In partnership with:
CME Group

89 Making a Simple System and Sticking To It with Robert Carver, Author & Trader – 1of2

"Most people are not as good at trading as they think they are." - Robert Carver (Tweet)

Robert Carver spends most of his book telling you that he is not a good trader. So why did he devote his life and a whole book to the subject?

Because he believes in systematic trading; making a simple trading system and sticking to it.

In this episode, we dive into his new book and explore why simplicity is the key, people think they are better traders than they are, and so much more.

Thanks for listening and let's welcome our guest, Robert Carver.

Subscribe on:

Subscribe in iTunes Subscribe on Stitcher Subscribe on Soundcloud

In This Episode, You'll Learn:

  • Why Robert decided to write his new book, Systematic Trading
  • Robert reads the preface to his new book
  • He explains one of the biggest investing mistakes of his career
  • The book that got him interested in the financial industry

    "From quite a young age, I was fascinated by computers." - Robert Carver (Tweet)

  • How he got a job with AHL

    "From 2006 to 2013 when I left AHL, it was probably the most exciting period to work in financial markets." - Robert Carver (Tweet)

  • Why he wrote a book with very little math involved so that he spoke to a large audience
  • The 3 types of investors that he defines in the book
  • How cognitive biases in humans make them bad traders

    "If you use a system with simple rules, you can overcome those cognitive biases." - Robert Carver (Tweet)

  • Why he spends a lot of his book explaining how bad he is at trading
  • Whether trend following and other divergent strategies work or not
  • Why sticking to a plan is so important

    "You need to make the assumption that the future will be at least a bit like the past." - Robert Carver (Tweet)

  • How the markets and CTAs will change once the interest rate changes
  • The differences between a subjective and an objective system
  • Things that people should avoid when creating a trading system
  • Why few people have a good handle on overtrading
  • Machine learning approaches vs. idea approaches to creating trading programs

    "I personally prefer the ideas-first method to creating trading systems." - Robert Carver (Tweet)

  • How systematic trading adapts

Resources & Links Mentioned in this Episode:

"I try to pick up on human behaviors that have not changed for thousands of years." - Robert Carver (Tweet)

This episode was sponsored by Eurex Exchange:


Connect with Robert Carver:

Buy his book on Amazon.

Follow Robert Carver on Linkedin

"Overconfidence is the fundamental human flaw that effects most people." - Robert Carver (Tweet)

Full Transcript

The following is a full detailed transcript of this conversion. Click here to subscribe to our mailing list, and get full access to our library of downloadable eBook transcripts!


Welcome to Top Traders Unplugged, where my goal is to give you the clarity, confidence and courage you need to invest like or invest with one of the top traders in the world. It is the stories that you never get to hear set out as the most honest and transparent account that I can make of what goes on inside the minds of some of the best investors in the world. Today you're listening to episode 89. If this is your first episode you've heard, you might want to go back and listen to all the earlier conversations. Before we find out who’s on today’s show, I want to mention that today’s podcast is brought to you by the Eurex Exchange. In today’s conversation I actually learn about a brand new contract from my guest that Eurex has just launched, which my guest finds to be very useful for many investors and which you can find out much more about by visiting the Eurex website. 

Today I’m talking to Robert Carver, who’s both an author and a trader and who spent a good chunk of his career at one of the biggest European systematic trading firms, namely AHL, before leaving in 2013 to write his newly released book. By the way, if you want to read a full transcript of today’s episode, just visit the TOPTRADERSUNPLUGGED.COM website and sign up to receive access to all of them. Now let’s get started with part one of my conversation, I hope you will enjoy it. 


Rob, thank you so much for being with us today, I really appreciate your time. 


Thank you for inviting me, Niels. 


My pleasure. Now, Rob you’re a little bit different from my usual guests which have been hedge fund managers or CTAs running their own firms and strategies. Today you are an author, and you trade your own money having left the bustling world of the city of London behind. But in your new book, Systematic Trading, A Unique, New Method of Designing Trading and Investments Systems, I think you bring such an important perspective on the systematic trading world and you break down the components in such an easy to read way that I wanted to make an exception to my usual line up and bring you on for an in-depth conversation on a number of topics relating to this. 

