Why Systematic? Why CTA? Why Now? These are common questions on the minds of many professionals in the Managed Futures field. On this episode of the Derivative podcast, host Jeff Malec (@AttainCap2) takes on a unique role as a featured guest on a panel recently hosted by RCM Alternatives in Chicago alongside Cohen & Company, and Mercer.
While Jeff is usually the one leading the discussions, this time, he gets the opportunity to be a participant alongside other industry experts. Together, they delve deep into the inner workings of various trading models. The main focus of the discussion is to understand how these models function, distinguish between pod-shops and multi-strat data infrastructure, and explore the role of AI in the industry. Additionally, they discuss the differences between systematic macro and trend-following approaches.
Joining Jeff on the panel are Joe Kelly from Campbell, Brian Proctor from EMC, and Rodrigo Gordillo (@RodGordilloP) from Resolve. The conversation is both insightful and engaging, shedding light on the world of Managed Futures Hedge Funds — SEND IT!
________________________________
________________________________
Check out our Trend Following Guide!
Check out the complete Transcript from this week’s podcast below:
Why Systematic? Why CTA? Why Now? A panel event with Mercer, Campbell, EMC Capital & Resolve
Jeff Malec 00:07
Welcome to The Derivative by RCM Alternatives, where we dive into what makes alternative investments go analyze the strategies of unique hedge fund managers and chat with interesting guests from across the investment world. Hello there, hope you’re enjoying the dog days of summer here we’ve got a good one for you today to listen to while mowing the lawn or sipping a lemonade or whatever you do. I got to sit in the guest seat during a panel last week little bit different from me, but it was fun. And actually I was really a little bit more of a listener as some of the great pros in the Managed futures hedge fund space. Were also on the panel and really talk more about their models and let them talk. The panel was titled Why systematic why now? We dove deep into how models work the difference between pod shops and multi strats data infrastructure and AI and finished with an insightful Convo from an audience question on how to delineate between systematic macro and trend following. Here we go with Joe Kelly of Campbell, Brian proctor of EMC and Rodrigo Gordillo of Resolve. Send it. This episode is brought to you by RCM outsource trading desk. Did you know RCM does the clearing and execution for several hedge funds ETFs and mutual funds like the one on this panel, utilizing futures options and more. That’s right. Check it out at RCM olds.com And now back to the show.
Bobby Schwartz 01:35
We’re gonna get started. Thank you everybody for coming. I want to thank RCM, Matt Bradbard, Cohen and Mercer for putting this together. We have a great panel today. Rodrigo from ReSolve Joe Kelly from Campbell. Brian Proctor from EMC, RCMs own, Jeff Malec and David from Mercer who’s going to be hosting and helping everybody channel through the conversation. Thank you very much for coming.
David Triplo 02:06
I will thanks again for for everyone making it here today. I mean, we will have a great panel with these three CTAs that have been around for forever in the industry and Jeff Malec from RCM. I’d like to start off with kind of getting an idea for the audience like how many of all of you have already invested in CTA Halfords so so as the of that half, it would actually be good to know the breakdown of how many of you are focused on more trend following CTAs versus other strategies outside. Okay, so definitely lower class on that side. So I feel like most people always get the idea of like when they look at CTAs and Mercer views, it is systematic macro as a whole is two different buckets. So we kind of view it as there’s more trend focus, which is 50%, or more in trend or multistrike in general. So just kind of to start off with talking to Jeff a little bit more. You good to know like how you guys deal with the universe in what are How should people be like approaching it for CTAs as a whole?
Jeff Malec 03:07
Sure, and welcome, everyone, thank you. I really look at it as we did some research. And last year when CTA has had been doing well ran some historical testing back to Jan of 2000. I think the numbers ended up the compound annual return was about four and a half percent vol 11% drawdown 18%, that was an average across four indices. We put that out on Twitter, and got a lot of comments back of like, what what the hell that doesn’t look so great. That looks terrible. Coming full circle, what we find attractive. And what the clients find attractive is that’s positive carry, that’s four and a half percent across that whole period in order to survive and get those those pops in the downturns in the OA in the 22 to 2014. So for us a little bit less. Am I doing trend following in my multi strat? How does that fit in and more of that return profile? Okay, are you pure trend following? Are you systematic macro? I don’t care quite as much if you deliver that return profile that’s gonna pop and have that positive carry.
David Triplo 04:15
Yeah, I mean, that makes sense. I mean, we’ve just seen that there’s been numerous environments work during every equity market, drawdown that it doesn’t necessarily mean that it’ll correlate well with each CTA. And in terms of performance wise, I mean, some have a better vaccine profile versus another. And during the discussion, we’ll kind of get more specifics on that for each strategy. So that yeah, so kind of talking about the equity market drawdowns and the different profiles from there. So like, let’s kind of moreso back up a little bit and just talk about it from a high level of just why now for systematic macro because you know, that certain environments are better for certain profiles, but just looking at it from a broad perspective is what are the main reasons in this macro economic environment? that we’re in that makes sense for, for pretty much everyone to be investing in systematic macro, maybe we’ll start with Joe.
Joe Kelly 05:08
Yeah. Hi everyone, Joe Kelly, I run the institutional business for Campbell, actually out of Chicago. So if we haven’t met, my wife and I moved back to town about a year and a half ago, so the office main offices in Baltimore, we run about 4.2 billion across what we consider kind of two multi strats. And about a quarter of the assets and pure trend, you can use in our mind multi strat, interchangeably with systematic macro, I caught on a couple, maybe a year ago, year and a half ago, there’s always this frustration and this tension between systematic discretionary macro and multi pm shops, and you tend to have the same conversations with a lot of the same investors. Multi Pm is obviously taken in a ton of money. And behind the scenes, what we’ve been told is, ultimately, you know, the track record, and the numbers are what they are, but they’re better at allocating risk across strategies than a lot of alligators. And so the way we sort of approach systematic and kind of the why now, we have this really cool what people call, I guess, I call it a periodic table, or a quilt chart that basically says, like year on year on year, we can’t tell you what’s going to work, we can’t tell you trend or systematic macro or short term is going to sort of be the dominant strategy. But through risk allocation, you know, we can systematically allocate to what works or D allocate to what doesn’t faster in our opinion than kind of the human brain. And even specific to trend strategies, one of the challenge challenges with institutions is when they need it, they’re always too late to the game in terms of board decisions, and getting it through the process, and then they buy the top. And so for us, or the message we usually give people is, you know, respectfully, during periods of nonzero interest rates, where there’s a lot of dispersion of global markets, and you don’t know, frankly, where to put your money. We think systematic has a role to play in that, you know, I think, I’m not gonna say smarter, you know, faster and in a more tactical way than any of us are on a discretionary basis, if that makes sense.
