From Hurricanes to Hedge Strategies: The Hidden Common Factors ft. Richard Brennan

[00:00:00] intro

You’re about to join Niels Kaastrup-Larsen on a raw and honest journey into the world of systematic investing and learn about the most dependable and consistent yet often overlooked investment strategy. Welcome to the Systematic Investor series.

[00:00:23] Niels

Welcome and welcome back to this week’s edition of the Systematic investment series with Richard Brennan and I, Niels Kaastrup-Larsen, where each week we take the pulse of the global market through the lens of a rules-based investor. Rich, it is wonderful to be back with you this week. How are you doing down under? Any big news in your part of the world?

[00:00:40] Richard

Well, hi, Niels. Great to be back.

And look, not really from our side of the world, apart from the fact we’ve been glued to the television sets looking at the carnage unfold in the Gulf of Mexico with the two hurricanes they’ve recently had. We call them cyclones down here, but they’ve had hurricanes up there, and that’s probably going to feed into the topics we talk about today. But my heart goes out to all of those that were affected by those significant events, both Helene and Milton. We still have got to sort of uncover what the full effect was.

But my wife and I have been sitting here glued to the TV thinking about our friends over there and hoping that they’re all safe. But yes, it just goes to show, in this world, expect the unpredicted.

But fortunately, as we’ll be talking later in our discussions, these models, these ensemble models that they use for forecasting techniques are certainly better than what they used to be with the old methods of determining outcomes of weather events. And hopefully that’s going to be an interesting topic today.

[00:01:53] Niels

Yes, indeed. I mean, I know you kind of already jumped into my next question, which is what’s been on your “radar” (pardon the pun) for this last few weeks?

And I agree. That would have been my topic as well. Clearly, my colleagues are in Florida, in Stuart, luckily, just a tiny bit south of where the path went through. And so, they’re not affected by it as such, other than I’m sure it’s been pretty windy and wet.

But I am quite impressed, generally speaking, when I visit Florida, which I will again in a couple of weeks, how calm they are about these events. And even when you look, you mentioned you’ve been watching television. I mean, even the people where you see their house is pretty destroyed and so on and so forth, and they’re preparing for the next storm to come, I still find them to be quite calm in a weird way.

[00:02:49] Richard

Yeah, that’s right.

[00:02:50] Niels

Yeah.

[00:02:51] Richard

I suppose they’ve lived with these events, over the course of their lifetimes, but they certainly seem to be increasing in intensity. So, these last two, who would have expected two on the tail like that, you know, two weeks apart or whatever. It’s quite amazing how these events occurred.

[00:03:12] Niels

I’m not an expert in these things, although I do seem to recall my colleague saying that actually the last couple of years have been relatively light on hurricanes or the hurricane season. So maybe again we’re just going back to the law of averages and then you get a couple of years or maybe one big year like this year so far. Hopefully it’ll end now for everyone’s sake.

Now speaking of hurricanes and unsettledness, I guess we should talk a little bit about what’s going on in the trend following world. It’s not been calm in that part of the world either.

So far October, we’re into October now, has been pretty rough I would say, or tough for trend following strategies no doubt (Certainly from my point of view I’d love to hear your point of view in a few minutes). Fixed income has been the most difficult sector.

I think that trend followers were just settling into some long exposure over the summer after a few years of riding the short side of that market. And as it often happens, once you get that push into a new emerging trend, the markets decide to provide a correction of meaningful size. And I think this is what we’re seeing the past few weeks.

It’s probably too early to say anything about this being a kind of a reversal of a bigger trend. I don’t know that we can say that. It’s certainly a meaningful correction and it’s inflicting some pain on trend followers as far as I can tell. Then we have currencies which some of them are also in a bit of a correction: British pounds, your own Aussie dollar, New Zealand dollar for that matter as well.

Energies are kind of mixed, probably not where the damage is too much so far, in terms of trend followers, simply because I don’t think signals are particularly strong in that sector at the moment. And also, grains are quite mixed at least the last week or so. Losing a bit of money in wheat maybe, and making a bit of money in soy complex, I would imagine. The bright spot of a fully diversified trend following portfolio seems to be meats, live cattle in particular. Unfortunately, that’s not enough quite to offset all the other stuff, but it’s there. And metals giving up I think a little bit of performance, not too much.

So, it looks to me like we’re in one of these periods where diversification doesn’t really help, and all the sectors are kind of ganging up on trend followers to make life a little bit difficult. But I love to hear what, since you trade differently than the strategies I follow closely, what it’s been like from your perspective.

[00:06:17] Richard

Well, I agree, it’s been a very difficult month’s start. Last month we managed to eke out a fairly good gain. This month, immediately from the commencement, it’s been a bit of pain for us, predominantly. Energies actually hit us. Everything, including all of the crude oil, Brent, all the energies, their trends reversed.

The coffee trade that we were having a very good time on was quite painful for us. But you’re right, despite the levels of diversification, you often find that subtly hidden beneath all of that matrix of that very large diversification is hidden correlations. And clearly everything sort of is reversing on us at the moment. There isn’t much joy on our scope.

Luckily, the last day we recovered a bit of the grief, but it’s quite a challenging start. It’s been quite a challenging second half to this year. As we were talking before this show, Niels, I think dispersion is going to be one of the major topics that come out over the course of this year with the trend following community.

[00:07:32] Niels

Yeah, no, absolutely. Of course, people can follow that dispersion as we publish our monthly updates. It gives a good hint on what’s going on.

