understanding in terms of the pricing of of gas and with this week's um I think bombing of uh perhaps the largest gas facility in the world in in Kata, I think it was. Um, and and and I think that's that's why it prompted this guest to be invited on a podcast to talk about it. Um, but it's quite interesting. Something I've never really thought about that the what makes up the price is is quite different between those two quote unquote energies that we just take for granted that, you know, they're either in the pump or they're, you know, being used for heating or whatever we use it for. >> Yeah. Yeah, it's quite interesting and also from a portfolio construction risk management perspective you know you think of them as like no energy markets but in reality you start looking into even statistical correlations you know there's >> pretty much zero like there's very little correlation between the gases and the oil complex you know the oil complex in itself is much more homogeneous and that obviously has significant implications when you build a portfolio so then you have to start thinking obviously nap gas is the most volatile probably of them all um specifically on the I guess on the on the right tail so it's it spikes quite aggressively Um you know for oil we can have a conversation on both sides of it because you know you can argue that you know there is a supply shock as it currently is the case but also there's a macro shock that can drive the oil price you know further down because of reduced productivity right so it you know can be more kind of birectional with fat tails I think that gas and maybe I'm just kind of front running myself here without having looked at the data recently but I think it's more of a right skewed um but that level of volatility specifically because it's very um very obviously tough to to to foresee um and it's also driven by >> you know by by obviously the you know the uh the winter um more than anything >> um you know it has implications of portfolio design right so you build a trend following strategy how do you think of those markets how do you think of clustering them how do you think of volatility scaling them you look into the big complex that has specific weights that are driven primarily by by um if you like production and liquidity it doesn't account for volatility So then you end up being exposed in risk terms more than what your nominal exposure suggests. So all these are important questions and all these are important considerations when we built a portfolio. Uh and and and I should say that the oil complex and moves that have happened recently maybe bringing that kind of closer to home for us. They were very well captured by trend following which in itself given the rise in the volatility it is reducing risk exposure precisely because yes it might be spiking but it's spiking so aggressively so that the volatility itself suggests look there is a trend but maybe you should be bit more you know conscious of of the amount of risk you deploy. So I guess in the broader context um you know the benefits of diversification and multistrat portfolio when it comes to commodity systematic strategies it keeps on playing out quite well like know the skew or value dynamics now we have discussed here um you know have been benefiting recently because they are benefiting from the petroleum sector. So in a way strategy level details and nuances um are always upon question the benefits of you know the diversification uh they know they have played out quite nicely uh yet to date at least from a from a QIS or systematic standpoint. >> Yeah. >> But you're very right on the on the volatility dynamics. You're very right on this one. >> Sure. Yeah. No, super interesting. I mean I think uh I think Bloomberg will be mentioned quite a few times because on my radar was not something I spent a lot of time looking at. It was just an interesting uh headline and that was that they had an article out um today or yesterday about I mean we we talk a lot or at least we refer a lot to these pot shops and the success they've had and all the assets they've gathered. Um but they actually had an article out uh talking about how traders are now ditching these uh pot shops and and much prefer to set up their own uh go their own way instead. Um and they have um they have you know numbers for 23 24 and 25 and in 23 according to their uh numbers uh or the numbers quoted in the article uh 9% uh of people uh would set up their own would leave and set up their own. In 24 it's 12% and in 25 it was 17%. Uh so um kind of an interesting little take on that and also uh they they I think they have some numbers actually from Goldman Sachs. I can see the sources Goldman Sachs. It must be really really good reliable data. Uh Nick, I'm sure you would agree um that um fewer investors are um are actually looking to invest uh in the space compared to before you know and and a few more expecting to actually decrease the allocation to some of these uh type of strategies. So just a little fun uh observation. The other thing that that hit my um um sort of inbox was just uh obviously the latest FOMC meeting uh and and the comments. It was interesting that Fed Chair Powell uh kind of does not really think we're we're in a stackflationary environment just yet. Um, I think he made the uh distinction that when the term was invented or or or coined back in the 1970s, the difference was really that we also had very high unemployment and and I think for him maybe that plays a role that we're not quite quite in the same scenario. Um, but um at least at least they're talking about the word stackflation even from central banks point of view. Um so um and and you know on that note I mean interest rates have been moving higher uh in the last few days in the last few weeks um and [snorts] um maybe some of the things we've talked about on the podcast not least um people like Jim and and um and and others in terms of perhaps the wishes of the White House in terms of lower interest rate is not what we're going to see. we're going to end up seeing the opposite. Um that that may well, you know, play out that way. >> I mean, I would not disagree. Um some of the discussions we're having at the moment and and I personally had to have with um uh you know, with our colleagues uh in the research as well as with um uh with investors. Um you know, the year as we said 10 minutes ago when we started um was very um I guess bullish in in in the in the sentiment. um and it's now kind of turning around. It's not I don't think it's yet in a place whereby we feel that you know um kind of a down market or an aggressive down market is coming. But you know to your point I think memories of 2022 um are coming back. Um I think the discussion around trend following