David Triplo 07:27
Makes sense. Brian, do you have anything to add to that?
Brian Proctor 07:31
Sure. Brian proctor from EMC capital, we’ve been in the business since the mid 80s. So traded through a lot of different environments. And obviously, you’ve seen a big evolution and changes in the CTA space over the years. So when I think of systematic, I always think it’s the, it’s a quantitative approach to trading. So it’s very rules based, you’re not making discretionary decisions, you’re taking all the human emotion out of trading, your own biases, fear, greed, all that stuff goes by the wayside. And you look to build trading systems that give you the optimal prices to initiate and liquidate trades. So that’s what we mean by very systematic, your, you know, your you manage your losses, you always have a hard stop for every trade you make. So that discipline is it’s good for any investment manager to have, whether it’s discretionary, or systematic. And then, as far as systematic CTAs, we, our strategies are all directional. So we’re always looking to capture directional price movement up or down and the markets that we’re trading, we’re never going to be doing mean reverting or counter trend types of trades. So in evaluating how well we’re doing in the current market conditions, if you’re, you know, if you’re allocating to us, you can look and see what our positions are, and basically say, okay, you know, these guys are doing what we expect them to do. And as Joe said, you know, we don’t always know where the next good trade is coming, whether from a specific market or a sector, the currencies or the grains or the energies. So when you get a systematic CTA, what you’re really getting is a very diversified opportunity set, and very unique markets that were going long or short, equally as well. So you know, some years like this year, the best markets tend to are like cocoa and sugar and the Mexican peso. prior years, it was being short interest rates or long lumber palladium. So you, you know, we’re in some very esoteric markets that you’re not going to get in Traditional investments and by blending us with those, you know, it’s just a better, more balanced portfolio.
David Triplo 10:10
And Brian, would you say because of the fact that every markets have its own kind of regime, so to speak, that the fact that systematic like trend following CTA has performed very well over the past two and a half years minus last, like four months or so. So would you say that it really should be used in a portfolio is a full context of it?
Brian Proctor 10:33
Absolutely. You know, the question of, Oh, should I make the tactical decision to invest now, or wait for a drawdown? And I believe me, I’ve had, we’ve had plenty of investors who wait for you to get into a drawdown and they say, wow, we’re gonna wait to see you come out of it now. And it’s like, so. So you make, you know, it’s, it’s, you know, you can make a tactical decision. But, you know, it’s more of a strategic decision that you have to make to say, we’re going to have CTAs, we’re going to have this diversification in our portfolio. And we’re going to stick with it. And we know the return stream is going to be different than other assets that we have. But in the long run, that’s going to help improve the returns and the risk adjusted returns of your portfolio.
David Triplo 11:20
Yeah, makes sense. And then Rodrigo, kind of same question for you for
Rodrigo Gordillo 11:24
Yeah, I think the question was why now. So Rodrigo, president and portfolio manager of resolve asset management, we run just under a billion dollars in a multi strat, CTA. The, you know, from the trend side, we’ve also started a company with Corey Osteen called Return stacked etfs.com. And I think one of the reasons why now is because we are in a global macro environment that we really haven’t seen in 40 years, right, we, for the most part for our careers, and all the instincts that we have garnered from our experiences as advisors or portfolio managers and traditional assets has been what I call a two dimensional game, almost like balancing on a board in a barrel, right? You go on either right or left bonds or equities, the Feds been managing liquidity or taking away liquidity without any care about inflation. Now, as a Latin American, I can tell you that inflation has been front of mind from the beginning of my my life, I emigrated from Peru in 1989, after a 7,200% inflationary thrust in six months that left us kind of penniless and emigrating to Canada. So I, I had front and center my whole life, the importance of protecting against inflation. And it turns out that the best thing to do that is to be in some commodities or be able to short some, some bonds. And nothing offers that directionality in all the categories that we see in the alternative space, then the CTA space, whether it’s multi strat or a trend or otherwise, you have the opportunity to take really interesting long positions in commodities to take really interesting short positions and bonds and equities, of course, and currencies. So that’s, you know, most portfolios for the last 40 years have prepared themselves for, for either growth or disinflation. Now, we need to add that third element. And that turns us into a three dimensional game, right? Because once you introduce inflation, I can tell you that the dynamics change, it’s a lot more problematic to really bounce a portfolio out. And you don’t know when inflation is going to hit, or disinflation, aggressive disinflation is going to hit we tend to call it inflation. Volatility is what’s here to stay, we’ve seen an inflationary thrust. Now we’re going to we’re seeing a disinflationary thrust, is it going to end now we’re going to hit that magic market? 2%? The truth is, we don’t know. And I would say timing, this thing is incredibly problematic. If we could time CTAs performance, we would have done it. That’s our job. Right? So I’m putting out a piece in a couple of weeks called timing the timer, you really have to think about this from strategic allocation and put it in as a third leg to your stool that can deal with inflation and bear markets in a way that bonds equities can. The last reason I think it’s it’s really exciting now for retail investors, advisors, small pension plans. And so what is is, you know, institutions have always had access to this, this ability to get the excess return to stack, this CTA strategy on top of your traditional strategic asset allocation. But retail investors haven’t left out. Now, return stacked ETFs, for example, has an ETF that has 100% bonds and 100% systematic trend. That’s, you know, you give them $1 You get $2 where you can instead of saying this or that and having to make room in your portfolio, you can say yes, and you can actually stack it on top and because of its of its low correlation, it doesn’t necessarily stack returns were one or the products that does this or or RDM is our other product was the mutual fund is risk parity plus systematic global macro, right? So we’re not the only ones doing this, there’s a wide variety of funds that have this stacking capability. But it really allows investors to, yes, and this to stack things on top to not sacrifice their equity and bond returns, but rather put things on top that can help them and their clients kind of thrive in this third dimensional game that we’re going to be playing.
David Triplo 15:33
Yeah, I mean, over the past 10 years, really CTAs kind of had a top performance really from prior 2012 to 2019. And how much of that has really been related just to the interest rate environment that we’ve been in? And like, how will our current regime really be affected from that? Like, you’ve kind of hinted at that I’d be curious to hear more from from Joe on that.
