What's up, everyone? Welcome back to the outlier podcast. I have a familiar face, Augustine LeBron here in red celebrating the bloodshed. No. I'm just kidding. That's actually something we'd say in the military, though. But we are gonna be taking a look at strategy development. And this is something that you and I had gone back and forth in DMs a little bit on doing a follow on session, and you expressed that you wanted to do something, like, more practical and kinda hands on using a live example. And I love that because it's really easy to talk about stuff and to try and show pieces of it. I think that's a lot of value. I think that's how a lot of people can learn way more effectively. So the vision for today is we are going into this cold. He and I have not reviewed any of this stuff together. This is honestly stuff that I just kinda cobbled together for the sake of a conversation. And what we wanna do is like a back and forth. I'm literally gonna talk about some of the stuff that I did for the strategy, and then Augustine is gonna ask questions, poke holes, look for follow ups. And the purpose behind that is from a listener perspective, you should effectively be doing what Augustine and I are doing, but by yourself. You have to do both sides. You have to do the generation of stuff, and then you have to do the exploration and external analysis of it. If you have people you can trust, all the better. But this is designed to provide a framework. So, Augustine Mhmm. I'm stoked, man. This is gonna be great. Like you said, I don't know at all what you're gonna do, but I think my intent here is just to ask, hopefully, useful and interesting questions that jog some conversations. See how that Beautiful. Alright. So I figured that we would start with just a little bit of level setting for those listening on, like, Spotify. You're not gonna see the screen, unfortunately. For those on YouTube, I'm gonna have the video so that you guys can see the the stuff that we're looking at. As a starting point, if you and I were to think about just a strategy, what is in your view? Like, what is a strategy? What what do you classify as a strategy? I would say it's a fairly systematic, at least in my world, a fairly systematic process that takes information about the world and transforms it into actions that you might take in the world. I think that's probably the most general way I can think of it. It's a it's a really interesting definition too, and I I just wanna double click into 1 part of it, the fairly systematic. That's a fascinating that's a fascinating starting point because if you go on a place like x, it you know what I mean? There's, like, this massive spectrum. Like, somebody is doing it off of horoscopes, which I guess you could argue is a system. It's kinda crazy, but it's a thing. And then there's the, you know, the more really heavy quant side where it's like, well, unless it's fully systematic, it won't work. So when you say fairly systematic, could you elaborate a little bit? Me I think I think of it as look. I'm not good at discretionary trading. I've never really done it. I've never really tried it. I'm not trained to do it. Like, it's just not a thing for me. And so I think that my area of expertise to the extent that I have any sort of starts somewhere along the spectrum of fully discretionary to fully systematic. Like, somewhere in the middle there, I feel like I I go from being uncomfortable and not really sure to, like, having some more confidence about, like, how to structure a thing that could be useful. And so that I think I think that's maybe more a reflection of where I'm coming from than some, like, hard and fast rule about what is systematic enough to count. I don't know. It's just like what I wanna say. Right. Right. And I and I think the context is super useful. So 1 1 more probing question on that. What is an example of a strategy that you're aware of or that you've heard or you've come across that you would consider not systematic? I think, and maybe we talked about it, in our last conversation, but I certainly know plenty of people who, are not very systematic in how they do their trading. Like, they read the newspaper. They read Bloomberg. They think about things, and they're like, I wonder if I wonder if this is, like if this makes sense. And they maybe maybe some of the time they do some trades, and those trades work out. Like, it's literally them just every day trying to be smart, and I'm sure they're looking for patterns and stuff they've seen before. Like but but it's not, like, systematized in the sense of, like, here is the framework that I'm using, and here are the rules that I'm going to apply. Here's the data that I have. It's much more, we'll say, based as opposed to, system based. So it's, like, maybe 1 spike 1 end of the spectrum that I'm familiar with that I mean, these guys make money. So it's just not something that I know how to do. Sure. Yeah. And and I think that that gives at least, you know, folks an idea of, like, the lens, right, that you're applying, which I think is useful. For me, I I find myself actually quite similar to you. I I'm absolutely I'm a discretionary trader, but I consider most of what I do parameterized. So it's not this completely open thing that I get to do whatever. There are pieces that are clearly defined, and I have to play inside this world. And it's a scale. There's some effects that I play that are very systematic. You just do the thing. And then there's other things where there is some creative license, and that's really, really challenging to balance. You have to track that very carefully because if you don't, then it's really hard to tell if your discretion is value add or not. And Exactly. Yeah. It it's a bit of a black box. Yeah. So Yeah. What I prepared for us today is 1 of my premises, especially for newer traders. And this is something I've come across in the last few really the few months. Normally, I wouldn't really have, like, a direction that I would push people in. I would normally just say, like, kinda go see stuff and see what you see. But then I realized from my view, especially as a, like, a private operator, there's a ton of shit that you have to do, and you have to do a lot of things right. And my argument is if you're doing all of those things, the process side of stuff, but the underlying market effect that you're trying to monetize just isn't there for some reason, That creates a really awful feedback loop where you could be doing all the right things, not getting the right outcomes. So when you go back to the things you're doing and change these even though they might be right, might be wrong, you don't know. You have 2 variables at this point. Yeah. So Yep. What I've started thinking through is like, well, if I were to start all over again, where, like, where would I try to go? And there's a couple things that we generally know. We know things like positive drift in the market. We know things like if there's a breadth of research on it that you can probably understand something about it even if you don't know what exists. So what I mean by that is if I don't know anything coming into trading and I'm doing research on the stock market, SSRN, or whatever, and I come across some different market effects that gives me at least some sort of foundation that is admittedly resting on other people's work and their analysis, but the expectation there is that they might be smarter than I am at that point. The reason why I'm couching all of this is because there are a handful of broad market effects that I think are really well researched. 1 of them is momentum. I talk about momentum a lot, and that's kind of become my default angle to put in front of people not ever saying trade this. I'm just a fucking Internet guy. You should do exactly none of what I say, and I'm very serious about that. You should do whatever's right for you. I I'm just an inter I'm just an Internet guy's guess, so it's even worse. Exactly. We're 2 Internet guys. That's that's exactly we're we're on opposite side of the world right now, just Internet people. And I think the the reason though that I think Momentum is a great place to start is because it's incredibly well researched. You might not inherently find hyper profitable stuff right out of the gate there. Although, admittedly speaking, you could do very well with something like momentum, again, depending on what, quote, unquote, religion you subscribe to, but there's a lot of research out there that would lead us to believe that it's at least somewhat promising. So Okay. By starting with a market effect that's well researched, that gets limits 1 of the variables that we're working with. Because as a new trader, you still have to get all the process stuff right, but now you're at least starting with the market effect that is generally documented and known that led me to pick momentum. Great. Now for the sake of ease and simplicity, I opted for a sector rotation strategy. This is actually something I used to run kind of a lot, and it's a really boring idea. But what it does is take the concepts of momentum and applies it to sectors. So what I thought we could do really quick is I'll share my screen, and I wanna walk you through some of the general, like, research that I think is worth looking at to show people, like, how to pull some of this stuff up. And then I thought maybe you and I could go back and forth on, like, this idea generation phase, and then we'll move on to, like, the, you know, strategy development phase, if that makes sense. Great. Let's do it. Great. Let me grab my screen. Should be coming up. Can you see that? I can see a black box. There we go. We got it. Perfect. Alright. Let me move this out of the way. Okay. So the 1st thing that I have here is a book. I'm not, like, I'm not, sponsored by this guy or anything. I know Wes.