My own trend barometer actually has been pretty good at, how should I say, not predicting because it doesn’t predict anything. But it’s certainly been very good at reflecting the challenges because it’s down to 20 as of yesterday, which is very low. And it went all the way down to 14 a couple of days ago, which is pretty much as bad as it gets. Despite the fact that, as I often caution, it does use a little bit of a shorter-term time horizon than traditional medium to long term trend followers. But still, it’s been pretty accurate in picking up the challenges. So, let’s wait and see when it starts to turn.

[00:08:19] Richard

I love your barometer, by the way, Niels, because last week when we did our state of trend report, when I looked at the barometer, it had there ‘as bad as it gets’.

[00:08:29] Niels

Yeah, that’s the little notion that we decided on many years ago. When it gets down to a certain level, we just put the little caption as bad as it gets. It’s still there. It’s still there.

Anyways, let’s talk a little bit about the performance. So, BTOP50, and this is as a Wednesday, I suspect yesterday was pretty flat, not very interesting yesterday, but BTOP50 as a Wednesday was down 2.84% in October, still up 1.23%. Unfortunately, it’s the only index that’s up because this SocGen CTA index is down 3.64% in October. Now it’s down 1.22% for the year. The SocGen Trend index down almost 5% so far in October, down 2.6% for the year, and the Short Term Traders index down one and a quarter ish for the month and down 75 basis points for the year.

Comparing that to the MSCI world, for example, that’s down 33 basis points in October, but it’s still up 17% and change in 2024. The S&P Global Developed Sovereign Bond index (that’s such a long name) is down 1.02% so far this month, up 1.38% so far this year. And, of course, the star of the show so far this year. The S&P 500 Total Return index up 30 basis points and up 21.18% so far this year. I’m not entirely sure whether it should be down 30 basis points this month. Anyways, it’s pretty flat, let’s call it that.

Any thoughts that provoke by having a positive year for CTA’s all the way through October and now we get into negative territory for the year?

[00:10:13] Richard

Well, the thing is, there is so much dispersion in our camp, I tend to think in periods like this, I don’t think we’re necessarily representative of the benchmark, and I certainly know you’re not representative of the benchmark in your performance this year without discussing numbers. But you know, when we do state these benchmark results, it’s got to be clear to people that not everyone is the same.

There is a fair bit of dispersion there, certainly enough to warrant further investigation into the group as a whole, because I think these indexes are dealing with the very large funds, a bit like the cap weighting we see in the S&P 500. It’s sort of looking like the cap weighting of the trend following index with the particular benchmarks.

But when you look at things like the Top Traders Unplugged index, which is a wider benchmark incorporating more bodies within it, we see that it’s not necessarily as gloomy and doomy as maybe the SG trend and SGCTA index. And of course, what that means to me is that whilst we can talk about the general sort of theme, in relation to these benchmarks, being a difficult year, etcetera, people need to do investigation and clearly see how much dispersion there actually is in the group before they make any choices.

[00:11:43] Niels

Yeah, I think that’s a fair point. Actually, you know where there’s also an interesting development going on this year. It’s actually in the replicator space. And of course we have Andrew as our flag bearer for that space, but his product has a massive tracking error this year. Luckily it’s to the upside for him, but I still think it’s tracking error, which is interesting to say the least.

Alright, let’s jump into the main show content. And the first one is a question. It’s a question that came in from Shahar. He or she? I’m not entirely sure, sorry, writes, “First, I’d like to thank you for an amazing podcast. I listen to it frequently and it helps me expand my understanding and knowledge about trend following. (That’s very kind. Thank you so much. Shahar.) I have a general, simple question, but one which I haven’t seen, at least to my knowledge, being discussed on the show. If you have addressed this issue before, accept my apology, and I would appreciate if you could reference me to the appropriate episode.

My question regards the topic of trend following broad market ETF’s like SPY, TLT, EEM, and using long-term long only trend following as an overlay. Very simply, to what Meb Faber outlined in his seminal white paper as a tactical asset allocation strategy, respectively.

What do you or your guests think about this approach? And do you believe that this is a compelling approach to limit the downside, although potentially sacrificing the upside?

Assuming one invests a large amount in one time and not with dollar cost averaging, would you consider this a valid ‘hedging simplistic approach’? For the sake of the conversation, I ignore the taxes issue.”

All right, now, I presented that question to you as our go-to professor on all things trend. So, I’m sure Shahar and myself and everyone else listening would love to hear what you think about this.

[00:13:52] Richard

Yes, well, it was a great question by Shahar. So, look, what I’ll do is I’ll provide a context for it first before I go into my opinion about it.

So, just to explain what Shahar was referring to, he mentioned using broad market ETF’s like the SPY, the TLT, and the EEM in a long only trend following approach. So, for those less familiar with the symbols, the SPY. The S P Y represents the SPDR S&P 500 ETF. So, that’s an ETF that’s representative of a broad slice of US companies, basically a measure of the largest and most influential companies in the US.

Then the TLT is the iShares 20+ year treasury bond ETF. So, this tracks long term US Treasury bonds, which tend to be viewed as a safe haven asset, especially during times of economic uncertainty. So that’s the TLT.

Then we have the EEM, which he mentions, which is the iShares MSCI Emerging Markets ETF. Now this covers emerging markets such as China, Brazil and India. So, these offer more growth potential, but often come with added volatility.

So, the first thing is by investing in these uncorrelated ETF’s, we naturally obtain benefits of diversification that help to smooth the portfolio returns. And having specific exposure across these three specific asset classes is similar to many strategies you might hear of in our trend following world, referred to as all-weather portfolios.

I know Eric Crittenden has got a long and short version of what you’re talking about, but because you’re investing across these three uncorrelated asset classes, where one zigs and the other tends to zag, they tend to offer this, what they call this all-weather portfolio protection.