Joe Kelly 15:55
Yeah, there’s there’s definitely different opinions over the years whether CTA has made all their money on on rates when you know, the salad days of the mid 90s, that we haven’t seen until now. Or, you know, whether we were maybe biased to a longer short rate roll up over 30 years, and we’ve actually done a white paper on, you know, CTAs performance during periods of rising rates. And, you know, it turns out that they’re, for the most part, you know, the folks up here, and the investable universe of CTAs that have done their homework. There’s, there’s no inherent bias towards long or short rates, there was a tailwind because a lot of these portfolios have, you know, minimum margin requirements, 30% margin requirements, so you have 70%, in cash earning risk free, we tend to position our returns on top of that, and a very similar way to that Rodrigo is talking about stalking. And so what we decided, I guess each of us will kind of tell our approach, what we decided is we don’t want to leave trend behind because we believe in the convexity, and frankly, we had had, at that point, 30 years experience in trend following, but we did have hires along the way that said, Look, just because we’re systematic, doesn’t mean we have to be a trend follower. We could also it was a gentleman named Bruce cleeland, who said, you know, we can apply this systematic process to, you know, diversifiers and, you know, the early diversifier was Carrie, we all know that that can be a great trade, but also can have a big left tail. We put in a cash equity infrastructure to try to smooth the ride in the future side by adding equity strategies. And then there’s this introduction of short term and qual macro around along the way. And so, you know, I have worked for some world class CTAs that have said, like, we want to build our business on trend, because it is the great diversifier, we made a decision at Campbell that we want to build that and not leave that behind, but want to smooth the ride for some institutions that maybe wouldn’t otherwise invest in systematic by adding these other strategies. And what that allowed us to do is in the, the quiet periods for trend is kind of compound on the industry by either introducing RV into the strategies, which I’m sure Brian will have an opinion about visa vie trend, or things like short term just to pick up sort of granular moves along the way. And so we have kind of three ways we describe allocation, systematically, you need to allocate to sort of a style, a market. And then the third is kind of is this directional, or relative value. And then those quiet periods relative value really dominated. There are challenges there are different tail characteristics to relative value, versus directional, there’s more leverage required. So it’s not, you know, this perfect scenario, but it is, every firm has to kind of decide how they want to grow over time. And that’s why I describe us as kind of two multi strats in one trend. And we’ve just seen more growth in the multi strats because it’s a little smoother ride. And to be fair, we gave up some convexity last year, so trend was up 35, maybe ish across the board, and our multi strats were more like 2025. So it’s all about the trade off and what works for the investor.
David Triplo 19:36
Yeah, Joe, you hit a big point there just talking about like the evolution of funds, like all three of your funds have been around for over 20 years and have had different steps along the way of like how you choose to evolve, and how some have chosen not to just to keep true to its roots in its own way. So maybe Brian, can you kind of discuss a little bit more about like your your program and how it has evolved over time. Sure.
Brian Proctor 20:01
Well, when we started in the mid 80s, the risk free rate was double digits. So our investors would look at us and say, Okay, we want at least two, maybe three times the risk free rate. So when you look at our track records going back that far, you’re gonna see some big swings, some big draw downs. And that’s just part and parcel of the market environment at that time. Over the years as rates got into the more moderate level, you know, we adapted our strategies, we use less leverage, we understood the institutional investor was looking for better risk adjusted returns drawdowns, you know, 25% was considered kind of the bar, you know, whereas 15 years prior to that we’d have a 40% drawdown, but we’d explain to our investors, you have to expect this if you want these kinds of returns in multiples of the risk free rate. So as far as the evolution of EMC, our flagship program has pretty much stuck to its knitting when it comes to directional trading. So some of our systems we describe as trend following or range dependent systems. And some of our systems are momentum based, and they all are optimized to a different metric. And each one of those metric metrics is different in nature. And it gives us a way of diversifying, how to get in and out of a market. You know, one system is long, two, maybe all, all systems are long. You know, that’s when the most risk is on. But we tried to diversify, how we initiate and liquidate trades along the way. And actually some systems, longer term systems can be long, you know, a certain market and the shorter term systems that are getting in and out quicker, are actually short. So they’re offsetting each other. So they operate independently of each other. And, you know, it’s just a way of diversifying how we do everything. I mean, we look to diversify in terms of markets, global locations of the markets that we’re trading, the number of systems, you know, to us, you know, the trend following diversification is what we think an alternative investment looks like. We also have developed products over the last five to 10 years that are blending long only stock bond and gold portfolios with active futures trading, it’s about a 3070 mix. So as Joe mentioned, that kind of helps smooth out the return streams and his appeals to other kinds of investors. So we have a couple of different programs that are available out there.
David Triplo 22:54
And for Rodrigo, kind of on the same similar point, at least, like I would just how would you say, more recently, what evolutions in r&d Do you think are important that investors should focus on?
Rodrigo Gordillo 23:07
Well, look, when we first started in the business, there was like anything, you start with alpha that turns into beta, then you have to constantly be putting r&d research dollars in manpower in order to have that, continue to add that alpha that turns into beta. And so when we first started, you know, we thought that, you know, a single factor based momentum or trend was, you know, all we needed, and then you realize that, you know, ensembles is probably better, though it’s really tough to decipher which trend pattern is going to be best in this decade. So we wrote a paper on ensemble methods and how they have the highest Sharpe ratio. And we can’t tell what’s better term between short term long term breakout systems, moving averages, etc. All the while, you know, luckily, our head of research comes from machine learning space, and we had a couple other people that had been making a lot of money using machine learning and not in the AI sense that they’re trying to predict the market but rather, once you have an ensemble of trend factors carry factors mean reversion, relative value, etc. The question is, okay, we have hundreds of 1000s of these signals. We assumed in our previous iteration, that we don’t have a skill in choosing them, so we’ll use them all. As we put more man hours on the machine learning side, what we learned is that you know, you there are statistical methods using machine learning and proper experimental design, which is just making sure your sparse data for testing in the proper way that’ll allow us to really, more than anything, prune out the factors and parameters that are less likely to be real. And when you start pruning those systems out, what ends up emerging is just a better outcome of you’re not you’re not cutting them all to to find the one branch or the one parameter that you data mined to have the best performance you still have a A wide variety of parameters, what you’re attempting to do is to really eliminate the ones that are clearly of no value. So the machine learning aspect is something that we really took us five years to develop. And we really pushed out three years ago on October 1 2022. And we continue to now use that experimental design to bring in a bunch of multi strands. And it can I just say one other thing. So that’s on the complexity side, on the simplicity side, there’s also a value in the beta. And so, you know, on the on the trend side, which you know, is a valuable thing, we decided to make the classic beta desoxyn trend index, you know, we do a replication strategy for trend and put it into an ETF at a reasonable price so that people can get exposure that, you know, return stack bond and trend. So I we’ve kind of gone, we’ve kind of bookended the whole, the whole process
David Triplo 25:59
Yeah, on the machine learning topic, I mean, that’s been a hot topic, really, for the past few years at least. So what would you say Rodrigo is like one thing that investors should be aware of, in terms of like misnomers and space, which I’m sure there are plenty of, and when investors are looking to, like, allocate to funds that are involving machine learning, AI, NLP, and just everything as a whole, how should investors like look at that framework?