is here. The discussion you just said it um about volatility uh in the interest rate market you know inflation volatility is back. Um the question as to how central banks will act or react to the current situation is here. Obviously some of that is quite mechanical with the oil moves that feed into I guess now cast of inflation. >> I guess the bigger question certainly I don't have an answer for is um uh is how much of an escalation we see uh and how much of that oil shock um is permanent or about to be subdued. Uh you know historically those events you know took some time before they they die out. eventually they do. Um but I think that's the biggest kind of question mark at the moment. >> But to conclude, I think the the stockflation um maybe the stockflation risk is is is probably elevated visa v kind of two months ago, right? >> You know what that was also and and obviously I'm not trying to make any forecast or political statement or any other kind of statement but just an obser both becoming very neutral here. Um, but at least as a Swiss person, I am generally quite neutral as people would know. But anyways, uh, joking aside, you know, there's one thing I thought about. We we talk about, oh yeah, but they can just, um, you know, the the war can stop any moment and and and and things will go back to normal. But I was thinking about this people who if you think about the Gulf States a lot of them is you know oil is so important but in the last two or three decades what has also become super important to them is tourism and I just wonder I mean are you likely to say to suggest another holiday in in in some of those countries even if the war stops um I don't know so I just think that these after effects in terms of of um you know activity, growth, whatever could be much more um impactful. Uh and that these because we know memories are long, right? If you just if you just been, you know, in an area where there's actually real conflict and bombs are falling and whatever, you're not likely to suggest that as your next destination for your family holiday, I would have thought. So, u it'll be very interesting to see what the real fallout is. uh in my opinion other than just higher oil prices at the moment. Anyways, let's uh jump over to the the trend following uh update that we uh we always do. Uh so I mean so far March uh not surprising has been a month of uh some corrections not nothing too dramatic I think. Um but obviously things like equities where markets have been selling off bonds market has been selling off precious metals quite severely actually. um in some of the markets where maybe it's uh an easy way for people to raise cash and therefore they're selling those uh markets in particular. I don't know. Um but yeah, so we have some corrections. This is of course completely normal especially after a long run of eight consecutive month uh for the industry to deliver positive returns. Uh so this may come to an end this month but it may not. I mean who knows the month is still pretty long. But is there anything that has stood out to you in the first few months uh from from um changes in positioning or um or performance-wise that's that sort of stands out to you? Uh so I think what is very interesting uh from my perspective is and that was more of a March effect rather than necessarily a year to date but I guess you make the extrapolation it can be like a year-to- date realization. um you know what we're discussing back in December and pretty much I would say over the last um couple of years that um kind of slower speeds uh would typically allow to navigate through Vshapes. Uh what we're seeing at the moment is almost like the flip side and perhaps because we're in the middle of the situation. Um but faster speeds at the moment um are substantially outperforming slower speeds. Um like for instance if I look into some of um some of our performance in March it's close to being flat flag frankly. So it's not that it been like an a negative performance and obviously commodities have been the big driver of performance here. Um and to your point it's um um it's the oil complex. So literally from heating oil gas oil brand you know this is really the spectrum of of performance. Uh but you also get some um you also gain some positive return from some of the some of the complex specifically you know some sort of short positions um that have delivered good um good performance month to date. Obviously, equities have been the ones that have been hit. Uh but net net um so far the month is not just within statistical ranges. It's almost like um I guess a non-event frankly from from a trend following standpoint, right? >> Um so I mean to to your earlier point the 4% fine I mean you know we can we've seen 4% up and 4% down. I mean I would not >> I would not get any concerns in this regard. Um so that that is maybe the one point I would kind of flag here that um I've seen a dispersion in the speed complex >> that is almost mirroring but antithetically what we saw last So let me ask you just to put some context on that when you say um it's been the the commodities in a typical QIS portfolio um what would you say the split is between commodities and financial markets or would typically be what you would expect because obviously if you have like we do at Don we have a higher weight to commodities um which we which we like but not everybody likes to have that uh of course so what would you say is your sweet spot in terms terms of um allocation between those two types of markets. >> If I if I ask one clarification, you say you have a higher allocation is that it is just market number of markets in the portfolio because the risk allocation is obviously dynamic, right? So we we obviously don't know uh what that's going to be. I'm just thinking if you put together a portfolio of of say 60 markets or however I I don't even know how many markets would be typical in a QIS. How many of those markets would be commodities uh as a percentage? Uh I I would say from the from implementations with a core universe that can have maybe 10 commodities out of a universe of like 2025 to a much broader you can have up to maybe even 40 markets and commodities. Um but yeah I mean if I were to make a guess and that's just a guess about the industry I would imagine something like a big universe maybe tilted more towards the liquid market so I don't know 15 to 20 is a reasonable number but risk-wise my uh and again that's more my hunch would be that u people allocate a quarter of the risk you know visa v equities and rates and and and currencies um there is an argument to be had that commodities are much more diversifying as an ecosystem them specifically if you have like a broader universe and therefore there is a reason to have maybe