So, the question then applies, now let’s apply a long term, long only trend following overlay to these ETF’s, which is similar to a strategy that was popularized by Meb Faber. He talks about it as a tactical asset allocation. So, it’s actually a paper he wrote called A Quantitative Approach to Tactical Asset Allocation.

So, in Meb’s research, what he did was he used a simple long-term moving average, often a ten-month simple moving average, to determine whether to be in or out of each of those three asset classes. So, he applied this, though across a broader range of asset classes. He applied it to things such as equities, bonds, real estate, commodities, currencies, all implemented through various EGFs. In Shahar’s reference, we’re referring to these three uncorrelated asset classes. Meb was probably speaking a bit broader in terms of exposure in different asset class areas.

But the idea of Meb’s was to stay invested when the price of an asset, like the SPY or the TLT, is above its moving average. But when it falls below, you then exit and move into cash or something less volatile like short-term bonds.

So, this is a long only strategy, meaning that you don’t short the market, or you don’t bet against assets when they decline. Instead, you simply step aside when trends go negative on you. So, the main goal here is downside protection. And the idea of this approach is to limit large drawdowns in your portfolio by exiting these assets during prolonged downtrends, which is why it’s often referred to as a tactical asset allocation strategy.

But as Shahar correctly points out, there is a cost to this of sacrificing some outside, particularly during sharp recoveries, sharp quick recoveries, because these strategies tend to lag in their reentry after a sudden market bounce. So, in this specific scenario Shahar’s talking about, he’s asking whether this strategy could serve as a compelling form of hedging, especially for a large one time investment and assuming no dollar cost averaging.

So, the idea is whether applying a trend following overlay could provide protection from downside risk without needing more complex hedging strategies. So, in this context, I’m going to start looking at the pros and cons from my opinion, in relation to this approach.

So, the potential benefits of trend following are the downside risk management. So, the primary strength of this long only trend following approach is its ability to sidestep these major drawdowns. So, by exiting positions during extended downtrends, you’re simply using a simple moving average.

It’s a very simple method, simple strategy to apply without having to get into the complex world of options and other sorts of tail risk protection strategies. But it only provides a degree of protection in a limited sense. For instance, if you wanted pure short-term protection, you would typically go into tail risk protection strategies like options, et cetera, which gives you immediate downside protection. This is relying on the prolonged downturn, something which trend followers are used to because it often takes a while for their models to adapt to the new regime.

If you’re imagining a long only strategy applied in this context, it takes a while to reach that stop and exit out of that position as it crosses that smooth, simple, moving average, and then you’re out of that particular trade. But by having those three asset classes together, it allows you to dynamically adjust.

For instance, out of the SPY, you might be still in the TLT, and you might still be in the EEM, or you might get out of the EEM, you might still be in the TLT, etcetera. So, it’s dynamic in that sense, and it’s adapting to which asset class is producing those long-term trends. But it isn’t a complete protection strategy in that sense of the word.

[00:20:18] Niels

Sahar used the word hedge, and he asked if it was a good hedging strategy. I think I wouldn’t in a sense, use the word ‘hedge’. I would just use the words ‘risk mitigation’, because I think that’s really what you’ve talked about. It has the ability to mitigate some risk. But I agree with you, it’s not really a first responder in a portfolio sense. There you have to look at other stuff. So that was just sort of my comment to…

[00:20:48] Richard

And if you think about it, as well, you can enter these awful periods of whipsaw risk. For instance, when you’re sort of going in and out of that simple moving average up and down, up and down, up and down. So, over the short term, this particular approach can offer whipsaw potential, which is another sort of downside or a con to this approach.

And I think the long only constraint is another con from my perspective, in that you can’t participate in the downside move. You’re out or into cash. And we do know that sometimes, with certain asset classes, there’s significant downside potential which you want to capitalize on.

And also, by having both a long and a short trend following strategy applied to your markets, you do get significant correlation benefits between having the two at the portfolio level, as opposed to simply a long bias on all of your strategies. So, there are those sort of considerations as well.

But yeah, it’s a good, valid strategy. Meb certainly gives a big thumbs up for it. So, I don’t see any real significant problem in it apart from some of the caveats we’ve mentioned.

[00:21:58] Niels

Yeah, yeah. It’s probably better than doing nothing in just long only and closing your eyes. The other thing I would just say, added to this, is, of course, when you talk about these sort of strategies in general, “Maybe I should do this. Maybe I should do that.” Very often we forget to mention that yeah, but you may not see the result for another 10, or 15, or 20 years.

So, this time horizon, in order to see the benefit of doing something like this, can be very, very long. And so, you really need to have a good starting age, so to speak, but also be very disciplined in the application of it.

And then finally, I would say to that, and this goes back to a lot of the conversation we’ve had over the years, where we talk about how we test things and so on and so forth, if you only look at three instruments like that, you may obviously be prone to the fact that they may have been in a specific regime.

We know for sure, with bonds, that even if you go back 40 years from today, most of that time, probably 37 of those years, mostly interest rates were coming down. So that strategy would have reacted in a certain way, whereas in the next 37 years, we have no idea if that’s the case. So, always take these things with a grain of salt.

And I guess in a sense, that we prefer (both of us) a bit more diversification in order to offset some of these challenges or risks that you’re focusing on, on a particular regime, for a specific asset class. I think we’ve done enough on the question, so I appreciate that. I’m sure Sahar does as well.

And then we’re really moving over to kind of the topics of the day. Now, some of them we’ve always already alluded to. That it is going to be about kind of complex systems, not just weather systems.

But before we get into that, we are going to stay on the topic of diversification a little bit and see if we can’t use Warren Buffett in the context of trend following. So, what are your ideas here?