Play quite well. He's super cool. And if I've had him on the podcast before. He's a marine vet, by the way. Shout out to my boy, Wes. And, yeah, he's 1 of those guys that's done, like, 1000000 things. It's kinda crazy. And so he, funny enough, is, like, he loves momentum. And he put out an awesome book. And, again, I'm not putting this in front of you to, like, cosign or anything. Nothing like that. This is a good piece of research that people could consider starting with with something like Momentum. And what Wes does is go through Momentum, like, what it is, why it exists, that kind of stuff. This is useful, but I think that before somebody might opt to get into a book about momentum, it might make sense to pop onto a website like SSRN and actually just go look at the research yourself. And Wes does a good job kind of combining different pieces of it, but there's really never a replacement for, like, quite literally putting your eyes on the thing yourself so that you can try and get more familiar with it. So And is like the, like, the OG of Momentum papers. It's literally took the words out of my mouth, and that's why this is number 1. I have, like, a bunch of different papers here that I think are generally good, but, literally, that's exactly what I was gonna say is that these 2 guys are probably 2 of the the names that you wanna look for. You'll find a lot of their stuff is old, but they also have updates on a lot of it, which is kinda worth cycling on. But what I did to get these papers is I went on to SSRN, Social Science Research Network. It's completely free. And I just typed in stock market momentum, stock momentum momentum, and then I filtered by a couple different things. I filtered by relevance, downloads. I filtered by new releases. So things that have come out, you know, more recently versus stuff that came out more aged. 1st, I wanna get your thoughts on that process of a new trader coming up with some sort of idea and then going to a place like SSRN or any other place that you might recommend. Yeah. And any other tips or tricks as they go through that? Yeah. I think I think this all makes a ton of sense to me. I think 1 of the things that I would certainly say is that, like and you probably are gonna cover this even more going forward is, like, you need to get your reps in. And so you're gonna download a paper, and you're gonna replicate it. You're gonna try to replicate it. You're gonna do a crap job of it. And then you're gonna, like, kinda figure out, like, why isn't my numbers why don't my numbers match their numbers? And and then you're gonna look at another paper, and you're gonna try to replicate that. And, like, you need to get your reps in of, like, doing this. This needs to be pretty straightforward for you to do after a while. This is, the we'll call it, like, the basic blocking and tackling of thinking about systematic strategies is, like, can I take an idea that I see and just kinda quickly spin some version of it up that's not horribly broken in some weird way? And so, yeah, like, dive in and let's do it. You know? Like, there you can learn you can read all you want, but unless you're in front of a computer actually trying to replicate things, you're not gonna learn the things you need to learn. And another thing that I wanna emphasize from what you were just talking about is the fact that things won't necessarily line up. If you're doing something like this, what would you consider or what would you look for is, a benchmark of, okay. I did a pretty good job here. Understanding the fact that it will effectively never line up 1 for 1. And for those that don't know, a lot of that has to do with, you know, the data source, the assumptions, the way they run it, like, all of these tiny variables that wrap up to difference. Part of it makes it like, part of it's really down to the quality of the paper a lot of the time. Like, if the paper just takes a bunch of mids and just run some, like, numbers or some some stats or something on on a bunch of mids, that's probably pretty straightforward to replicate. If they're actually, like, trying to do a good job, they're probably doing things like figuring out what borrow costs if you're gonna get short something. Right? So that's, like, something. And then they're gonna try to figure out, like, what trading costs are. They might even try to, like, model market impact in order to get a sense of capacity. Like, these are all things that better and better papers are should try to do, and and these are all things that are progressively harder and harder for you to do the same thing they're doing in order to replicate. And so I think it just kinda depends a lot on, like, what the quality of the source paper is. And I think something that's worth keeping in mind, and I I wanna label this before we even look at back tests and stuff, is the fact that doing all this modeling and stuff, it's really designed to help understand something, get an idea of some things, like, worthwhile to spend time on. I think it's really, really important to emphasize that, like, let's say we replicate this perfectly and we get good results. It looks good. That doesn't mean that what you have is actually good in going to work going forward. It simply means that what you have might be promising. It might have utility. 1 of the biggest pitfalls I think somebody could make from the beginner lens doing something like this is, like, going through replicating this, being like, okay. Bam. I'm I'm super close. This thing's gonna work. You've kinda set yourself up for a pretty bumpy ride from there. Okay. As you're going through different kinds of papers, I also like to kind of strongly consider the the background and the basis of, like, who wrote the paper like you talked about. And, again, that's why these guys are here, Jaggedish and Tippmann. They're kinda like like, literally, the OGs of Momentum. These guys have been looking at Momentum for a long time. It's kinda like Barbara, Leon, O'Dean. They look at a lot of options, behavioral stuff. You'll find as you look at research for a while, you'll, like, find pockets of people who do things. Yeah. It's definitely like a pockets business for sure. Like, you start to recognize the names that are referred to in papers that you read. Like, that's usually like, if you look at the bibliography paper and you're like, hey. There's 3, 4 papers by these guys, and then you see, like, another paper in this area and you see, like, the same references, probably you should read those source papers. Right? And you'll find that there's some, like, pockets of people that are just sort of, like, the the the known authorities or at least the the high status, if nothing else, authorities in a given area. Massive tip. Massive tip there to emphasize looking at the citations and seeing what you regularly see there. That's huge. I wanna flip through a couple of these really quick. Some of these, I think, are good. Some of them, I actually don't think are good. I just wanna get your like, if you pull this up and I can zoom in a little bit more if you need. If you pull this up, you downloaded this. You just found, you know, momentum. You downloaded a paper on it. I haven't seen it before. So let me just read the abstract. So this is what I'm gonna do. I'm just gonna read the abstract. Blah blah blah. Novel alpha momentum strategy in stocks. 