Rodrigo Gordillo 26:25
Like this is a complex topic, but I think one of the things that we’ve seen in the last five years is a lot of people come out from non financial places like Google, you know, very good example of a professor that came out of Waterloo in Toronto, in Waterloo, Ontario, Canada, one of the hubs for machine learning, that came out from a place where you have a lot of data, like so much data that you can get statistically significant results that are unlikely to be data mined, and apply that model to markets. And it turns out, you know, random for a few months realized that he’s making a bunch of money goes live, lo and behold, it’s a it’s it was just a bunch of noise that he was hitting. Right. So I think one of the key things is, there’s a lot of people with a lot of conviction about machine learning AI that come out of that space and think they have something, but they haven’t learned that there’s not enough data in financial markets to really attack it from the same problem of big data that Google and Facebook and and Tesla, have, we have sparse data. And for that, it requires a much more thorough experimental process of testing the data with holdout sets and, and, you know, making sure that you’re not fooling yourself, because you’re the easiest one to fool as they say that they, what’s the sign that the? Anyway, it’ll come to me in a second, but a scientist once said that, you know, you don’t want to fool yourself, and you’re the easiest one to fool. It’s important for people allocating to ask the questions. How do you as far as you’re telling me about how much data you have, and how it is that you are using your back tests in order to inform what you’re doing in live performance? And that’s a key question that everybody needs to ask. And if they cannot answer it, then you might be getting into trouble at the end of the day, where we’re not allowing the machine to tell us what to do. We’re handcrafting the parameters, we have a wide variety of things that we know have fundamental reasons to exist from frantic carry, and so on. And then we’re allowing the machine learning process to help us guide the parameters a bit.
David Triplo 28:34
Yeah, Jody Everly? Any comments on that? Because I know you’ve kind of focused more on short term more recently and have some interesting things you on on the website, on your website that kind of go into some pretty intricate details to say the least on that space
28:49
Fineman? Yeah, that’s fine.
Joe Kelly 28:51
Yeah, it’s, um, some of it has, I think you I think that was really what’s really well said, around shrinking Parameter Sets and finding, you know, sort of hundreds of parameters instead of millions of parameters and those techniques for shrinking them down. And then and then treating sort of the meaningful pieces. I usually say two things. First, I have to talk my team out of saying we’ve, we either don’t do machine learning, or we’ve been doing machine learning for a long time. And now we’re just calling it machine learning. I’m looking at your smile because you’re probably have the same conversations. The consequences between the gentleman Rodrigo is describing coming out of Google and applying you know, whether you’re going to find somebody from your high school class, and whether you’re going to make somebody enough money to retire are entirely different. I watched the AHL sit in front of a state pension eight years ago, and they they had a facial recognition story around AI and they had a center at Cambridge, I believe it was that they renamed the building of the HL and none of had anything to do with the consequences of making or losing money based on an unsupervised, you know, piece of machine learning, which nobody in the room wanted. So the way we kind of describe it is, you got to think about the consequences, you have to think about your skill set. What we do do is supervised learning have a limited set of parameters that is being innovated on, but it’s not self adaptive. And so, you know, I haven’t met an investor yet, at least that I deal with that wants me to put things into kind of a black box. And in terms of what people wanted to ask or what people should ask. There was another Steve Benson, who one of our big beers, maybe the biggest peer in our space salespeople had been in saying, We’ve got hundreds of years of data, we’ve traded fluid through inflation. And now we’re, you know, 100% ai, and the CIO of this state plan said, I already have a, I have it through this manager. And if you want to the CEO of that firm today, who is now retired, not to give it away too much, he would have fired that person immediately. It was pure marketing, and not blessed by research. So I would listen to the researchers. Most of them are going to be very honest around the fact that there is an opportunity in supervised learning, but the consequences of what we do we take very seriously across this, this group. And so you know, applying that to financial markets, I think you just got to have a healthy skepticism around.
David Triplo 31:45
Yeah, I mean, we’ve talked a lot about just data in general. And I mean, that’s a core component of any systematic macro strategy, whether that’s going to be price data or fundamental data. I mean, up here, we kind of see the full gambit of that. So maybe Rodrigo, can you kind of go into a little bit more about how you’re like kind of sourcing your data a little bit, how you’re making sure everything’s like playing, because I mean, that’s always such an important part, knowing that there are a lot of datasets that are definitely not reliable.
Rodrigo Gordillo 32:14
A data scientist is a sexy name, but it really is spending 90% of your time fixing your dataset, and 10% of the time testing on it. I mean, the amount of effort and energy that has gone and continues to go into creating systems to clean data, you know, futures contracts aren’t continuous, you have to stitch them together, you have to make sure you stitch them together in the right way. There’s multiple contracts that you can use through your testing environment. And so the infrastructure that’s necessary to be able to run your data to test on your data to continuously clean your data. And as you emigrate from an old trading infrastructure, where your team is now redesigning, faster, you know, newer language, so you have to, you know, continuously improve continuously get faster, all of that process takes years and years of, of time, and team that knows what they’re doing and a backup team that that has redundancies to make sure that you’re not screwing things up. So it’s it’s this idea that, that you get a lot of people want to be do it yourselfers, you know, I’m just gonna go on my own and try it out. It’s, you know, in theory, it’s easy to swallow some trends. In reality, you can get hurt if you’re not doing all of that back and work.