[00:24:13] Richard

Okay, Niels. Warren Buffett is very well known to have said that, “diversification is protection against ignorance. It makes little sense if you know what you’re doing.” And as soon as trend followers hear this, their ears pick up and they say, hey, we diversify extensively. What are you talking about, Warren? But I think for an outlier hunter, it does make sense, and I want to explain why it makes sense.

For many investors that hear what Warren Buffett said, they typically think this is a rallying cry for focusing on concentrated positions in areas of deep knowledge rather than spreading investments too thin across a wide range of assets. At first glance, this philosophy seems directed at those who excel in traditional stock picking. But its deeper implications extend to the world of outlier hunting.

So, as outlier hunters, we operate in an environment where random price noise dominates the bulk of the market distribution of returns. And our goal is to capture those rare but massive price moves called outliers, that exist in the tails of the market distribution. So, to achieve this, we rely on trading uncorrelated trend following systems that remain inactive during periods of noise. But then when we start entering into these trends that materially evolve and endure, we progressively concentrate our capital into those trends as outliers emerge. And we’re doing this with an ensemble of trend following systems.

So, this is where Buffett’s philosophy intersects with our own, as we’re concentrating on specific market patterns. Warren Buffett is specifically focusing on stock picks, but we’re focusing on these things called outliers. When we see an outlier, we start concentrating our risk towards those material events.

So, while the outlier hunter doesn’t know where these outliers are going to be, of course it spreads the net very, very wide in its initial level of diversification. But rather than investing in everything across a diversified universe, think of it more as a watch list. In other words, we’re prepared to pounce on any opportunity that exists in this extensive, diversified watch list because we don’t know where these events will unfold.

We might put our finger in the water with very small bits of, with certain things as trends start developing, but we’re not going to be concentrating our capital until these material events unfold. So, a lot of the time, our portfolios are inactive, and so effectively, we’ve got this watch list we’re watching all the time with our systematic processes that are ready to pounce and waiting for these material trends to emerge.

And when these material trends emerge (and we’ll talk about this later), it’s a bit like looking for lows or depressions in the Gulf of Mexico, which are the kernels of potential, major, catastrophic weather events. We’re looking for the kernels in trend, which are the commencement of these material trends.

So, we don’t know whether these are actually going to become outliers. So, we therefore only invest, as a very small amount, with our uncorrelated systems. We’re applying multiple uncorrelated systems.

What I mean there is, we’re trying to avoid investing all of our capital at once into one trend that may or may not unfold into an outlier. So, we might deploy things like Donchian channel breakouts, Keltner channel breakouts, Bollinger Band breakouts, et cetera.

A range of different uncorrelated systems that we know, through our design, are uncorrelated, and they will not implement or activate at the same time. So, one finger goes in the water with, say, a Donchian breakout, a bit later, a Keltner channel breakout, if that trend persists, will activate. A bit later, if the trend still persists and endures, a third system will activate 4th, 5th, 6th. I’m applying, now, ensemble models, which were 10 looking at 20 now.

So, as you can imagine, if these trends become what I’m calling nonlinear in extent, so what I’m meaning is that when you’re out in the powers of the distribution, the magnitude of these moves can be many orders of magnitude greater than the normal machinations of the market you experience in the bulk of the distribution. Those moves tend to be what I call almost random perturbations, or mean reverting oscillations in the bulk of the distribution of returns.

But all of the liquid markets we trade, we plot the daily moves of the markets that we trade. You’ll see that they all follow what they call a leptokurtic signature with these big fat tails. And there’s a boundary where these tails start emerging from what I call the edges of the Gaussian distribution, edge of the normal distribution. From these edges, these thick tails emanate. This is when we start wanting to activate those systems.

So, we’re waiting, waiting, waiting for all of this noise to be over, all of this mean reversion to be over. When one of these signals arises in this extensive watch list we’re monitoring, one system comes on, two systems come on, three systems.

Trends might dissipate from that point in time. So, you’ve had three systems. You’re only going to get a three system bet loss. But as it progresses, four systems, five systems, six systems.

So, the results we’re getting, at the individual strategy level, are representative of the nonlinear results we get of the price move. And so, if you could imagine it, we’re getting nonlinearity with every single system we deployed.

So, if we’ve got 20 systems, that is some huge multiplication of the potential nonlinearity that exists in the price movement itself. So, this is why I refer to our approach as being nonlinearity squared. We are maximizing the potential of significantly boosting our equity with these outliers through these events.

So, someone like me, I’m a bit of a weird guy, I’m not the normal trend follower, I’m really focused on these tail properties. So, I go a bit excessive. I concentrate my risk towards these events, and this is where it intersects with Buffett’s idea. I think that what we’re looking at, instead of a knowledge of the individual stocks, I’m relying on my knowledge of how complex adaptive systems work and my knowledge of serial correlation, which is driving these nonlinear events through this amplification process, this positive feedback process that we’ll be discussing in the next topic where we talk about John Holland and complex adaptive systems.

So that’s basically how I see Warren Buffett’s comment. I actually think it is endorsing our approach. It’s not something to say, hey, he’s wrong. I actually think, in my opinion, and in my particular approach, he’s spot on. We don’t, for instance, diversify just for the sake of reducing risk. No, an outlier hunter is diversifying for the specific purpose of being able to focus its capital towards what we call the events that matter.

So, when we’re saying the events that matter, we think it’s this nonlinearity, that occurs out in this tail region, that are the things that really give us an edge in comparison to other methods. We think this is a thing that’s driving the performance of our returns. So that’s how I’m seeing it, Niels.

[00:32:33] Niels

Yeah, I appreciate that.

And I think, as you rightly say, I think that your philosophy is obviously incredibly aligned with trend following, but it’s also a little bit different than many of the larger trend followers as well. So, I have a couple of follow up questions before we jump into that.