3 factor alphas. Estimate using daily returns. Okay. Like, immediate like, okay. Daily returns. What year is this? This is what year was this paper? Somewhere? Like, if it's if it's only returns and it's a very recent paper, I'm like, this sounds this already sounds sketchy. Like, if you can't get your if you can't get your hands on, like, 5 minute bar data in 2026 as an academic researcher, like, how much effort are you really putting into this is maybe, like, the 1st question I'd ask myself. Alpha has power in predicting cross section. So this is like a cross sectional momentum thing instead of a time series momentum thing because there's kinda 2 different varieties. Alpha momentum dominates price momentum only in The US. Okay. That's weird. Connecting both strategies, behavioral explanations, blah blah blah blah blah. That's kind of academic stuff. Under reaction to firm specific news, while price momentum is primarily driven driven by price overshooting due to momentum trading. Price overshooting due to momentum trading, that doesn't like, that immediately doesn't make a ton of sense. Like, momentum is an underreaction thing. Like, mean reversion is more of an overreaction thing. So there's a couple things there that I don't really understand, but, like but that's fine. That's there's like, that doesn't mean they're wrong or anything. It just means, like, okay. These are a couple of things that I don't quite understand reading this abstract. And so then what would I do? Well, honestly, the 1st thing that I would do now after having read the abstract is that I would make a decision. Is this worth me spending more time on? And if I say, yes. If I decide, yes. This is more worth me spending more time on, the the very next thing I would do is I would say, if I had written this abstract, what would the paper that I write look like? Like, how would I implement a paper that has this abstract? That's, like, the 1st thing that I would do. I could really think deep very deeply about what data I would need, what sort of, experiments I would run, what kind of tests I would do, blah blah blah blah blah. And I kind of sketch it out in my head, like, what is this thing gonna look like? And if I'm excited enough about the abstract, I might just actually go do that before reading the paper at all, like, just based on the abstract. And I just because then I'll have, like, a reference point of, okay. This is what I thought when I read this. And then I'll go read the paper and see what's the same and what's different. Because that gives you a very, like, clear lens on, oh, that's what's kind of unique and special and whatever about what they're doing or, you know, something like that or, oh, that's weird. Like, I totally accounted for this effect that they didn't. Okay. That's interesting or vice versa. Right? So I think that's maybe a long winded answer to your question about what I do in this situation, but that's what comes to mind. That's super useful. And I had 1 follow-up. You mentioned this, like, key threshold where, like, you decide if it's worth it or if it's not worth it. I can

Infer some of the criteria that you would be using in that scenario. But I'm thinking things like if there's a novel idea that you're not familiar with but looks intriguing to you, like, what other kind of things would you use as your decision point? Yes. I'm gonna continue. No. I'm not. Yeah. I think, like I said like you said sorry. Yeah. Definitely. Because there's some, like, kernel of a new cool idea that I hadn't thought over that might be something novel to me. Definitely probably the biggest thing. And then I would definitely look at the authors and their their affiliations and kinda make a decision about whether I think these are, like, credible enough that it's worth me spending the time to look at this versus me spending the time looking for and hunting for all of the ways they've made a mistake. Because I think it's worth saying, Eric, I don't think you disagree, most published finance trading research is crap because if it weren't crap, they would go trade it. And so, like, there's a high bar in the world of finance academic stuff that that just exist because of this sort of adverse selection. It doesn't mean, like, all academic finance is crap. I'm not saying that at all. I'm just saying the bar is high. And I think of papers as idea jumping off points more than things that I read, replicate, and then do. Like, I think of them as sources of interesting ideas more than anything else. If you see somebody with, like, asset management, hedge funds, like, anything like that in the name versus, like, a like, a university, a school, how do you interpret those different origins? It depends. It's a great question. I I don't know that I have, the correct answer, but here are some things that come to mind. If I've heard of the place, that seems, interesting. If I've not heard of the place, the bar is higher. So that's 1 thing. Having heard of the place isn't either you could imagine it being a proposer. Oh, this is a place that's good. I should read this paper. You could also imagine it as a disposer. Oh, this is a place that's good. If they published it, it can't work because if it did, they wouldn't be doing it. So you you can kinda see it both ways. And so you end up, like, seeing the names for a lot of things, like, for a lot of these like, for example, I don't know where he works now now, but, Antti ilmenon. Mhmm. Where does he work now? I think he's just still at AQR. I'm gonna assume he's still at AQR. Yeah. He's at AQR. Right. So any paper by Anti ilmenon, I'll I'll, like I will almost default to reading because he's written really good stuff in the past. He's he's very experienced, etcetera. And so he's at AQR. Like, yeah, that's, like, that's a good guy. Right? Like, I'll probably read that. The thing about universities is if it's university affiliations, sometimes it's some dude's master's thesis, and those are just shit. Like, the bar to a master's thesis being good enough for me to want to actually think about is very, very, very high. If it's if it's somebody who's, who who's, like, a professor at a university that I've heard of, that like, then the bar is lower. Like, they're probably legit. Right? Like, they're probably doing something reasonably sensible. And so, yeah, it's just sort of the soft thing that that I think you develop over time when you