even a higher risk budget uh to reflect that that the tagging of that universe being commodities is not as similar as the tagging of equities being equities because in equities the you know the commonality and the homogeneity is significantly higher. So the the clustering we do in commodities is more of a tagging exercise and I know that some of the CPAs would think of commodities as um kind of energy and metals and a one of them being in you know an asset class in isolation. So in other words, instead of having four asset classes, you have seven, but three of them span the universe of commodities, whereas the correlation between those classes remains as low as it is between let's say equities and currencies or you know bonds and a so on so forth. So I think it's a significant philosophical divergence between models that is not heavily typically discussed. I think we speak about you know the speed and the risk management and kind of correlation structure. I don't think we speak too often as to how many or in what context commodities are part of a trend following portfolio and it can have a significant um effect and specifically over the last you know eight months to your point it was it was um a very strong precious metals rally and now oil rally so it's been a big >> and actually I would say probably ever since 2023 when coco started sort of making noises uh in the portfolio and then came coffee and live cattle and all of that stuff. So, um it's been an interesting um period for sure. We may touch on this a little bit because we will review uh a paper that uh talks about sort of the importance of market universe for sure. So, so we'll we'll probably touch on that. So, um yeah, no, so from my perspective, my trend barometer finished at 45 uh last night. So, that's kind of neutral. So, obviously uh not really showing any signs of of stress uh even though performance uh is probably down a little bit um so far this month. Um nothing major in terms of changes as I can tell exposure-wise uh just yet. Um so no real changes of sector exposure perhaps maybe fixed income being uh one where we may see some some changes soon or have seen some markets um you know going from from long to short uh or some more markets for sure. But anyways, performance-wise, as of Tuesday this week, uh yesterday, I should say, was probably a negative day for sure for the industry. I'm pretty sure of that. But before that, uh on Tuesday, the Btop 50 index, um uh sorry, the S SG uh CTA index was down 1.38% for the month, up 6.82% uh so far this year. Uh stockchain trend down 1.77 for the month, up 6.87 for the year. And the short-term traders index as you suggested uh down just just a little bit quarter of a percent um not quite keeping up for the year up 3.55% but still doing better in March. Msei World on the other hand having a tougher time down 4.64% 64% so far this month down now for the year 3 uh 1.34%. The US aggregate bond index also no uh diversification benefit there down 1 a.5% uh in March uh up 17 basis points though for the year and the S&P 500 total return down 3.61% 61% as of last night. Up almost 3%, sorry, down almost 3% [music] I should say so far this year. [music] Before we jump into the uh topics, uh we have uh a couple of questions from Tim that [music] we're going to deal with. But before I do that, I just want to mention to everyone that if they haven't been on the website recently, uh there is a new version of what we call the top traders ultimate guide. Uh and that's essentially a guide to about 600 books now. Um and uh hopefully there will be some of those that will be uh inspirational. Um and uh we put it all together in one uh resource. So you can go to toptraersunplug.com/ultimate and then you can get the guide uh for free. Anyways, um Tim, who has been following the podcast for a long time and is very kind to uh um share and like um our content uh on social media, send a couple of questions in. Um and I think it's quite an interesting question. So he writes, "The options markets have experienced incredible growth in recent years, especially postcoid and the widespread of use of uh zeroday options. uh trend followers most mostly only use futures price data as their inputs to their models. Have we reached a point or will we reach a point in the near future where trend followers should be looking more closely at the option markets and the information within those? And then he says, "As a follow-up question, do you think that by ignoring options data, the performance of existing trend following models could deteriorate or that incorporating options information would produce improvements on existing models? So who better to ask than u than you, Nick? [laughter] >> Okay, I would need gem on this one, but um you know, I'll give it a go. I'll give it a go. I'll give it a go." I mean the observation that um that options markets have become um um to a good extent dominated by the zero DTS is um is not just something that that we did in the news. You know you look into the hard data um you know in anticipation obviously of the discussion I kind of pulled up some numbers. So um currently the volume in S&P options um 60% of it is zero DTS and 40 is anything from daily up to you know longerterm tennos and in that zero DT complex more than half is dital participation. So we can make the argument that it's a place that you know almost like know democratized access to to to kind of intraday leverage um and equally um access to those um I guess cheap bets that you can you know put on the day uh out of the money put or out of the money call depending on the on the view um for a significant premium to be had um or significant payout to be to to be had at a very small premium. uh I guess the question on how this I guess market evolution in the option space can to a certain extent impact and followers. I would look at it in two ways. The first one is how we calculate signals, how we think about establishing market trends that we kind of capture them and then we end up deciding to go long or short. But secondly, I would also look at it into the use that those um options and derivatives can have in risk managing trend following portfolios. So if I go to the first one, right? If I go to the first one, it's it's it's all about the marginal or not impact that dealers would have when hedging their gamma, which is a consequence of investors selling or buying options. So if an option you know if if investors are selling out of the money options you know to generate some carry for instance then dealers would typically be like no long for example and therefore in this activity they would be buying when the market drops and they would be selling when the market goes up. So they have a tendency to min revert to to to force a