So, you mentioned that we have ten different systems, I think, and this gives you a way to build up ‘concentration’ if they all activate in one market. I think just to maybe explain for people who may be new to the show, I think what many trend followers will do is not necessarily having ten different systems, but we’re going to have many, many different permutations of maybe a few models or systems, and that also kind of has somewhat of the same effect, but in a different way.

My question to you, though, is just purely sort of a curiosity. What’s the maximum concentration you can have in a single market if they all kick in?

Because I think, for me, this is one of the big, interesting points about you picking Warren Buffett as an example. I mean, Warren Buffett obviously, more recently had, what, 40% of his entire portfolio in one stock? Apple. Maybe it was even more.

I read the other day, I think it was just a headline, but I think the article came from Motley Fool, that Bill Ackman, another well-known billionaire in that world, he’s the founder of Pershing Square. The headline says he has 53% of his hedge funds, which is $10 billion, in just three stocks.

I have come across some information, public information, of course. It’s the story about a really profitable partnership, investment partnership. One of the partners was called Nick Sleep, and I think the partnership was called Nomad Investment Partners. It’s closed down now, but in that documentation of how they became so successful, they were also incredibly narrow in their selection and had huge amount of risk in a single or two stocks.

So, from your perspective, how much is enough? How much is too much when it comes to your trend following strategy, in terms of what limit or what maximum risk you can have in any one of the markets you trade?

[00:35:16] Richard

Okay, so I have done a fair bit of research over the last… We’ve got a research project on at the moment, which is why I’m considering lifting my ensembles from 10 to 20. Now, this is as opposed to being applied to the old CFD markets, Niels, this is now in the futures market.

So, what I’m finding now is that I’m capping out at 20 at the moment simply because it’s margin; the margin on these instruments and the limits of capital, which is restricting my ability to go further. So, if you could imagine with some of the markets that I trade, like palladium, et cetera, they’ve got excessive margins. And as you are concentrating risk on them, the margins get explosive, very large.

So, the reason I’m capping out at 20, my fundamental viewpoint is that, like market diversification, and because I’m focusing only on these outliers, it’s a different ball game to when you are focusing on different types of trend. So, if you can imagine, with the outliers that I’m targeting, correlations tend to go out the door.

Outliers naturally destroy correlations in markets, normally. They are these things that a lot of people refer to as anomalies. I’d argue that, well, actually they’re a natural part of the market that exists. But a lot of people referring to them as anomalies do so because they actually destroy and disrupt correlations that exist in the market.

So, if you could imagine when I’m trading these things, that I get better results – the more uncorrelated systems, the smaller bits and the more uncorrelated I have in the systems, riding these things called outliers. For instance, if I had one system riding that outlier, because these outliers are typically very chaotic in nature. They’re totally unpredictable. We don’t know if they’re going to have significant volatility explosions. We don’t know their endpoint. We’ve just got to ride these things. So, we need maximum freedom under a couple of caveats.

One caveat is we can’t afford to leave all our profit on the table. So, in other words, we must, with each of our uncorrelated systems, have trailing stops behind them at all times. So, as we progressively get more and more involved in the outlier, if it turns against us out of our favor, we can’t afford to leave too much profit on the table. So, the way we achieve this, what I call the loose pants methodology, is through an ensemble, as opposed to, say, a few systems.

With a few systems, you might have very wide stops or very wide trailing stops. But with an ensemble of 20 uncorrelated markets, the sum total of those being applied to the outlier actually does have very loose pants. In fact, much looser pants than maybe two systems with wide stops. So, you’ve got to look at it in the context of not only the individual trade in the system, but how they all marry together.

Now, what I’m finding in my research is that the more I diversify with my uncorrelated systems, the less (How do I put it?) hindsight bias is required. What I’m finding is that CAGR, on these systems, on this ensemble, does lift. Yes, immediately the lift is large, but as you go, and more and more diversification, yes, it does start tapering away, but there isn’t a point where it stops. It just continues to get less and less in its incremental addition to the overall result. But what I do find, that’s CAGR. What I’m finding is the drawdown, because of the uncorrelated ensemble, is always improving as I’m putting more and more uncorrelated systems into the enterprise.

So, if you could imagine, what I’m finding, therefore, is that if I plotted the number of systems I deploy against a graph with CAGR on the light left axis and drawdown on the right axis, I find that my drawdowns are continually improving and my CAGR is incrementally nudging up. It’s not going backwards.

So, this is why I’m saying, theoretically, I’d go further than 20 if I could. But what’s stopping me is the margin limitations on the portfolio.

[00:40:05] Niels

I think another thing that may be stopping you soon is I can’t think of even 20 different trend following approaches.

[00:40:13] Richard

Yeah, well, you are right in one sense. You know how you said a lot of people follow this in a similar way, but they might use different parameters as opposed to different systems.

So, we actually have eight core systems. So, they’re eight core systems, but the variations in that are via parameter variations as well, if that makes sense to you.

[00:40:35] Niels

Yeah, that makes a lot of sense, because I always say to people, there’s only, like, five or ten ways you can do trend following. And I’m thinking, “What? Did Rich discover another ten?” So, anyways, that’s a big relief.

All right, well, let’s move on to really something we’ve addressed before, but maybe we’re going to do it in a slightly different way today. But it is so important to understand this concept in order to understand why trend following is such a relevant strategy for investors, in my opinion. And there’s no one better than you, Professor Brennan, to dive into complex adaptive systems. So, over to you.