read enough. And something to emphasize there that you're intuitively doing is, like, thinking of why is, like, why is this person doing this? What's the reason for what's the incentive here? And some, like, normal incentives for asset management places is for legitimacy or to get their, you know, recognition up, get their brand up, or universities just as you're talking about. Sometimes it's people that have a passion project in a thing. Sometimes it's people that are just doing a master's thesis, and they gotta do something, and they pick this thing. The reason why those underlying things become important as you read through these is you'll actually see a lot of that represents itself in how they do the the rest of it. There are some people that just wanna make simple assumptions, which I'm not gonna lie, I did for mine because it's faster. And you can, like, rough out an idea, get it through, and be done, but there should be a really big difference between, you know, what that study shows and what you might actually find. Yep. 1 last question I had for you. Mhmm. As you kind of flip through different studies, And if you put yourself in the lens of a beginner Mhmm. Like, how many do you think you would at least wanna look at before you might have an idea? And the reason why I'm asking this, I know there's no number, but if you don't know what you don't know, you don't know if it's 5 or 500. You have no idea. Yeah. I that's a good question. I don't know. I'm gonna think back. Like yeah. It's a good question. Part of it certainly comes down to your background. Like, if you have a background in data science and programming and stuff, like, you're just gonna go faster, and you'd be able to do more of them and and just, like, get get your it's like doing your scales when you play piano. Right? Like, it's just you'll you'll be able to go through it well, and and you'll get fast fast. And if if you don't have that kind of background, it's probably gonna take you longer to, like, go through the 1st few and, like, really sort of understand the mechanics of where do I get the data and how do I clean it and, like, make sure that I'm not looking forward in some weird way, a causally, like, all this sort of stuff. Like, it'll just be it'll take you longer. So I don't know. I would I would guess, like, probably minimum 10. Like, just kinda go through things and and kind of get a lay of the land, maybe. I I don't know. I I don't I don't wanna pin myself down to a number, but it's gotta be gotta be a few. I think another, like, way to think of it is as you're reading like, let's say you read 3, and then you read another paper, and you start to see stuff that's familiar. You're like, oh, yeah. Okay. Like, I I've seen that before. I understand why they're doing that. As soon as you can start kind of, like, more directly following what's going on preempting what they might do, that's kind of a good indication that you've built at least some level of familiarity. Because the 1st few that you're gonna read, you're just gonna be trying to figure out how to read this shit because that's a that's a thing onto its own. Yeah. That takes a little bit of time. Is there that I actually haven't thought much about. Do do you know of any, like, good resources on, like, how I I I've read this from, like, my undergrad and grad. I'm just kinda used to it, but I don't know if there's any resources. Learning. Right? Like, you just you learn it in in university and especially in grad school, like, how to read a paper. Like, because you you just read so many of them. And and and sometimes your your PI or your your adviser gives you some tips and helps you out and teaches you stuff, and sometimes they don't. And, like, it's it's 1 of those like they say in golf, like, you dig it out of the dirt. Like, this is this is digging it out of the dirt. Right? I think another thing related to what you're saying about seeing stuff that you're familiar with after a while is, I think another good sign is when you read something and something clicks and you're like, oh, I completely misunderstood that paper I read 5 papers ago. Let me go back to that now. And you read it and you're like, oh, I understand what's going on here now. And I think that's a that's a good sign that you're making progress, I think. Like, if you never have to go back and reunderstand something you thought you knew, I don't know that you're making too much progress. Right. Right. And 1 last question on the research. I I know that we're spending a bit of time here, but this effectively serves the foundation for everything else. Mhmm. How do you think about the relevance of things that are older versus newer, that kind of relationship? Like, how do you bake that into your, context as you're reading? Yeah. I think yeah. You have to start with the, I think there's a there's a Linde effect. Right? Like, the stuff the papers that that are 20 years old that people are still talking about, there's probably a reason, and so you should probably know that. Right? Does that mean that the specific strategies or the specific things that were studied in that paper still work if they ever did? Probably not, but it kinda doesn't matter. The point is, like, you're sort of, understanding the landscape a little bit. You're building the map of, like, where the where the cities are. Right? And and probably you're not gonna find big game in the cities anymore, but it doesn't matter. Right? Like, yeah, at least nobody know where where the cities are and and where to, I don't know, where to get your provisions for for when you go out hunting big game. Right? Maybe that's, like, an analogy that I just came up with. I don't know if it's any good, but it's kind of a thing that I think about. But, yeah, like and then you look at newer stuff, and you're like, oh, yeah. This is, like, pretty new, and I haven't seen this before. Okay. Like, that might be something interesting. There's definitely an effect where if there's a good paper that comes out, like, it probably did make money, and then, like, it immediately stops making money. Like, definitely, there's an effect there. And so maybe being aware of that isn't terrible. Right? Definitely. And, another thing I would add that's kinda useful with stuff like this is you can download a ton of these papers, throw them in a Gemini or something, and have it batch the groups by decade and then summarize the findings from the papers. And it's a way to, like, just quickly look at trends. Is something that somebody talked about in the 19 twenties even remotely relevant right now. And the funny thing is sometimes, most of the time, you'll find that there's, like, small artifacts but largely go away. But then every once in a while, you'll find persistent things that were literally talking about the same ship from 100 years ago. And then you kinda have to ask yourself, why? Why? Why? Why?