min reversion behavior. Conversely, if you have the opposite activity, you can be in a negative gamma situation and leverage ETFs is perhaps one reason why this this dynamic can play out uh that you end up buying as the price goes up and sell as the price falls and that is exacerbating trends. So you being a trend follower, you might be in a situation whereby either volatility is suppressed and min reversion is I guess a the artifact of of this auxiliary activity or that prices accelerate at a higher volatility dynamic. So then the question becomes in the absence of those dynamics how would trend in itself perform and you know would be established for us to be able to capture it and I think there is a question here to be had you know should we somehow take that into account not too clear to me how but there is certainly a good question to be had specifically for shortterm managers I think the longerterm managers you know if your signal is very close to zero anyway your risk exposure would be small to be honest with you so does it change the Um if you like the the um the measuring of uh of the signal probably not. Um the other point which perhaps is equally interesting is that you know on specific days that those options specifically uh now not just the zero DS but more broadly um you know have an expiration you have this kind of so-called pin risk. Uh so there's a bit of gravitational power towards the you know the strike price that um that the open interest is large. Um, again, it can have an impact as to how the signals themselves are documented for a trend follower. For a medium-term trend follower, is that impactful? Not very clear to me. Just to be just to be open. Um, with that, you know, at the same time, should be ignored. Certainly not. But frankly, I think more value at least in the present market environment of utilizing those tools to maybe take intraday exposures when the market goes against you, right? So suppose you're holding equities and you know equity starts selling off. Uh so your long position in your long-term model would either take some time to revert or during the day it's going to suffer. Maybe there is a premium to be spent on a zod option to just cover that exposure. And that is more how we can use those more shorter term to risk manage the portfolio without cutting the exposure but rather adding to it at a cost a kind of short-term protection. Um so think know it remains to be seen what the outcome can be but these are some of the some of the quick thoughts. Maybe the last one I would have I think there was a paper recently but you know I hope I don't misrepresent it now. It just came up to my mind actually. Um I think it's by Greg Vilov and some colleagues. Um I think it's a project perhaps sponsored by the CME if I'm not mistaken or CBOE. >> Um doesn't really matter. The point I think they're making they're looking into zero volumes. Um and they find that dealers are typically more long gamma than short gamma. And therefore there is an argument to be made that short-term reversions have become more prevalent than you know than than intraday trends. um which again in itself um can have an impact as to how markets trade intraday. I think we've seen a good amount of reversals uh recently and and over the last few years in daily is that caused by ZD is there is certainly a a a force into it. Um so that's how I look at it. Point number one how the models can become aware if that is required. Number two, can we actually make use of them for risk management? I think the latter is more of a direct use case. a former. >> Okay. >> Maybe. >> Okay. Maybe. >> So, so I would say I tend to agree with you on this one. Um, and but I think about things in in much simpler uh non-quanty terms. Um so in in one in one way I would I just make sort of a simple observation saying yeah the last period of time it's been definitely more challenging for say short-term strategies and maybe that is exactly because there is a force out there that is much more convergent uh and has this sort of mean reverting um interest and that could well be from uh all the people around you know surrounding the the the zero DTA options markets and and how they uh manage risk and so on and so forth. That would be one observation could also be the pot shops. Uh but they're probably also part of maybe the zerodt options market. I have no idea. But the other thing um spotted specifically to Tim's question about you know should we use that data? And I'm thinking what would would we use it for? We're trying to find trends that last for 3 to 6 to 12 to 24 months and a zero DT option volume. And again, not I'm not a quant here, so but I'm thinking, okay, that gives you a sense of people making a bet for a single day. That's not really going to inform me what the price is going to do over the next 3 6 9 12 24 months. And but even for risk management purposes, why would we start trading options when we have plenty of liquidity in the futures markets? And I think one of the trend following I mean one should never say never but one of the trend following mantras has always been you know you you test what you trade and you trade what you test right and and and we are we are using futures data to build all our models and so we should trade futures um and by the way we adjust positions on a daily basis it's not like we're making massive bets any any one day unless something really crazy uh is taking place. So I'm I'm less certain about the use case of option datas for classical trend following models. Um but as I said I'm not a quant so I could be completely wrong here. I mean I can see the hypothesis that uh periods of min reverting behavior can create excess turnover without necessarily being any particular reason for it. I think most of our models do some sort of short-term smoothing anyway. So no maybe we can look at the symptoms of the whip sowing dynamics ultimately if averaging helps a bit moderate that as it has historically done with let's say asynchronicity possibly it's captured indirectly but I can see the hypothesis there I can see the hypothesis of maybe some sort of option implied information like no skew um providing some sort of sentiment indicator that I don't know that in itself might be like a a penalty to the specific direction of the market or maybe multiple um so you know there could be nuances >> anyway full stop full stop there's no comment in my sentence I think I I think I think these are like some of the hypothes forward as a consequence of team's questions but as long to get those questions >> allows us to think about stuff that we may not um think about on a dayby-day basis >> always do all right well let's jump to the papers now we had kind of several choices I think we've narrowed it down to