[00:41:18] Richard

Well, I think you’re right, Niels. I think that trend followers have this way about them in their philosophical approach, how this non predictive philosophical approach, which is basically reacting to market moves, it’s not predicting. It’s reacting to them. I think this was what sets us apart. So as trend followers, we’re not trying to predict what the next move is. We’re simply responding to what the market is telling us.

So, if you could imagine, if we go back to the indigenous tribes of America and Australia and all of these areas, we’re doing effectively what these tribes have known for centuries. And they know that by looking at the overall system, the relationships that exist in that system, by understanding how everything knits together as a coordinated, correlated whole, you’ll start realizing that this world and these financial markets, there’s no sort of anything that exists in isolation.

They’re all related to each other. They’re all correlated to each other. And this is where the philosophy of complex adaptive systems fits very snugly in terms of the way we identify with and manage risk in these markets.

So, I’ve recently had my head buried in a couple of books. One is a book by a gentleman by the name of Jeffrey west, and I really encourage your listeners to read it. It’s called Scale. It’s a magnificent book. I think Nassim Taleb did a big forward on it, but it’s a magnificent book looking at the scale invariance that occurs in systems and why it throws ideas at us that we don’t necessarily think about in the way we process information.

For instance, a good example is when you look at little ants and they’re lifting grains of rice or they’re lifting objects, we’re thinking, if only we scaled them up to our scale, how strong would they be? But the bottom line is that that will never occur because of the scale invariance.

And we see this scale invariance in how fractals form. And fractals are centerpieces of complex adaptive systems. So, you know, in our man-made world, we’re used to building things; constructing engines of cars, putting bits together, assembling pieces together. But that’s not how systems grow and unfold. Things grow from a seed, and then they don’t use things like what we use.

We use rulers, timepieces, to basically construct the objects that we design and build. But that’s not how nature does it, and it’s certainly not how markets work. They build from seeds or kernels. And then, in a scale invariant way, as they progressively evolve, they’re using simple processes, simple processes all the time that maximize energy efficiency, metabolism, anything you like. It’s maximizing their ability to survive in a very adaptable environment.

So, these fractal structures become incredibly important when understanding how these principles operate. And just to explain so you might be familiar with our circulatory system, which is a fractal design.

The trees and the venation in the trees and the forests are all fractal designs. The coastlines that surround all of our continents are all fractals in nature. And what this means is that if we zoom in with our microscopes and zoom out, we see this self-iterative recursive structure at all scales. In our man-made world of circles and squares, when we zoom in, we start losing the object nature of that circumference or the square. We start seeing flat Euclidean spaces.

If we look at a perfect circle and we slowly zoom in at a perfect circle, as you zoom in, you lose the curvature of that circle to eventually you’re looking at something that appears flat or Euclidean. Now this doesn’t occur in natural systems, and it doesn’t appear in complex adaptive systems because as soon as you zoom in, you get this same recursive structure unfolding at all different levels of scale.

And so, another book that I’ve read that’s fascinating is through John Holland’s work. And John Holland, he was a big writer and educator in complex adaptive systems who was a seminal figure at the Santa Fe Institute. And he died, unfortunately, in 2015. But he wrote a magnificent book called Looking Through the Lens of Complex Adaptive Systems: Uncovering Signals, Boundaries And Emergent Market Behavior. And at the core of these complex adaptive systems are these three processes, signals, boundaries, and emergent processes.

So, I’ll just step through signals, and what boundaries are, and then what emergent processes are. And then I’m going to look at how we can apply this to trend following.

So, signals are the pieces of information that agents. So if we can imagine, all complex adaptive systems comprise agents as their fundamental units. So, in our world of finance, the agents are the traders, or the investors, or the institutions, they’re all the entities that exist in that complex adaptive system.

Now, signals are the pieces of information that agents use to guide their decisions. And these signals can range from environmental cues. For instance, cells in the body are agents in the body, and they’re working on signals that they receive from the body.

Traders, investors are agents in the system, and trend followers are agents in the markets as well. And we’re relying on signals for executing our trades; entries and exits, we’re relying on signals.

So, a cornerstone to these complex adaptive systems is this presence of signals which agents use to guide their decision making abilities. So, John Holland explains, these signals are not just simple data points, but more importantly, they’re interpretations of that information by the agent.

So as humans, in these complex adaptive systems, we interpret these signals from the way we interpret them. We might have a different interpretation to what you might interpret them at, to another person might interpret that. That’s limited by our cone of understanding. So, it’s an interpretation of these signals.

And so, this therefore leads to a variety of different manifestations of how we respond to those signals. Because it’s an interpretation, it’s not the same. So once these signals come in, we’re not all doing the same thing when we’re getting these signals. So therefore, these interpretations of this information by agents of these signals, these trigger actions within the system.

So, in financial markets, signals include things such as price movements, technical indicators, economic news, shifts in market sentiment, and for trend followers, these signals are vital as they form the foundation of our entry and exit signals. For example, when a stock’s price crosses a moving average, it can be interpreted as a signal to buy or sell.

But Holland emphasizes that the strength and clarity of these signals can vary. And traders, therefore, must constantly interpret and adapt to the changing signals in order to thrive within the market ecosystem. The good thing about, well, the complex thing about these markets is they’re always evolving and adapting. Signals are always evolving and adapting.

It gets more complicated than this because now we’ve got to look at boundaries. What are boundaries and why are they important?

So, boundaries are the constraints that shape and limit the behavior of agents within a complex adaptive system. So, John Holland, he highlights the importance of boundaries because they define the scope of what agents can and cannot do.

So, this helps the system remain stable. So, in financial markets, boundaries take the form of, for instance, risk management rules, regulatory frameworks and system parameters. These things define what can and cannot happen. They’re actually providing a structural constraint or limiting the degree of freedom with how we respond to the signals that we’re receiving within these boundaries.