David Triplo 33:39
Yeah, we’ve talked a lot about kind of more Alpha sources throughout this conversation. But we really haven’t talked too much on risk management, kind of within this space, it’s definitely more set up differently. The fact that it’s a systematic program, compared to most funds within just a hedge fund universe as a whole. So maybe, Brian, if you could talk a little bit more about like, how your you view risk management within within your fund?
Brian Proctor 34:03
Well, I would say risk management is the number one priority of any CTA properly managing and thinking about risk is why I think see TAs are able to last 1020 years or more. Trading Systems can come and go ideas, you know, using machine learning to we always said it was unsupervised learning, and we would open our parameter sets. And let them you know, the algorithms determine what the optimal Parameter Set was we used to we used to sit at a table and say, here’s the optimal parameter set for each one of the parameters. And the parameters are, how do you initiate a trade? What kind of price action How do you liquidate a trade? And then maybe a couple of filters in there that say, here’s, here’s when you want to stay out of the market when this kind of volatility is happening or whatever. So every system will have maybe anywhere from eight to 1112 parameters, the fewer the better. But in terms of managing risk, we we do it at every, every level of the portfolio. So I think one of the things that our firm does that might be unique, I’m not quite sure what other CTAs are doing. But we have a very short look back window when determining the volatility for each market in our program. So we basically are looking at the last two weeks of trading data. And we’re looking at the average true range of all those markets. And on our trading platform, we have a risk page that will show the volatility and how it’s changing in each market from day to day, it could be good, or it could be shrinking, or it could be expanding. So that’s the first thing you have to take into account, because when you’re sizing positions, like we do, you have to normalize the risk across the market, you might take 50 Korean contracts, but only eight Japanese bond contracts, because you want to risk the same amount of capital per trade in each market that a signal is being generated. And so looking at individual market volatility, at a very short look back window, is important to us, we feel that’s a better way to respond to quick changes in volatility in the markets that we’re trading, you know, global macro event, unforeseen happens overnight. And obviously, you have to be able to monitor that. And any new position you could have existing positions on involves increasing. And that’s good value. I mean, we always think of volatility kind of as a two edged sword, good volatility and bad volatility. So when, when, when your position correctly, and volatility is expanding, those are the trades that in the long run, create the returns for CTAs. And if it’s bad, vile, while you’re liquidating that system pretty quickly, so
Jeff Malec 37:06
another quick question for you. How do you how do you protect against a historically volatile market being very non volatile in those two weeks?
Brian Proctor 37:15
Well, there are times when a market say like the Euro Yen, when the price of the early on was flatlining, we would not allow the contract size to get above a certain level because we recognize inherently that something could happen to make that market move 20 standard deviations overnight. So in that, in those rare instances, you know, we put a limit on the number of contracts we’re going to trade take, but as far as protecting yourself from unforeseen market moves, I can think of one in particular when the Swiss franc decoupled from the euro. And it was a 26 standard deviation move in like one second. And, you know, so if you don’t limit the amount of risk that you take in each market, and each sector that you’re trading, you obviously open yourself up to concentration risk. So I think we lost four and a half percent on the short Swiss franc position we had that day, which was somewhat offset by a long Euro Swiss position. So natural diversification happening there. But so you really, you know, you look at the volatility, you look at each individual market and say, here’s the maximum amount of risk that we’re going to take when we’re initiating the trade and evolves expanding and you’re making money. Great. I think another thing that we do a little bit different in the way we manage risk is when we get to certain profit objectives, we’re up 1015 20 25%, during a good run, we have a mechanism in our programs that actually start scaling existing positions, taking small profits gradually over time, so that it helps us improve the drawdown from Peak equity, and helps our standard deviation. And usually those periods where we’re making those outlier profits vial is expanding. So we’re getting our positions, kind of back to where the current vial is, instead of the vial when we put the positions on. So those are, you know, some of the things that we’re doing kind of unique in the risk management space, but NACTA the first thing they should talk about when you’re talking to them is how they manage risk and how they think about risk at all levels of the portfolio.
Jeff Malec 39:39
And David, I’ll just jump in on an allocator perspective. Right, I’m looking at all three of these guys to allocate to I want to manage the risk across them. I don’t want EMC to be driving the portfolio when Campbell is has less vol so what we’ll do there is look at your max drawdown your annualized vol and your margin usage base They get a risk score out of those, if you’re twice as high as him, I’m going to be doing twice as much of him. So that involve weighted across risk weighted across those three managers. And then also that you were saying ensembles of your, we like to get ensembles of ensembles, right. So you have an ensemble, you have multiple timeframes, you have multiple strategies. To us, the best method is to blend all three of those together in an intelligent way, in a risk weighted way. And then we can get an example of ensembles and have even better performance.
David Triplo 40:33
Yeah, Mercer kind of takes the same approach is that not necessarily just within systematic macro, but within our hedge fund portfolios as a whole through our OCIO groups? One last question for me, and then we’ll kind of open it up to the audience and quick one, discretion is sometimes used across a certain CTA, some feel it’s more appropriate than others, some just prefer to let the systems itself always kind of correct for the situation. So just kind of quickly, does your program use discretion at all? And like, how frequently if if ever?
Rodrigo Gordillo 41:14
Yeah, so this question for us is almost exclusively at the risk management level. So we do allow our systems that have been, you know, again, all the risk manager and all the work has been done up front. But things like the yen, where, you know, it starts to flatline, you know, that we number one, we’ve already thought about what maximum level of volume of exposure, we’re gonna get to each sector. So that’s predetermined. But for us, when we looked at that, we said even That’s too risky. So from a risk management perspective, we said, we’re only going to, we’re going to take it off for now, as we look at the global macro space and see if that’s going to blow off or not. What is taking off that contract due to a Sharpe ratio going back? Is it a big impact? No, it’s It’s not a big impact. What’s the impact? If it goes against us? Even with those limits? It’s big, why don’t we take it off until such time as they d peg that situation? Other things are when we look at the the historical risk characteristic version of our strategy. And when there’s certain days that are above a certain standard deviation of the history of our strategy, we immediately put the risk management team together, to sit down, get the whole team to figure out what’s going on, how are we doing against our peers? Is it a resolved thing? Is it a, a CTA thing we’re all struggling through? Okay, so those are the kind of checklist that we go through in order to decide whether we want to take some risk off the table. So if we find that we’re offside and nobody else is, we’re going to go deep into the system and understand whether we screwed up somewhere, and we’re going to take risk off the table before we go in to do the work, we do the work, we find something, this has never happened, this is all hypothetical, but it’s in the checklist, then we will fix it and then take the risk back up. If everybody else is struggling with it, then, you know, there’s really nothing we can do. The system is automatically designed to take risk off the table. So there’s been two or three occasions where in the meeting, were like, Okay, we’re gonna take 30% risk off. And the system is already doing that the next day. So we do have discretion, but at the risk level.