Is this thing still here in some way? There's normally a reason for that. So anything else that you wanna talk segue to the sector rotation thing. Literally perfect. I was gonna say, is there anything else that you wanna hit on papers before we get into, like, reviewing a quick outline of stuff? Let's look at it. Let let's let's dive in. Beautiful. Alright. We're gonna take a 2 2nd break. We'll be right back. Alright. We back. Okay. So what I'm gonna do is share my screen. And the 1st thing that I wanna do is level set a little bit more. So this is just looking at spy. I'm just gonna zoom it out really long term. And the main thing is I find when trying to think about in effect or something, it can be useful just to, like, visualize the thing and try to section it up into different pieces. So if the premise here is that there's momentum and we think of momentum as the propensity for something to continue doing what it was doing for 1000000 different reasons why it can be the case Mhmm. Then I simply think looking at something can be useful. Right now with what I'm looking at in S and P, I'm not actually trying to necessarily derive anything specific. I'm literally just looking and seeing, does this even seem to remotely hold true? And a good example of that is if I look at S and P and then we instead look at all. Well, if I look at VOL, I would have a very different perspective in terms of momentum and what that might look like in something like this compared to this. This is really the main point of me mentioning this again is if, like, you're new and you don't know what you don't know. This is just a way to get some sort of visual input on what this stuff looks like. That's general orientation on eyeballing. Don't know how you feel about that. Actually interested in your thoughts, and then I'll get to the strategy outline. Yeah. Certainly, in the case of Vault, like, we all know that it's sort of mean reverts to some long term average or something, whatever that means. In the case of, like, Delta 1 stuff, it's not obvious that it does anything. Like, other than obviously, the S and P 500 has gone up for the last 80 years in general. Is that a fact about the S and P 500, or is that a fact about America? Like, I don't know. But, yeah, it's like yeah. Vol is different from Delta 1, but I think I think that's pretty clear. Yep. Exactly. And the the main point there is, again, from I try to at least for this, I'm trying to think of it from, like, a beginner lens that like, they might say, well, what about, like, oil? What about corn? What about wheat? Right? Just to understand how something like Momentum might look in different kind of assets. I actually have found that kind of mapping process even at an intuitive level can just slowly build context and help generate ideas. So for example, if I came across a paper that's talking about momentum in volatility products, depending on, to your point earlier, who's writing it, I'm either gonna think maybe I've missed something huge and there could be something really cool in here, or I'm just gonna dismiss it and say, yeah. I don't really care about this that much. Yeah. Okay. Certainly, null the null hypothesis in terms of delta 1 is always that, like, there's nothing. Right? Like, that's that much is clear. I think it's yeah. I think that's a a a good way to frame it, and I think it goes in that kinda reemphasizes what we were talking about before as well with, like, backtest in general, not necessarily proving anything. Right? Backtest just being in a a way to rough out an idea that might be worthwhile. Okay. So for the S and P 500 sector ETF rotation strategy, the idea behind this is taking the momentum factor and then looking inside the S and P and the various sectors and determining if that holds true and if there's any opportunity in here. And the thought process being, if we look at the market broadly now, we can see, you know, the market's been rolling off. It kinda flattened out for a while, yada yada yada. But if we take a look at the way that the sectors have performed, it's not uniform. And we can see clear artifacts of something like energy has actually been doing really, really well for the last several weeks. I mean, almost year to date, essentially. And conversely, if I look at other stuff like financials year to date, they're doing absolutely abysmally. So the idea is the S and P 500 summarizes, obviously, returns from all of these. But the question becomes, do we think if we gain exposure to these smaller subset of sectors that are doing well for upside and doing poor for trading to the downside, which isn't covered in this strategy at all right now, this is just long only, That might provide an opportunity. So that's the purpose behind applying momentum to something like a sector rotation strategy. K. This sounds good. What we do in the beginning is outline why we think this thing works, so what the profit mechanism is, and then defining the stuff that's gonna be eligible for this instance. Now the trading universe can be really complex depending on what kind of strategy you're running. Again, as I said earlier, I picked something really simple for this simply for a exercise for us to work through. So I picked sectors because they're easy to get. You can get sector ETF minute level data for free from all different kinds of places, and that's exactly why I picked these. The more detailed you choose to get, let's say, for example, I go through this and I was like, well, you know, maybe I can add options to this. Maybe. But as soon as I seek to add options to this, I have just massively complicated quite literally everything. So it's something that you have to do over time if you wanna use stuff like derivatives. But, again, the purpose of this is to give an idea of what this process looks like and, you know, go back and forth. So that's how I defined the trading universe. It's by ease of expressing the effect based on what I can observe from all the different kinds of research. I think that that effect lives in the sectors and then also within the sectors within the industries and stuff. And for the sake of the exercise, I picked something that was easy enough for me to rough out and look at before adding a ton of complexity. A question for you. How do you strike that balance between an idea? Because some of them, you might immediately wanna bolt on a lot of, you know, curious ideas that might meaningfully help things. They might drastically overfit, but you want you might wanna do that. But, obviously, it adds a lot of complexity. Yeah. So I think I think of it in terms of baselines. It's a very sort of machine learning way of thinking about it. But, like, let's just say that my baseline is buy and hold the S and P 500. Like like, let's just say. Right? So I'm gonna do, like, what is the next simplest thing that I can do? And then I'm gonna compare that to buying and holding the S and P 500. So in this case, I'm sure you're gonna go through exactly the details of it, but, like, it's something fairly simple with very liquid tradable things like the XLs. So, like, I'm gonna I'm gonna define some strategy that that gets long, some XLs some of the time, and then sells them back out at some point, etcetera, etcetera. I'm gonna run that over some time period, which we should probably talk about as well and say, like, okay. Does this beat the S and P 500? If it doesn't, then, like, I should just put my money in VTI and chill. Right? Like, that's what I should do. Okay. So it outperforms S and P 500. Okay. Great. After cost, after my time, all of this. Okay. Well, what's the next thing that I want to add to this, and does that do better than my new baseline? Modulo, like, overfitting risk, all this stuff. Right? So I I like, I think you wanna go through it stage wise as opposed to throwing the kitchen sink at it from the very beginning. I love that. And you mentioned 1 thing that I wanna talk about for a 2nd. You mentioned, you know, after cost, and then the 1 that caught my ear was my time. 1 of my 1 of my underlying premises is that, like, new traders, your time is effectively not worth anything. You're learning something brand new. You don't get to say, like, is this worth my time? If you feel like you wanna say, is this worth my time? Then I think that's a pretty good indication that, like, trading probably isn't the place. Like, your time might actually be too valuable for doing this digging because that's what most of this is gonna be. A lot of digging and a lot of finding absolutely nothing. So Yeah. And I I would I would even, like, I would put a finer point on it. Like, I would expect that most of the people who are coming into this new have a day job. Right? Like, this isn't the thing that's putting their food on the table given that it doesn't put any food on anybody's table yet. Right? And so if you're considering doing this, like, the probability of this end up ending up working out for you is very low. And so, fundamentally, this has to be about