So, for instance, a trend following strategy might implement boundaries like position size limits, stop loss orders, or leverage caps. And these boundaries, they’re very important because they define the possible opportunities that exist in what I’m calling this ‘state space of opportunities’.

Now, what I’m saying is this is not a, a stochastic state space where anything is possible, like a random move. These abilities to move in these spaces are constrained by the boundary constraints. And it’s a bit like thinking of, let’s think of what are these boundary constraints?

Think of a large, big building with many different rooms and corridors. So, to get from one area of the building to another area of the building, we can’t just go in every direction, the structure guides us. We can go into one room, into one corridor. Each of these boundaries are constraining the ability to adapt within that context, to move within that context.

And this introduces the notion of path dependence, because we’ve talked previously about that, with stochastic modeling (stochastic modeling, which is a hallmark of the way most people think about traditional financial markets), with stochastic modeling, this teaches us that really markets are ergotic. In other words, they’re not constrained by these time ordered constraints or spatial constraints. There’s no sense of path dependence.

But when you get these fractal structures; the signals and the boundaries all interacting, we get this curious arrangement where suddenly there is an order, a deterministic order, to what can happen. And there is a degree of freedom in what can happen, but it is in a very ordered way, an ordered, structured way.

And this brings us to the concept of what they call strange attractors in complex adaptive systems. (It seems a really curious word.) Complex adaptive systems are defined by the fact that whilst they exhibit this fractal structure and whilst they exhibit these signals and how they interact with boundaries…

So, boundaries are not just something to tell what can occur within and what can occur outside of that boundary. In other words, they’re not just what separates internal from external, like a cell wall, they’re actually dynamically interacting with the signal. So, if you can imagine, in the human body, for instance, we’ve got boundaries at all levels throughout the human body. And the signals that occur throughout the human body occur in lots of different ways. The signals are not all consistent.

For instance, we get chemical signals at the cellular level, but at the brain (and understanding a lot of systems, another boundary level), the signals are electrical. So, these boundaries are shaping, what are these signals, what they can do, how they’ve got to be transmitted. There’s an interaction that’s occurring between the boundaries.

The signals and the way all of these boundaries nest within each other, they create a huge correlated fabric in every complex adaptive system, like a human body. So, there is no sense of autonomy or independence in any part of our body, even though we’ve got these, what appear to be boundaries around the cell. There are feedback loops that emanate between boundaries to the next level, to the next level. And these feedback loops progress in, what I call, an emergent way.

So, what starts off as a chemical signal, ultimately, as it goes through the body, becomes an electrical signal for the higher order boundaries that exist for organs, for brains, all of these things. So, you can start seeing the nesting of relationships that occur in these complex adaptive systems and the signals and boundaries that occur.

So, in the context of a financial market, we see these ecosystems happening all the time. And in these ecosystems, there are different signals that participants are reacting to. And also, now we’ve got to look at, right, recognizing that we’ve got all of these independent agents working within their boundary confines. How do they all react together?

And this is where we can start seeing when we basically zoom out and we look at what are the impacts associated with this boundary reacting with this boundary, reacting with this boundary? We see that it’s a reaction of positive or negative feedback loops that occur; how these signals within these boundaries are being transmissible to the next overlying structure. They’re influencing, or they’re creating a positive feedback loop, or a negative feedback loop.

So, a positive feedback loop is when signals reinforce each other and a trend follower, one of these particular participants that actually focus on positive feedback loops, these enduring trends, is caused through positive feedback amplification, nonlinear amplification. You’ll see that they’re path dependent, their serial dependency, in how these loops unfold, these positive feedback loops, and they amplify.

But a mean reverter, or a convergent trader; negative feedback loops. What they’re doing is that they are explicitly going against the trend. They are therefore destructively interfering with that positive feedback loop.

And so, there’s this continual battle going on with positive feedback loops and negative feedback loops. But this is what stabilizes the entire system; the reaction of all of the participants, how they were all reacting together. This is what’s giving the system this stability.

So, we can’t obviously all be trend followers, as the system would explode with positive feedback. We must have these other ecosystems in the financial market that basically allow these financial markets to continue to operate, be robust, resilient. And this is how these systems remain stable, because at all scales of the market. From the individual trader, and as you zoom out, zoom out, zoom out, this is occurring at all scales, and this is what is facilitating the stability of the financial market system.

It’s interesting when we look at things like outliers, and here’s an argument I’d like to say, Niels. So, in the context of traditional models, people tend to view outliers as aberrations or anomalies.

But I’d just like you to think of, let’s think of a rainforest. And we think of the rainforest. The dominant form, in a rainforest, are the leaves of the canopy trees. All with this fractal structure. We’ll see some branches which are sort of bigger fractal structures. Then we come to the trees themselves, massive big trees. When you look at the frequency of these massive big trees against the frequency of the leaves, etcetera, you tend to start thinking that they’re aberrations, anomalies. But really, that is a natural part of this nested hierarchy of systems within systems.

And in financial markets, they are not aberrations or anomalies. They are a natural part of how this complex adaptive system behaves. They are a natural part of the fabric. They’re to be expected much more frequently than what these traditional stochastic models would have you believe. It’s just a way that these complex adaptive systems adapt and evolve over time.

So that was just a fascinating couple of books I’d encourage readers to get into if they want to go down this path. And the good thing for me is that I’ve been wrestling. So, as you know, Niel’s, I love my science, and I’ve been wrestling with two aspects of science. And there’s this controversy that exists between how we reconcile the very big with the very small.

And what I’m talking about here is the very big with probably our best models, such as general relativity, versus the very small, and here I’m talking about quantum mechanics. And as you know, Niels, there’s a big sort of battle going on at the moment. How do we reconcile these two approaches? And at the moment, it’s irreconciled.