David Triplo 43:21
Joe, how about you? Same question.
Joe Kelly 43:26
We try as hard as possible to remove discretion from the whole process. There’s outliers, you know, this last year, we removed the Russian ruble, just before, you know, the invasion, just to take that out of the entire process and portfolio, not necessarily from a political stance perspective, but from a liquidity perspective. Before that, you know, when the tsunami hit Japan, we removed sort of all the Japanese markets because we didn’t know when the exchange were going to we’re going to open or for how long they would be closed. So those are easy ones. Those are ones where, you know, it’s liquidity driven. It’s always taking risk off the table. It’s kind of the known unknowns, like election risk and unknown unknowns, you know, like COVID, that drive us probably all kind of crazy around, you know, the temptation to use discretion. And throughout those periods, we have sort of been dogmatic about things like ball floors around now loading up on contracts that have flatlined from a notional perspective, we have this wonderful pet of risk named Grace Lowe, who’s co head of IC, and then that I see meets every single day and if there are events, as Rodrigo was saying, that may look like outliers either to Campbell or to our piece of the industry. Grace is always in there trying to generate some systematic approach to That event or environment and a good a good example is something like, like using implied volatility of the pound going into Brexit, instead of realized volatility allows you to systematically kind of cut your position because implied vol is reflecting the uncertainty of the event better than realized. And then the event happens, implied volatility collapses to realize and you normalize your position. Again, it’s just a way of systematically taking risk down by using a different input going into an unknown event. And so we do use discretion. It’s kind of a five person I see committee, but it is extremely infrequent and always around taking that liquidity risk out of the portfolio.
Brian Proctor 45:48
Well, I, and I’ll chime in on that. First of all, when I think of discretion in the research process, there’s always some discretion going on, you’re you’re figuring out how much risk weighting to give a market or, or the discretion what kind of trading system parameters do we want, that will, that will affect each of the systems that we’re trading. So in that sense, you know, there’s a regular amount of discretion, but it’s all in the building block process. Once you have the system set, then you don’t use discretion on when, what signals to take or whatever. Second of all, I agree with Joe, there’s, there are times when we, we use discretion only to reduce risk in the portfolio. So what, you know, there’s election risk, or whatever it is, we are in 2008, when you came in on Monday, and, you know, banks were out of business and, and we were up, you know, a substantial amount on that day. And we knew the volatility across all the markets had was probably 20 or 25%, higher than the previous day. So we’ll make a decision to say, Okay, we’re going to cut all the positions that we have in the portfolio by 20 25%. Just because we know that even the short look back window we have for the volatility of the markets that were trading, that you know, today is 20, or 25%, more volatile than it was yesterday. So that would be a very rare instance of when we decide, okay, we should just cut some of the risks because it’s, it’s so elevated right now. And then finally, the last place that we use discretion fairly regularly is in spreading commodities from the commodity markets from one from the front month to the next month when expiration comes because, you know, 50% of our classic program is essentially in commodity markets and not financial markets. So we’re constantly monitoring the spreads in the markets that we’re trading. And we can, we can look and know if the spread is working in our favor, we’ll hold a position a little bit longer. Before we spread, you know, research always has us spreading on a certain day, X amount of days before, before personnel to stay and things like that, and rolling. Yeah, rolling. But, you know, in reality, sometimes you can hold on to a position because there might be a near term squeeze, because of some kind of shortage or something. And if you’re, if you’re positioned correctly, you don’t want to just roll the market just because the calendar says why you should be rolling this soon. You know, you can watch that and, and benefit from that. So we do use some discretion, when spread markets between commodities get a little bit stretched.
Jeff Malec 48:42
And from my standpoint, we don’t want any discretion. Like if you’re saying no, right, because it invalidates the the back test. You don’t know when those discretionary moves were made in each of those events? And how much would you have made in October of oh eight, if you left those on? So right the more discretion that’s used the less I can trust the the track record.
David Triplo 49:03
Yeah, I mean, it makes sense and I mean, even with the rolling aspect is like WTI spreads going negative in certain times. So it’s like you have to be very cognizant of what month and how you’re handling your portfolio Well, I can pretty much ask questions all day as is pretty much what I do with CTA so I I’ll go ahead and open it up to the floor for anyone that has questions.
Jeff Malec 49:31
Don’t you guys want us to turn the light show back on? Yeah.
49:38
30 years ago. Racing’s alive you guys I’m targeting 30% Return Are you starting to see now with rates going back up restricts What’s that now guys are just having fun you guys come back at 2530 Because as you guys all know me since I’ve been doing this because institutionalization of trying to find another 30% Guys done, as you know, 20 times the Prime Day. And now it seems like investors are saying they want word.
Brian Proctor 50:10
We haven’t seen that yet. But it could be around the corner. If if rates, you know, they’re fairly stable now. Actually here in the US, I should preface that. I mean, in the UK, they just did kind of a surprise 50 basis point hike. So we know we haven’t had investors coming back to us saying, hey, the risk free rate is 6%. Now we want you guys to shoot for higher returns. But you know, you have to be cognizant of that if you have a strategy that’s only making single digit, you know, eight 9%, and the risk free rate is six, then, you know, then you have to have a conversation with your investors about possibly increasing leverage or those kinds of things. But we haven’t seen that yet.
Joe Kelly 51:02
Yeah, I don’t know if you guys have had a resurgence of fun to fund interest. But you know, when you have a fun, fun netting vol down to three, giving you a one Sharpe versus risk free. It’s a tough business. And so we’ve seen an uptake, family office fund to fund Endowment Foundation that, frankly, were very happy taking all the equity risk in the world over the past 10 years. And now they’re massively under way to do our space. Mercer, you know, is a great partner and you know, in that consultant space, a great group to sort of carry the torch. But it takes time, it takes time for them to come back to what they didn’t believe in for 10 years. And so I think we’re all probably fascinated by a two year sales cycle with risk free rate at six. Even if you’re putting up 1012. You’re far outstripping their low vol portfolios that are netting down even with a good job. So we’re seeing at least a resurgence, if not of people asking for more return, you know, the people that used to be in our space coming back to the space,
Jeff Malec 52:15
who suggested the risk free rate in their sharps in their decks. Anyone? Have any of you adjusted the risk free rate used in your Sharpe ratio in your pitch deck? Yes, it’s good, but probably not as high as Yeah. 5%?