You enjoying this. Like, this has to be fun for you. It really does have to be fun. Like, yeah, I can't stress that enough because, like you said, your time spent doing this is worth 0. Just like this the time that I spend playing tennis is worth 0. Nobody's paying me to play tennis. I just do it because I enjoy it. It's fun. And so that should be the basis of this. And if you're not finding it fun, if you're not finding it, like, worth the interest of, like, digging into these random weird things and, like, trying to understand it, then, like, I 100% agree with you, Eric. Like, it's probably not for you. And that's okay. There's nothing wrong with it. Yeah. Like, I think it's an important thing to say that that that you just did. Beautiful. We move on. After summarizing some of the concepts of the paper, I'm just gonna overview them just because I I know that I I don't wanna go too over time. I can double click into anything that you're interested in. But the may there's a a series of big questions that need to be answered. Things like how do I pick the things to go long? Do I maybe wanna go long short? How do I figure out, you know, what's stronger than other things for things like cross sectional momentum where I'm looking across different kinds of things and blending that also with time series? I also need to figure out how long I'm gonna hold these things before I make some sort of adjustment. I also need to figure out if there's any specific criteria that would lead me to bypass the base rules? So let's say we'll talk about it in a 2nd, but let's say that the base rules are to buy and hold the top 3. Is there a scenario where if the 3 top performing, but maybe they're not even positive in the window that I'm looking at? So even though they're the top 3 performing ones, do I still buy those? Mhmm. So what I'm emphasizing here is after defining the trading universe, these aren't a bunch of questions that you're just gonna magically know. There's stuff that you're actually gonna kinda pick up as you're going through the research. You're gonna see like, oh, okay. They used a 12 month look back minus the most recent month. Why do they do that? And stuff like that. So I'm just planting that seed that can be working in people's brains as we kinda move through. So the overall methodology that I outlined is is kind of, like, pretty straightforward. We'll just I'll go back up to this in a 2nd, but this is kind of the the meat and potatoes of it. What's gonna happen is on a regular cadence, which I'm gonna talk about when that is, there will be a ranking sort of S and P 500 sectors using different look back periods. How do I pick the look back periods is something that we're gonna talk about via testing. But I know that in order to rank and sort these, I need to have some idea of what's doing well compared to others and contextualizing that. Common waiting functions at a high level for momentum, 12 month and 6 month are typically stronger for continuation. 1 month and 3 month are typically stronger for reversal. That does not mean and, again, this is all just stuff from general research. Do not take anything that I'm saying is law. I'm just trying to plant ideas. 1 of the things that I think people mess up with momentum is they'll say, like, okay. 12 month is the strongest. I'm only gonna use 12 month. That's, in my opinion, a mistake in something that I started in this 1 to show you what that can look like. But what you'll typically find is ranking across multiple look back periods and giving something a composite score tends to work a little bit better. So what that means is I might look across 12 months, 6 months, 3 months, 1 month, and I might apply a certain waiting understand what you mean by that. You mean, like, from now to 3 months ago, from now to 6 months ago, from now to 12 months ago. Perfect. Right? And so what you're basically doing there is, like, triple counting the the return from now to 3 months ago, double counting the return from 3 months to 6 months. So you're basically constructing, like, some sort of, like, exponential waiting through this sort of composite score is basically what you're doing, which makes sense. Like, it's, like, a very sensible thing to do. Yep. That's exactly right. So the the stuff that I think we should talk about are, 1, after outlining the the broad idea, there were those series of questions that we need to look at in the beginning. And some of the questions will look like, well, do I wanna use a 6 month look back or a 12 month look back? Again, I gave a little bit of a cheat sheet on how some of these might look, but if you're going into this new, you won't know any of that. So you have to figure that stuff out. The other questions you have to ask yourself are, like, how many of these things am I going to hold at a given point in time? Am I gonna buy the top 1, the top 2, the top 3, the top 4? And the way that I answer that is essentially by locking 1 variable and then tweaking the others. And I iterate through all of these different conditions so that I can answer a handful of questions. How many of them are am I going to hold? For how long? With what look back period? I effectively wanna figure out which of those groupings and combinations tends to hold well over the entire duration and then segmented durations and then forward looked durations kind of to get both in sample and out of sample for the thing. I wanna emphasize a key thing here. I'm not inherently looking for the quote, unquote best return. That's actually very frequently a trap. 1 of the things that is important to prioritize is what is the most robust over these different durations. And what's most robust very frequently won't actually be the top line best return. So And maybe maybe, like, a a more like, a machine learning way link machine learning y way to say that is that 1 that looks the best is probably overfit a little bit to the data that you happen to have. And so looking at measures of like, some robustness measures, a way to sort of protect against your protect yourself against this kind of inherent property of looking at past data, which is the tendency to overfit. Definitely. And you can use things like random forest information ratios, and a really useful tool that I was referring to before is actually segmenting the returns, especially with momentum. Because you can have stuff like, for example, this little red line that looks incredibly good. This is a 12 month look back holding the top 2, except it requires absolute positive momentum. So what that means is as I go through my rebalance period, which for this is at the end of the month, which is kind of picked for a reason, as this rebalances at the end of the month, this looks good. But if you start changing some of the small variables so let's say, for example, instead of the end of the month, it's the 3rd week or the 2nd week or the 1st week. I wanna observe how this changes, which for this, it actually doesn't change a whole lot. But by looking at these different variables, segmenting them, and then observing how they change as you make small adjustments can be really informative, which there's a big difference, though, between a, quote, unquote, small adjustment of looking at the holding the top 2 versus the top 3, which I went through pretty much all the iterations out anyways, but I just picked a grouping here. If we think of this logically, it's for a pretty