So, the idea being that physicists would like to see this unification of these two great principles from the big to the very small. They’d like to see how there’s a seamless process from going from small to big. But at the moment there’s this massive divide, and we can’t reconcile the two. But when we start looking at chaos theory, we start seeing a way forward.

And this is represented by two good thinkers. One is a bloke called Steve Wolfram. Steve Wolfram is Professor Steve Wolfram. He’s worked on a lot of Wolfram mathematics. A lot of viewers would be familiar with Wolfram. He’s a really interesting fellow. And he’s trying to basically reinvent the universe with computer based models. And he’s starting with what we call cellular automata.

And through this rules based iterative process, he’s allowing the computer, in an iterative way, to basically build how a complex adaptive system works. There are two things to understand here. There’s a branch of theory called theoretical maths, and there’s a branch of theory called computational maths.

So, what we’re saying is, throw away the theoretical maths. So, the theoretical maths is the thing like perfect circles that don’t exist in reality. They’re abstract mathematical ideas. Let’s look at things that we know can happen, iterative processes, step by step by step instructions. So, let’s start looking at almost like an algorithmic process of how the universe unfolds. And let’s go to what we call computational maths.

And when you go to Steve Wolfram’s idea, starting from this very simple structural relationship of fractals; basically rules based algorithms that decide how the system iterates going forward, you start seeing that all of the structures of this universe unfold through this process of emergence, as these models get more and more complex. We start seeing theories like QM, quantum mechanics, theories like relativity. They unfold out of this process, of this principle of emergence, as we slowly, progressively iterate the universe.

And the theories of time, why time is ordered, can be understood with this iterative manner. It’s not a case of a linear projection in time. It’s the unfolding state of the universe with a rules based process that iterates and iterates and iterates.

Now, initially, when I heard this, I thought, well, that’s a different novel way to view the physical universe, et cetera. But the more I’m getting into chaos theory, the more I’m convinced there is a unification of Steve Wolfram’s views with the ideas of chaotic fractals, things such as strange attractors.

Now, in our universe, when we start looking at strange attractors, we realize that whilst things can be very deterministic; deterministic rules do this, go left, go right, go left, go right, what we find is that when we are applying nonlinear mechanics under this deterministic process, we find that the ability to predict into the future is limited or impossible because of these strange attractors.

So, I talked about signals and boundaries before. Now, what the boundaries are doing is they’re interacting with the signals, and the signals are interacting with the boundaries. In other words, the small is guiding the big and the big is guiding the small. There’s actually a deterministic way that this system unfolds as it iterates through the course of time. So, it’s not exploring all avenues like a stochastic method. It’s a rules based deterministic process that is iterative and sequential. But the result is that there is no repetition over the entire sequence.

So, when we understand what does predictability mean, is it means that there needs to be a repeat of a sequence somewhere for something to be predictable. But what we find with this chaotic structure is that there is no repeating sequence. It’s this continuous unfolding.

And then we see, in the world of Steve Wolfram, he refers to these systems that are so complex that they are what he calls computationally irreducible. Now, what this means is that for an observer embedded in this universe, if they wanted to understand this system, it’s not that it can’t be understood. It’s just that they would need a computer that lasted the length of this universe to iterate in such a way to be able to predict what the system is going to do next. And of course, all observers in this universe are limited, and they cannot do that. So inherently, it still is unpredictable, but still is in a deterministic fashion, if you know what I mean.

So, this is exploding my mind in that it’s making me realize that what I viewed as randomness, what I viewed as objects that occur in isolation, it’s all fading away quickly. I’m finding that just examining this in detail, the more I’m reading, the more I can see how it all interconnects together. And the more I’m amazed by the power of this process. So, I don’t want to wax lyrical anymore, but it is a good topic, this one.

[01:05:16] Niels

It’s a fantastic topic, and no one better to guide us through that than you, Rich. Now, given the fact that you do like this, it also means that we actually ran out of time just going through topic number one, but that’s fine. That doesn’t mean that we can’t enjoy some of the other things that we were meant to talk about today.

Now, of course, there were some analogies, actually, to what’s been going on in Florida, with hurricanes and so on and so forth, but we’ll pick it up what we find relevant in a few weeks when we speak again. But this was tremendous. And as always, I learned something new, as I’m sure all of our listeners do. And the more, as I said in the very beginning, the more we can really appreciate how this system works, I think the more we can appreciate the strategy that we employ to try and make sense of the markets that we see. Very interesting indeed.

Now, next week I have another great person to guide us through some of the intricacies of trend following, namely Katie Kaminsky. So, if you have any questions for Katie, then feel free to email them to me. Info@toptradersunplugged.com I’ll do my very best to bring them up.

And of course, as always, if you enjoy these conversations and all the preparation that goes into them, not least by all the cohosts, do show your appreciation by heading over to iTunes, Spotify, Amazon, wherever you listen to podcasts, leave a rating and review. It really does help, and it actually encourages us to continue this journey

From Rich and me, thanks 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 of each other.

[01:07:18] Ending

Thanks for listening to the Systematic Investor podcast series. If you enjoy this series, go on over to iTunes and leave an honest rating and review and be sure to listen to all the other episodes from Top Traders Unplugged.

If you have questions about systematic investing, send us an email with the word question in the subject line to info@tradersunplugged.com and we’ll try to get it on the show. Remember, all the discussion that we have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance.

Also understand that there’s a significant risk of financial loss with all investment strategies and you need to request and understand the specific risks from the investment manager about their products before you make investment decisions. Thanks for spending some of your valuable time with us and we’ll see you on the next episode of the systematic investor.