Brian Proctor 52:36
Yeah, I mean, it depends on what window you’re looking at point
Jeff Malec 52:39
five, you went from zero? Yeah. And
52:46
so just the definition track is governed by system and track. Point is a course on trend following the rules of sex in america to become systematic, and the definition is one product products. What makes a difference in strength following multiple products, with a whole bunch of rules based versus just
Jeff Malec 53:12
I think you mean becoming systematic macro?
David Triplo 53:17
Yeah, I mean, I would say for the most part, how Mercer would define it is, every anyone that’s unsystematic macro is 100%, systematic with the few exceptions that they were talking about in terms of typically reducing risk. There’s almost no exceptions to that and Mercer’s universe of how we’re defining what systematic macro is. And we used to call it CTAs. Beforehand, managed futures, all three of them are basically the same thing. But systematic macro is kind of used, because it’s a little bit more all encompassing for a few outliers, so to speak. And that’s when with within our universe, and we have our two subsets, one being kind of trend oriented for more higher convexity profile more often than not, and higher skewness versus the multi strap style.
Jeff Malec 54:03
I’ll take a shot at that. So these guys are all registered CTAs commodity trading advisors, that’s their registration status. The kind of consultants label is managed futures, which is encompasses all that trend following is the primary prime primary strategy within managed futures. I’m just I know you guys do more than pure, but I’m going to use you guys as the example there of EMC more pure trend following now when I get over into Campbell’s area, and I’m doing more carry more multi strat stuff. Now I’m starting to morph into a systematic macro manager that’s doing more than just trend following I’m doing some long equity, I’m doing some carry, I’m doing some other things. So I think, to me, those lines have really blurred in the last three to five years of historically trend followers who have now morphed into doing way more strategies multi strat versus more discretionary macro guys coming down the into the systematic space. And it’s kind of blending in the middle there.
Joe Kelly 55:04
Yeah, it’s a good point. And we’re trying to go upstream and systemize fundamental ideas. So the way we think about it is it’s non price data. So G 20 economic release data is a big thing for us point in time for 30 years, when you’re building economic surprise indices that either you can trade directionally or mean reverting based on revised data. That’s all non price, right. And so that is where we really see things like systematic macro be a better definition, there might be price used in trend, I think universally with some conditioning factors around the environment, flip that on, its on its head and use fundamental inputs, of which there are like, you know, hundreds, if not 1000s, if not hundreds of 1000s or more parameter sets, you can use that or non price to determine a relationship. And maybe a good example is Russia invades Ukraine, generally, you would go long dollar and short, pretty much everything else on a risk off event directionally as a trend follower, our macro system saw the commodity squeeze from the grain complex, and realized that EMF X exporters, were really going to go bid in the face of a risk off environment and EMF X importers were going to get particularly hurt, because you know, they don’t have the economies to afford this squeeze. And so rather than Joseph just going risk off our risk on directionally, our systems actually, both of those are verse dollar, but actually went long those EMF X exporters, short the importers because of the grain squeezed and what that was going to mean to the macro environment. And so it’s just a another different set of inputs that are non price. For us, it’s a combination of directional and relative value. And really, ultimately, the way we try to describe it in front of investors is I should be able to explain to you every single idea in this portfolio, the way a discretionary macro trader could, and we just use computing power to systemize it. Rodrigo, I know you have something,
Rodrigo Gordillo 57:19
I think let’s talk about the evolution of CTAs. Back in the day, it was you had a chart in front of you, and you were basically looking at the charts. And as that line crossed over, you were calling and picking up the phone and making your trades, you know? Well, you can, but as time has gone by that has become less of a, you know, charting thing and more of grabbing a computer, putting that data into zeros and ones. And this is what we do we use zero and one this situation so that we don’t have to look at any screens. The triggers are just happening in the background. So everything that you said with regard to looking at factors that are fundamental price driven, and then codifying them, so that we can we can have a lot more of them in executing. I think what differentiates this category of systematic macro versus fundamental macro is that yes, a lot of these guys that are fundamental macro Are you are CTAs, they are trading similar contracts. But they’re more forward looking. They’re listening to the Fed notes and deciding how to position their trades, they’re less they worry less about that equal risk contribution and have more conviction weightings than they do equal risk contribution weightings. And so there’s there’s some people that are really good at that we’re not, you know, they’re trying to this parse out what’s going to happen based on language and data and grabbing all those parameters and trying to see the future. Whereas what we’re doing is we’re understanding what has happened in the past, given the certain parameters or certain patterns that we’re seeing across all of our multi factors and then executing that religiously. Right so I think there’s room for both if you find the right strong fundamental manager, I don’t have any skill set in finding that manager and, and I was bad at it. So I decided to codify it with my team.
59:17
That doesn’t create a better clarification saying is all based on price data, race data and fundamentals to support the systematic putting on just on a price databases, your stop losses and trend breaking will add impact and will be non discretionary to get at what Brian Paul did not participate in Mexican pesos, right but he had that pain for 85. Then sell off right? How would the systematic pick up any different than a trend follow one really nice trend side? Either way? I mean, it’s got like you do like okay, it’s got the cavity it’s got those things. We’re straight.
Brian Proctor 1:00:03
Well, we in the peso, as an example we have, in our flagship program, we have force systems that are operating. And like I said, each one of them has a different way of getting in and out of trades. So some of them, we designed the system to try to get into a trade and stay in no matter what the volatility is. So, you know, to capture, you know, the most efficient way to capture a trend from point A to point B. So that system is much more accepting of volatility over time. And then we have other systems that look to complement that that have, that are optimized to a different metric, to a Sharpe metric, instead of absolute return metric. So that system is designed to get in and out and back in and back out regularly tried to be nimble. So, you know, we look at each system independently. And, you know, each one has kind of its own unique way of positioning itself in a market that’s trending, consolidating, possibly the end of the trend is beginning to appear. But we could still be long our longest term system, even though it’s it might be at a two month low. And, you know, other CTAs are net short, at that point. So
Jeff Malec 1:01:29
I think that’s the systematic macro would use the rate differential or the GDP print or something to trigger the same trade, and totally agree, yeah, they’re probably both going to be in the same trade, which is why they get lumped together by the investors, right of like, hey, these tend to kind of end up in the same place, they tend to both be positive skew and have those pops in embed periods for traditional investment. So investors tend to lump them together, even though they’re doing different things behind the scenes.