two or three that we liked in particular. Now the first one and by the way courtesy uh mostly by our friends over at man Group um because they are they have been very busy um recently um and they have put out some really interesting one although the first one is a probably a few weeks old um and I think Al and I quickly touched on it that's why I was delighted when I saw you wanted to say a few words about it and even more delighted to say that I think I might get one of the authors authors of the paper on the podcast along with Katie in a couple of weeks. So, we could also maybe leave some questions um for him. Um you never know. So, as I said, I may have touched on it already a little bit. It's a paper uh called a trend following deep dive, the optimal market mix for a trend follower. Now, of course, the topic itself is uh relevant because we talk a lot about it um over the years. what is which market should we trade? Uh do we trade uh too many too few? Uh and and all of that and we all have our different views and and we favor different uh solutions but it's not always we've been very good at uh eloquently describe the benefits of doing you know one choice versus another choice. And I this is what I love about the paper. It's it's a very visual, easy walk through um in terms of what people should be likely to expect from choosing managers trading different market universes. I think they they made that in in a very eloquent way. Um [snorts] of course, if you want to read the paper, you should go to the um man.com and the insights. That's where you will find the paper. Um so um anyways, I very much look forward to hearing your thoughts about it. uh Nick and then we'll we'll we'll take it from there. >> It was a very good read, right? And I think so. Why did that resonate quite well with me? Because we always discuss how defensive your trend follower is, what's a universe, you know, a core market, a broad market, alternative market, whatever market. I think the reason why I like this particular one is because it connects a use case to the universe that you use. So we typically say you want to be defensive, be faster, you want to be longer term performing, be slower, but you know bear in mind you're going to get a lot of beta risk in your portfolio like you know we always have this conversation but I don't think we have ever touched upon the point of you want to be defensive this is the universe to do it with and I think that's why the paper is interesting because it says or maybe taking a step back when we speak about managed futures um there is this kind of duality of objectives I think cliffas wrote about it a couple of years back uh and I keep on using now this kind of duality term. Um it's one of the very few systematic strategies that deliver long-term positive returns and kind of downside protection some sort of a crisis alpha to use Katis term some sort of a reactivity when the markets are falling but not obviously sharply but you know in a kind of a medium term. So this duality of objectives frankly I don't think anything but trend following commodity curve to basically say the the one that we started from. So it's typically defensive as well because kind of shorting in front of commodities. Um so in a noninflationary recession it actually performs well. Uh and then I guess interest rate volatility. There's nothing else at least in my mind that uh doesn't showcase a trade-off between kind of reactivity or defensiveness and and some sort of a kind of cost associated with it. So if we look into this duality which is performance and defensiveness, we can stretch it out and say, well, how can I maximize my defensiveness and how can I maximize my kind of long-term sharp ratio? And surely there's going to be like a trade-off now between the, you know, I guess the choice I have available while still obviously maintaining the duality. So it's not about making, I don't know, negative long-term return or making non-reactive kind of a profile. So I think there is a commonality here in the solution that being that the duality is preserved but it's more about kind of the major and the minor in the objective and they make the point that look if you really want to be defensive and defensive is more about generating positive return when the market goes through a a stress situation then you should have a market universe that is more kind of core and more kind of standard and more mainstream and perhaps more liquid >> because no surprise is when you have a massive correction, it is not that assets fall, it is principal components that are falling. So equities as a risk materializes um perhaps I don't know in the duration space you have central bank intervention and and you just want to be long bonds or maybe some of the cyclical commodities except maybe precious metals are suffering from a drop. So ultimately in a period whereby asset prices get squeezed and maybe correlations um are shifting to the extremes you just want to trade the principal component. So if you stick to a core market universe even for the same specification of speed and correlation so on and so forth you'll end up maximizing your kind of crisis alpha or defensiveness. Um and I I think Andre would actually be quite um uh quite opinionated on this one because I think that's part of his uh of his pitch, right? That you know the universe actually quite quite quite short but very much representative of the macro moves. Conversely, all the alternative markets you know maybe the alternative commodities or you know we can go into kind of EM equities or even equity factors that they also utilize in the paper. You know this expansion of the universe by design brings idiosyncratic risk that with the assumption of trendiness would provide longerterm diversification and therefore return and therefore maximum long-term sharp ratio is achieved with a broader universe with more independent bets. Now this is not as defensive as the tide universe would be. It is still defensive to my earlier point but longerterm has a better sharp. So now this poses the question you know you [clears throat] being asked allocator what do you want to solve for if you want to solve for defensiveness how should you look into it is it like a speed discussion is it a universe discussion is the risk management discussion but there has to be a discussion if instead you're looking into it more as an alpha seeking overlay then maybe a broader universe can be more associated with your objective and and and you know you can meet that objective more more successfully so with the recognition that you might not have the best defensiveness you know in in in D. So I think I'm posing here for for you to reflect but this is how I think about Well, I mean I just wanted to ask your