good reason. All that's happening here is concentration risk is going up. Looking back in time, this is saying that, yeah, like, the technology sector clearly did pretty well for some time. This perpetually has you in pieces of the technology sector, which is going to do this. That does not mean that holds true going forward, and it does not mean that if you just hold only tech that you're then gonna be better off because now you're fighting a series of metrics, which I was gonna pivot to next. So I think couple of things that I think that that are worth emphasizing in what you just said. Like, the, robustness to things that shouldn't matter is it is really, really important, like you said. Like, if it's the end of the month and it looks great, and if I if I scoot it by, like, 3 days and it looks bad, that's like and we're done. Like, that it's just it's not a thing that's robust. It's not there. It's just an artifact of the data. I think another thing with respect to tech, I think, is a really good point. Like, you're looking at data since 2000 or something. I can't quite read what is what's the It's 2000. Yeah. It's it's effectively from the launch of all the ETFs, so it's a life off of that. Like, from the start of 2000 to now, like, tech's been on, like, a world historic run. Right? Just like just as I'm sure I don't know. Like, electricity companies were on a world historic run-in, the forties and fifties. Right? Whatever. Right? There's always gonna be some sector that's doing really

Really well. And so, again, if we think that the strategy itself has some edge as opposed to it's picking up some secular thing that doesn't, replicate going forward, maybe the thing you wanna do is you wanna remove the tech sector entirely from all of your data. And, like, does this effect still hold? It might not hold quite as well, but it should still hold. Like, there should still be something there. And if it if there's not, then again, you're like, well, okay. I get it. Like, this thing is just a very complicated way of getting long apple. Great. Right? Like, I could've done that. PA. Right? Completely agree. Yeah. I completely agree. And when looking across some of these different metrics, there's a couple traps, and I actually think 1 of them is, an institutional trap compared to, like, a retail trap. 1 of them is Sharp because I think I think sharp is useful. It's contextual. But from a retail trading lens, I I'm not evaluated by my sharp. I can have an awful sharp on something that's very profitable. And at the end of the day, like, what I care about as a retail trader is actual money in my pocket. If it looks really rough to get there, then they might look really rough to get there, but you have to take that in the context of what it is. If you have something with a terrible sharp, you're probably doing something that's risky. And you have to The other thing about, like like, Sharp, like, there's a gazillion ways of sort of misunderstanding Sharp. I think in the case of strategies like this, especially if they have, like, a, like, a filter of some type that says, like, you shouldn't do anything, is, like, if the strategy is only in the market, it's only, like, has a has a risk on whatever, 10% of the time or something during the year, then you could easily, like, print money during that 10% and be out of the market 90% of the time. And, like, on paper, that thing's sharp is bad. Right? But it's great. Like, it's a great strategy. Right? It's incredibly capital efficient. Like, I can use the capital that strategies strategies not touching the other 90% of the time to do something else. Like, it's a great component of another thing, but, like, a headline sharp number might tell you, oh, this is not worth doing. Right? And so you have to really think very carefully about what it is that Sharp is pointing at and making sure that it makes sense for the thing you're looking at. Great point. And speaking of, like, different kinds of volatility metrics, obviously, there's Sharps or Tino, Kalmar. Like, there's all these different ones. Mhmm. How do you think about, you know, using different versions of them and, like, are there some that you use for certain circumstances and not others? Yeah. I think I would say, especially for retail, I think we talked about this in the last time we talked, the biggest mistake retail makes is they trade too big. And so thinking about drawdowns is actually really important for retail. Like, if you're this incredibly highly capitalized entity, I mean, you should still think about drawdowns anyway, but it's not existentially bad if if, like, if it's not the primary thing you're thinking about. But I think for retail, the number 1 thing you should think about is how do I avoid, like, losing money. Right? And in particular, how do I avoid losing too much money? Because if I lose too much money, I'm done. I'm out of the game. My wife is like, get out of there. This is wasting our savings. You know what mean? Like, it's just you can't live in a world where where, like, you are exposed to very large drawdowns. And so even if a strategy is really good, lord knows there's tons of strategies that involve selling options that look like making money for a lot of the time and then, like, bam, like, bulldozer. Hell, yeah. Right? And so those are good strategies for a certain kind of participant, but probably not for retail. So just like, that's another thing I would say I would emphasize is thinking about drawdowns and making sure that they are whatever historical drawdown you see, like, tax them more onto it and then think about whether losing like, just really put yourself in that position emotionally, like, financially, etcetera. Like, okay. I just lost 20% of my portfolio in the last 2 days. What am I gonna do? Am I gonna ride it out and, like, hope for the best? Am I going to liquidate and go home and and go play tee ball with my kid? Like, what are you gonna do? Right? Because I think this is critically important. And 1 thing just to emphasize there a little bit more is I I do find a lot of people, they just kinda brush over what they will do. Right? They'll think of what their idealistic self would do. They'll say like, oh, yeah. I I would ride it out. So, like, what you said there, I cannot emphasize that enough. Like, literally close your eyes and picture opening your account and seeing the numbers. Right? Like, really go through the exercise seriously. Because it's so easy to brush it off and be like, oh, yeah. You know, I'll I'll do the right thing, or I'll I'll just wait it out. And you'll be introduced to a very different version of yourself that you've never met before. Look. It's it's the Mike Tyson quote. Right? Everybody's got a plan till they get punched in the face. Like, that's, like, that's everything in trading. Yeah. I I like that. It's a really good analogy just for life, really. So continuing long, I do wanna be sensitive of time. There's a couple, like, key I've got I've got a hard stop in, like, 10 minutes, if that's alright. Perfect. Perfect. Okay. So let's talk then quickly about compound annual growth rates and some of these other metrics. Mhmm. If you'll notice using, you know, 6 month look back, 12 month look back, there's a clear trend. We're holding the top 2 has a better compound annual growth rate than holding the top 3. We kind of discussed logically why that might be the case. There's pros and cons depending on how concentrated you wanna be. It's something to consider. The main thing that I want to pick your brain on a little bit is if we compare some of these compound annual growth rates against the spy compound annual growth rate, they're not that different. And they actually only are quite different if I'm holding the top 2 versus the top 3. And fun fact, I hold it little bit so I can see the numbers a little bit better. Yeah. Yeah. Yeah. Sorry. Sorry. Sorry. There we go. So the the top 2, the compound annual growth rate is, 10%, and the top 3 is 8%. It's almost directly in line with S and P. Yep. 1 of the things that I think is worth discussing is how do you contextualize the compound annual growth rate? Right? This isn't Lambo money tomorrow against something like the reduction