David Triplo 1:01:57
Yeah, we’re definitely seeing more of that, in our just overall CTA space, this MAC MAC was a whole just more people using systematic data. Sometimes it’s more forward looking versus backward looking, depending on what data and how you’re using it as well. So sometimes, maybe half or Plus, I’ve seen in some funds, at least, where they always do correlate, but then there’s definitely others where might have the opposite signal, which is reducing their positions and times where other CTAs that are full price based trend might be doing better without looking at additional factors. So it’s just another way to like diversify your type of trade, instead of just being more focused toward one or the other. And there’s not necessarily a right answer to that
1:02:42
described the volatility in those two, it’s still price data that you’re looking at. Right? I would think that the yields systematic, about the fundamentals will be more like China decoupling, and that’s what I am onshoring. That’s how you pick that up. You know, for picking up that is completely driven outside, looking at volatility, right, you’re gonna have this complete shift that you haven’t seen before.
Joe Kelly 1:03:16
Yeah, they might be trading copper outright. And we might be trading copper based on movement in the Chilean peso. Or we might be trading the Chilean peso based on the move in copper. And so I guess my point is that you can treat each market individually, or from a Mac, the way we think about macro is, and this both directional and relative value, the effect of one market on possibly a completely different market, so not curved trading, the way people used to think about kind of macro, but you know, how does a given interest rate regime affect the equity markets of a given country, and we’ll look at one as the input to trade and other. And so as I said, it’s not just the price of that individual market, the input might be something related economically, that’s driving the behavior, even though we might have the exact same position. I think that’s, as Jeff said, The hard part is you guys all made money last year, how do I value more? How do I distill the quality of those returns? And I think that’s hard to do as an investor.
Rodrigo Gordillo 1:04:26
I’ll give you an example of from a systematic macro. It doesn’t necessarily need to be fundamental macro data, like we trade seasonality, right? So seasonality is looking at historical price data, looking at the seasons and the patterns day or the week, week, or the month, month of the year, etc. And oftentimes that makes a trade that’s complete opposite of what trend is on trend might be saying, hey, long go long gold right now. It’s an amazing trend and our other systems seasonality saying that’s a terrible period of revolt, you shouldn’t be shorting two units of Gold at this point plus we have a mean reversion thing that saying that that’s an A six standard deviation event. So that’s there’s a wide variety of it’s kind of like systematic macros also you can say multi strat, but I think the category has been used a systematic macro to mean things in the CTA space that are outside of just pure trend. And some of them are fundamentally driven. Some of them are just different ways of looking at data,
David Triplo 1:05:23
and just our evolution of like how Mercer’s even called, it has even changed in really five, six years just because of evolution in the space as a whole more fundamental data is being used, more different versions of that stat ARB is sort of becoming part of systematic macros, how we define it, there’s just so much evolution where even we cut out like how we viewed diversified trend, because we used to have a different category of like 50 90% trend was one thing than 90% Plus was pure trend. And now it’s kind of like, how do you define that if you’re looking at price versus fundamental data, that’s still for, that’s still looking at trends or something else. And it just became more for us of looking for one profile and or another in terms of risk return profile, and all the characteristics around that.
Jeff Malec 1:06:14
I got one more you guys want to go drink? Just to your point, Rodrigo. You have the gold trend, you have the seasonality short, and Joe, if you have all these multi strats, how do you protect against just becoming the risk free rate, right, of all these different models doing all these different things? And you’re just you know, the No, no, the opposite of like, you have so much diversification that you just end up with the risk free
Joe Kelly 1:06:40
rate. Yeah, when I when I came in, in 2016, there was a little bit of a an attitude towards anything that diversifying is good in the portfolio. And I would say that there was some lesson learned around, you know, there used to be this thing around model count in the industry, if you had more models, somehow you were better. And I never believed in that. And so over time, we’ve sort of set a bar around saying like, look, in some models, a point five Sharpe is is outstanding, especially if it’s lowly correlated to the rest of the portfolio. More recently, in some of the short term research we’ve done, you know, that David referenced on the on the website, we’re finding negatively correlated models, to other short term models that are north of a point five sharp and so, you know, as an allocator, you have some of the same challenges in terms of how to put forward put a port forward portfolio together and achieve a risk target and a certain return. For us, we just upped the bar on a minimum requirement for a model to have, you know, Sharpe characteristics, correlation characteristics, you know, some models come up to the Investment Committee, and they’re better implementations of what we’re already doing. So that can be a good thing. We try to kick out the ones that aren’t additive earlier in that model development and peer review process. And so it’s been a learning curve, and we’ve definitely, to be fair, you know, at times fell into well, it’s, it’s, you know, uncorrelated, but its return is expected to be 0.0 sharp, and yes, it’s uncorrelated. But that’s like saying, you know, CTAs, are uncorrelated in 2016, and we’re all down 10. And, you know, you’re buying an uncorrelated strategy, but it’s not a good strategy at the time, given the market environment. So setting the bar higher around return expectations, correlation expectations with the current portfolio, and not, you know, we put in four or five new models a year, we kick out two or three. So you know, we’re not in an arms race and on the back of that we have to do a lot of work around cost modeling that Brian referenced when he does a two week look back on volatility that has different metrics, and then you got to think about infrastructure and all the things I mean, it is a it takes, you know, takes an army and it’s not just those pure researchers, you’ve got to have people around them that evolve the process to reflect that and so you know, when you think about evolving your belief systems it’s not it’s not small those lessons have to filter through pretty quickly if you’re gonna adjust over time
1:09:40
what he said that was great.
David Triplo 1:09:45
Anyone else or are also thanks again for everyone for coming today. And yeah, appreciate your time and hope you all learned a lot.
Jeff Malec 1:09:58
Okay, that’s it for the show. Thanks to Cohen & Co, RCM, and Mercer for sponsoring the event. Thanks, David for moderating and acting as the de facto pod host here. Thanks to Jeff Burger for producing – Tune in next week. Peace.
This transcript was compiled automatically via Otter.AI and as such may include typos and errors the artificial intelligence did not pick up correctly.