opinion as well. Um, and that is would you agree? It's a leading question you can hear. Would you agree that actually that's exactly what I think this discussion brings? Because previously I've always felt that the crisis element protection element was really always a discussion that related back to speed. >> Yeah, >> 100%. And I plead guilty that you know I was all always going back to the speed discussion or maybe the you know the amount of controlling of your equity risk and you can do it with data or maybe some force constraint exposures all that >> somehow it I don't think it was very natural to think of the markets as directly at this as this one says I mean yeah we could control equity risk maybe we kind of thought about it but not to that extent um so I think it's interesting to to your point I think it kind of shifts a bit the discussion >> complete keep keep going. >> Um, look, no, I mean that that's pretty much it. I think they have another um, you know, another interesting kind of um, uh, I guess path that they follow which is um, >> you know, in addition to the maximum sharp price and the and the maximum crisis alpha portfolio. >> Um, looking into also cash efficiency and obviously the fact that we trade futures, you know, futures are, you know, instruments that trade on margin. So you don't have to spend, you know, $100 for a $100 nostal exposure. you just pay a fraction of it which is determined by the exchange. So the question then becomes obviously how is that margin determined and you know that margin should be financed from actual dollars. So the lower that margin is the more capital efficient the portfolio can be. So they make you know a a supposition here that a portfolio that is more cash efficient and therefore you can get higher level of volatility for the same dollar value of national uh sorry of of margin um is a portfolio that most likely would have liquid markets because that's no that's a component of um of of of the margin determination. Now that in itself brings you to a universe that is more of a core universe. So it's not a surprise that the cast efficient portfolio from that perspective resembles the maximum crisis alpha portfolio. Why? because them two for a different reasons want to capture the micro moves, want to capture the more liquid assets ends up, you know, allocating to the same ecosystem of of um if you like of assets. So like anyway, for me that's more of a byproduct of the discussion, you know, very nice to see it. Uh it almost says if I were to kind of reverse the argument, if you were to have the maximum crisis shar portfolio, it is more likely that your margin requirement would be lower. But yeah, to me to me the biggest uh the biggest point is is um is really the the association of the objective to to the universe. That's that that that's really >> another stat from the paper that I thought was kind of interesting uh was they they managed to to say that there are approximately 900 different markets we could trade. That that that surprised me that they could find so many. Um >> this is very true. I mean the the one the one thing I would point out you remind me I had forgotten about this one. If you look into equity styles, so they use equity factors and and they have 60 mark 60 markets or 60 factors. I mean something I'm sure they would they would attest to it and I guess you can ask one of the co-authors when when when he's here. There are no 60 factors that can be seen as to a good extent diversifying in the equity space. Um you know we can talk about earnings to price and book to price as valuation ratios. We can think of those as two factors but technically the amount of cross-sectional correlation is very high and they are value descriptors. So in a way 60 is almost like at least that's my understanding maybe I'm wrong but 60 here is almost like a signal by signal characterization of how many degrees we have to rank single stocks by and build single stock equity factors. But that is slightly at at least in my mind dissimilar to saying I'm gonna have 60 commodities >> because I think in the equities world maybe five maybe six factors >> probably explain the cross-section of returns. So I think 60 commodities versus 60 descriptors of equity returns would give you much less of a diversified universe in the equity space rather than in the in in the commodity space. So uh but you know setting that aside 900 is 900. So outside of that, I cannot really pose significant um questioning on on on on >> that's fine. We we'll look forward to having uh >> but yeah >> having one of the authors on the show in a couple of weeks and we'll we'll probably uh touch on this uh again no doubt but we're going to stay with man group as I said they've been really busy writing and you identified another paper um which is a little bit different I think uh in terms of the topic it's called um alpha trend and agentic research workflow so not necessarily specifically about building models but maybe processes um that can help do it easier or perceived to help to do it easier. Um what what were your takeaways from from the uh the the questions they raised? So I picked this up to discuss not so much because it talks about trend following obviously it came to my I guess to my inbox um as um as the output of of of the research at man. Um I think to me this poses like a bigger question as to how we see those new technologies and AI and Gen AI and LMS uh becoming not just tools that we use uh but eventually become core components of the research process. Um and and you know for the sake of um I guess of of those that are listening the paper talks about kind of designing an agent that can build a trend system. So to a certain extent no running a trend following strategy or any systematic strategy is a very deterministic process. You have your data you clean your data you build a signal you estimate risk you throw that into an optimizer maybe a ranking mechan methodology whatever that is you get target weights you have some liquidity control you have some volatility target and here you are the quantities to trade on a daily basis or whatever basis. So that process is very deterministic and certainly you can think of a world whereby single agents do that job for you and there's a kind of a supervisor that allows you to command this whole ecosystem. Um and they make the point that you know the current um kind of chat chat bots that we have are probably kind of more shallow in the amount of depth but very broad in the topics we can discuss. You know this model of conducting research is much more targeted. So very narrow in terms of scope but very deep in terms of um uh in terms of analysis. Um so in that context you know again to to to their paper you know they kind of give it out they