in the max drawdown. Right. So so you could think of it a couple ways. I think probably the the simplest way, I'm not saying it's the best way, but certainly the simplest way is, like, let's renormalize these things. Assume I have, like let's just assume I have free leverage. Right? Let's renormalize this so that we match the drawdowns. Right? Like, if I were okay with the drawdowns and I'm okay within the spy, what leverage would that imply in the strategy? And then I look at the growth rate and then, you know, hopefully, it's better. So you can see quite clearly. It's actually kind of interesting. Right? Like, the the growth rate like, the the compound growth rates are kind of pretty comparable. The sharps are actually pretty comparable as well. Like, maybe just a hair better most of the time, but nothing really to write home about. But it does a lot that are on the drawdowns. So much so that that, like, a 50% spy max drawdown, like, when did that happen? When did when did spy draw down 50% in the last 20 years? I don't think it did. Unless you're talking about 2008, but even then, is it 50%? I I think it was, like, '47 or something. Okay. Alright. So let's assume that that's a real number. I I would I think that's probably a little pessimistic for Spy, but whatever. Let's just say that it that's fine. I think, like, you're looking at like, you can kinda leverage this thing almost 2 to 1, right, and to keep the same drawdown as spy. And so now you're looking at a, like, double the, like, returns. Right? So that's, like, maybe 1 simple way to think about it. Yeah. It looks like in 0 0 9, we went from 2,500 down to 1000 on the total return index. 900 and it's 7 I remember clearly. Like, I remember, like, it was yesterday. We we hit, like, something like 790 sometime in October 2008. Yeah. Know what I'm thinking of. I remember Yeah. That The the cumulative drawdown in 2000 was 47.41%, and then in 2007 was 55.25. Okay. So that's probably that's probably it. Right? So yeah. So I that that's maybe 1 way to think about it. And there's 1 last thing that I wanna cover quick then before we but, like, these are just different heat maps to help align looking at different look back periods, different holdings, and stuff to try to get a sense of, you know, what what does what. Now the other question I had is when you're looking at something like this and it becomes, like, really, really apparent that in some instances being in the, quote, unquote, correct thing, the correct sector obviously, becomes the underpinning of the strategy. If we think of it from a logical perspective, sector rotation, like, that's kind of what you're betting on. Sure. As you decrease, like we talked about before, the number of holdings and you become more and more concentrated. Yep. How do you think about striking a balance between being concentrated enough to do well

But not being so concentrated that unless you're perfect, you lose. Yeah. I guess it kind of part of it, I think, comes back to, like, what is the source of your edge? Like, you always need to come back to thinking about this. Like, what is the what is the inefficiency you're figuring out and exploiting? Because that's sort of gonna guide you. Like, if I'm gonna I'm gonna say something fairly, like, facile. If the inefficiency you're figuring you're able to monetize is the fact that you figured out tech was gonna be awesome for the next 20 years and nobody else did. Okay. Great. That like, that's the thing you could tell yourself, but that's not what this strategy is doing. Right? And so, again, like, if it's super, super dependent on tech or whatever, some something that might not look like a thing that this kind of strategy should be exploiting, then then that's probably a pretty good signal that you shouldn't be that concentrated because you're sort of fooling yourself a little bit about how much edge there is. And so, again, like, a lot of those robustness kind of things will will sort of give you guides. Right? Like, if the top n looks really, really good and the top n plus 1 looks terrible, that's not great. Right? Like, if the top like, it should be kinda smooth. It should always be kinda smooth. Right? If there's, like, a big jump big, like, discontinuity that's, like, telling you there's something here that that's not necessarily replicable. Like, this should be kinda smooth as a function of these hyperparameters of your strategy. And and if it is, then, like, it's an easier question to answer. Right? Like, well, how much, like, how much risk? Like, it's probably, like if if you're going for a wider universe, then that's probably gonna be a little bit less return, but probably a little bit, like, more consistent return, hopefully, right, ideally. And so you can sort of calibrate that. You can kinda think about the risk versus return efficient frontier that you wanna live on. Right? Mhmm. So that's maybe a couple things that that I'm thinking about. Yeah. Yeah. I well, I think they're useful. It's it's like, is anybody, especially from the newer lens that's going through this that you could tell, it's like, there's a ton of shit that you have to think about, and there's a lot of things that you have to consider in order to have something that you're comfortable with. And I know that you have to run from a strategy perspective. Like, as you go through just even this as a general framework, if you were to give some thoughts to, like, a newer trader on if they were to go out and try to build their own strategy, Mhmm. Just some general ideas that you would offer to them. Yeah. I I 1 thing we we didn't talk about, which we could do a whole another show on, is, like, data. Right? Where does the data come from? How much do you trust it? Like, what are the pitfalls of using historical data? What are, like, the survivorship survivorship biases of data, like, corporate actions, dividends, special dividends, spin offs? Like, it's just like a zillion things that can happen to equities that your dataset might be representing properly and might not be. And, unfortunately, that's a very bitter bitter personal experience kind of thing where you get burned on it, you get burned on something else, and you get burned on something else. And, eventually, you just have a list of things you could just gonna make sure you're not gonna get burned on again. And, hopefully, that the list of things you do get burned on gets smaller and smaller over time. That's, like that's your best case scenario. But I think it's a very important thing to think about. When you're starting out, like, you're gonna get a dataset. You're gonna look at it. You're gonna be like, oh, this is awesome. And then you're gonna realize, like, oh, this dataset has some survivorship bias or some look ahead bias or something. So, like, don't go trading the 1st thing that you think is good or do it, but just do it small because there's gonna be all sorts of ways that you're not gonna realize you're getting bitten. I love that. That actually would be really cool to do as a follow on session just kind of like, you know, data analysis concepts or something like that for trading because you're right. That is it's that's its own bucket and and then some. So, Augustin, thank you so much for coming to hang out, even poking through at a super high level, you know, of a really basic boring strategy that doesn't really do anything that cool. The the main thing that I think we both were hoping to at least achieve is is giving people a look at some of the things that you have to do when you're going through this. And I think we're both pretty confident in saying, like, this is the very tip of the iceberg. This is like some super, super rudimentary, just basic stuff to help get the ideas going. So thank you for your time, and thank you for hanging out. I enjoyed it. I enjoyed this stuff. I I would even do oh, wait a 2nd. I just did. Yeah. And for folks that wanna get a hold of you, you're on x anywhere else that you wanna direct me towards? That's pretty good. Yeah. Just hit me up on a DM if you wanna talk. I'm always happy to chat. Beautiful. Alright, Augustine. Thank you, sir. Enjoy the rest of you all your evening, and we'll see everybody later. Take care.