give it like um a breakout signal um and they deliberately remove some good feature out of it and then they obviously ask the model to find what this feature could be u by describing it quite um uh I guess describing it with words and yes the model does pick it up and does outperform but you know then there's another feature that you know should not be um value accredititive and then the model indeed finds that it's not as useful as you'd expect it to be. Um and and broadly speaking they make the point that you know the model can actually go and operate as as a human being good but equally I think that's even more important that you know human judgment remains extremely extremely important not just from how you frame the questions and how you guide the process and how you interpret the results but also kind of guarding safeguarding against multiple testing overfitting like things that we've been discussing for years and how the culture of a research team is more important than the research team itself. Um I think they make the same point but now the research team happens to be like um kind of an agentic model rather than an individual or like a team. Um so I'm kind of bringing that up maybe the last point. Um they they kind of use different um different models like you know cloud and and GPT and so on so forth uh with the same prompts and and they found that the outcome was actually quite different uh but not necessarily worse or better. Uh so one for example was more of a kind of a single direction-minded you know with with with very correlated outcomes. The other one was a bit more dispersed in the way that it kind of treated the data. But broadly speaking the one thing that at least remains at least my personal view is that we still require this kind of critical thinking. Um I think the synthesis that those models can do is is insane. like I'm personally surprised every single day uh you know by using the tools but I think having some sort of a disciplined evaluation of the outcome and having critical thinking of the outcome is extremely extremely important. I'll give you this example unrelated to the paper. Um you know I I went to one of those engines um recently and on purpose I asked the following question. How has the first two months of the year been for trend following of all things? because I know the answer and I know that you know January and February were probably you probably one of the best two-month periods to start a year for trend followers right and you know the first sentence that comes out and I don't even care about what followed was like the start of the year has been mixed and I'm like you're so wrong [laughter] um so to to to that point I think some level of supervision is is more than important here it's actually essential specifically when you end up kind of managing money on behalf of of policy holders and and pensioners and and and is insurers. Um >> and on top of that actually I think a lot of what um >> companies >> I think that the value of what uh our esteemed researchers do nowadays is actually trying to keeps things simpler. Uh meaning uh I have a feeling without being an AI expert that they love to expand on things. That's kind of what they do. They find sentences and words and all of that stuff. But actually when you build a trend following system, yeah, you have a lot of options, but um but actually the skill and and the experience um goes towards actually stripping things down so you get the the cleanest uh signal, less noise. Um and uh and I'm not so sure that AI is really, you know, the DNA of AI may not actually be very compatible with that. >> We shall see. I I think the um the amount of um evolution we've seen in the last few months is is extraordinary. >> If I were to speak about like my personal experience and how I use it, it's just um for the same task. Uh I I I used to get rubbish. >> I'm getting high quality outcome now. >> Um so it's it's it's quite impressive. We shall see where time where time goes. But um I I found very interesting that you know they actually ended up kind of writing actually the second report on how they use AI for research purposes and obviously kudos to them. >> Um I guess opening up and and making that a topic of discussion rather than saying oh we use this model here's the line you know go trade with it. Let's make it three for three. Uh Nick, because you uh identified a third man uh paper um the quant renaissance >> exactly three for three. Talk to us about uh this paper and why it caught your your attention. >> No, this is now shifting gears away from trend following. So this is about uh contequity. So that's from the numeric crowd at uh at man. Um so I'm spending a good amount of my time to you know with uh with our equities. um kind of single equities uh strategies. It's a very interesting space because it's maybe one of the places that systematic invested started from. Obviously, trend following is probably the longest living from the 70s, but if I were to pick what the second one is, probably equity factors is the second one to come. And it could have equally been a an 80s or a 90s gig. um you know when Pam and friends came about um kind of the whole factor uh and then obviously uh who was a civ rose with a so this space of of equity factors driving returns and therefore being rewarded for a specific risk exposure um could have been a an investment mantra back in the '90s but it only became much more popularized when you know when we had you know high performance computers to turn all this data and the cross-section of stocks and you know did corporate actions and so on and so forth. Um so that space obviously became very very popular and then coming 2018 to 2020 it suffered from this kind of quant winter uh that had significant consequences also for the for the for the buy side managing those u those strategies. Um you know lo and behold postcoid uh significant resurrection uh so the multistrand the quant equity you know the long short space um has had very strong performance um and and what this report tries to bring into I guess into the discussion is how much of a risk we have of a repeat of a winter but also maybe how more resilient the space from a research product uh and and and and uh investment management has evolved over the years. Um so why did we have the winter back in the days? Frankly, there are two reasons. Um they bring about I would possibly just agree without having seen them. One was that the macro regime was not very accommodating for some of the factors obviously with the interest rate markets operating as they did back in the late late part of 2010s. You know, value was certainly impacted quite significantly. Then you had some reversions and the momentum did not play out well and the overall complex you know was just not accommodative of the micro environment. The second one is probably possibly some factor crowding. So the the theme became too quickly too popular to having not just perhaps alpha decay consequences but also having forced unwinding um events when some sort of a macro shock or funding liquidity driving uh driving some of the um driving some of the performance and and and then you have this circus of of of unwinding and obviously one is causing the other and so on and so forth. So some of the far cell dynamic and and you know if I were to quote some of the some of the stats in the paper they say that during stress periods you know classical macro factors can explain more than 50% of factor return variation which is quite substantial if you were to think that you know normal times is more like know 20%. So that was what happened back then and now the question is obviously if we bring this world to today how things have changed. They use an internal model they have for regime identification. So no surprise you typically have kind of recession early expansion late expansion and overheating. So the four typical regimes that the equity market or broadly the macro market goes through they utilize that to understand better the quan winter uh and associate if you like different premier different different environments. But then they make the argument that with us understanding better how the micro cycle works at least through the lens of factor investing. Their argument is that now we as investment practitioners you know as managers um are utilizing more dynamic allocation in the factor space and less static and I think that's something we also see and you know we have seen the QIS space doing and I think um it is now much more of an expectation that you have a dynamic allocation mindset rather than a static one. Um so there is some value that uh dynamic allocation can bring with I guess the caveat of uh concentration. Um there is obviously new data. So what historically used to be your quality score and your momentum score and your value score can now be either enhanced or expanded in the factor space by kind of geocation data and credit card data and you know you name it and patents and uh kind of sentiment from the buy side from the sell side you know using ML models. So there are new data to be utilized uh which inherently provides some diversification. So they put some quotes that for example you know alpha models back in the days um you know would have um correlations of like know u 70 to 80% these days more like know 40 to 50. Uh so there is some diversification at the signal level. Um and lastly um they make the argument that you know going through the more dynamic allocation going through new data going through more modern techniques uh designing portfolios or or crafting alpha scores eventually provides some sort of a macro regime resilience. So what was back in the day a macro regime dependence can become a macro regime resilience with the use of all those technologies and data sources in the quantity space. I think you know equities because of their dimensionality uh they have been the most researcher and they will continue being a space that return can be harvested. Um so I found this one quite um quite an interesting one both for myself as well as I guess for for for this discussion to be to be brought forward. So that that that's the whole story right is the quant winter something we can see again and if that is the case you know what drove it back then how the ch how the how the changes that we've seen in the space have maybe reduces probability um if the micro regime is not very accommodative so that's >> in a sense we could link that to trend following right the people talk about a trend following winter a few years ago as well right I mean have we >> I thought I thought you said in early March >> no I was thinking about the 2010s where people were complaining. I know. I know. And and where people were sort of um disappointed with returns. But when you think about the macroeconomic environment, low inflation, stable inflation, uh very little movement on GDP and so on and so forth, of course, um that's not necessarily the best environment. The question is, of course, with all the research we do, should that happen again, would we cope much better with it? That's kind of the same um argument or question that uh that only time will tell I guess. I mean I think I think we've discussed that maybe that was the that was the reason why we first met in the very very first place right you know this whole discussion I remember back in the days having about fundamental certainty and uncertainty and how the Fed put brings a lot of fundamental certainty I would not claim that currently we're sitting in a fundamental certain kind of environment >> doesn't look like it so >> doesn't look like >> certainty brings reversions uncertainty allows price trends to continue because a price trend in a certain environment provides information about where the fundamental value should sit whereas certainty brings you back to your prior which is very informative of valuation. So that I think this is the dynamic at least in my mind that the fedwood was was bringing to the trend space and therefore the challenges that came about. No, absolutely. This was great. Uh Nick, thank you so much for spending time preparing for all of this, finding these papers um and discussing them, of course. And uh for all of the for all of those of you listening um you know feel free and and please uh do uh head over to your favorite podcast platform um or YouTube and leave a uh rating and review to support uh the channel but also as a thank you to Nick and all your other co-hosts who do a tremendous job every week in uh preparing for these conversations. Um, before I wrap up, uh, let me just say that, uh, if you have questions for next week, which is where I will be joined by Yav, um, then send them to info toptraders.com, uh, just like Tim did, uh, this week, and I'll do my best to bring it up in our conversation. Um, with that said, from Nick and me, thank you ever so much for listening. We look forward to being back with you next week. And until next time, as usual, take care of yourself and take care [music] of each other. Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the [music] best way to stay updated is to go on over to iTunes and subscribe to the show so that you'll be sure to get all the new episodes [music] as they're released. We have some amazing guests lined up for you. And to ensure our show continues [music] to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it's the best way to [music] show us you love the podcast. We'll see you next time on Top Traders Unplugged. >> [music]