For me it’s a little bit funny because when I look back at all the podcasts interviews that I’ve done so far, it’s precisely one year ago the same exception for Katy Kaminski who had also just published a great book on the world of trend following. So for full disclosure, I have tried to read all of the 300 pages, but I may have skipped a few in order to make today’s deadline. From what I’ve read I really recommend anyone interested in this subject to grab a copy of it, or perhaps two since Christmas is coming up, and for those who think I’m getting paid to say this, you're wrong, this is purely a recommendation based on my assessment of the quality of the book.  

Now many people in the alternative investment industry are very familiar with some of your previous employers, in particular AHL which has been an institution in the CTA space for many decades. Before we jump into your background, I just have a simple question that I try to ask all of my guests in order to appreciate the different answers that there is to this question.  

Basically it goes something like this: When you meet people that don’t know you and they ask what you do, how do you respond? How do you explain what you do? 


That’s a very difficult question to answer and I normally say that either I am a writer or that I’m and independent trader, depending on who I’m talking to and which answer I think will get the best response. But I basically do both of those things, the last thing about trading systematically is, if done properly, once you’ve actually designed your system it takes up very little time and that leaves you more time for thinking and writing, which is actually what I enjoy doing more. That’s what I would describe myself as.  


 Now at this stage, I would normally move on to my usual questions but I want to do something different today since you have just published your book. I can’t think of a better way to kick things off then by setting the scene, and you reading a little bit of the book in order for people to get a sense of your writing style which I very much enjoy. So I would kindly ask you maybe to read a little bit of the preface and then a little bit of the introduction to the book, if you don’t mind. 



“I’m very bad at making financial decisions. Like most people I find it difficult to manage my investments without becoming emotional and behaving irrationally. This is deeply irritating, as I consider myself to be very knowledgeable about finance. I’ve voraciously read the academic literature, done my own detailed research, spent 20 years investing my own money, and nearly a decade managing funds for large institutions. So in theory, I know what I’m doing. In practice, when faced with a decision to buy or sell a stock, things go wrong. Fear and greed wash through my mind, clouding my judgement. Even if I spend weeks researching a company, it’s still hard to click the trade button on my broker’s website. I have to stop myself buying or selling on a whim. Based on nothing more than random newspaper articles or an anonymous blogger's opinion, but then, like you, I’m only human.  

Fortunately, there is a solution. The answer is to fully, or partly, systematize your financial decision making. Creating a trading system removes the emotion and makes it easier to commit to a consistent strategy. I spent many years managing a large portfolio of trading strategies for a systematic hedge fund. Unfortunately, I didn’t have the opportunity to develop and trade systems to look after my personal portfolio. But after leaving the industry, I’ve been able to make my own trading process entirely systematic resulting in significantly better performance.” 

I will now read from the Introduction. 

“It was the 23rd of January in 2009 and I was in my London office. Although I had a desk overlooking the Thames, I was usually too busy to appreciate the view. My day job was managing a portfolio of systematic trading strategies for a large hedge fund. But right now I was focusing on my own bank balance. Data was about to be released indicating how the UK colony had performed in the last three months of 2008. It would be bad news, the official confirmation that we were in recession, but nobody knew how bad. This didn’t mean extra work for me however, since a bank of computers would adjust our clients’ portfolios automatically when the news arrived.  

So I decided to devote some rare free time to trade my own money. With a stressful fulltime job, I was not a particularly active trader. But very occasionally, an opportunity came up that was too good to miss, this was one of them. In my research, I’d found historically, when people’s fears were confirmed by terrible economic numbers it was often the best time to buy and this was potentially the worst news I’d seen in my lifetime.  

Careful analysis showed the banks, hardest hit by the financial crises, should rebound the most if things improved. I was particularly attracted to Barclays. I had traded for their investment bank a few years before and their balance sheet was in relatively good condition. But I also looked at investing in the other major UK banks. In all, I was prepared to risk 10% of my portfolio on the four banking stocks. Then the figures came out - they were worse than expected, with GDP falling by 1.5%. Barclays dropped 15% almost immediately. Taking it to the lowest level I had ever seen. I waited for the market to stabilize and prepared to trade. 

But I hesitated, everything had happened as expected. I should go ahead and buy, but what if this was wrong, what if the financial industry really was imploding as everyone else seemed to think? Panicking I quickly changed my orders, knocking a zero off each one, so that only 1% of my portfolio was at risk. It was one of the biggest mistakes of my investing career.” 


Thanks very much Rob. It’s funny when you read that, it really does take you back to that time of 2009 which was a scary time to be in the investment world. Now this was of course just a little bit of a background and I wanted to ask you, to take us a little bit further back in your own career so we can appreciate and understand how you came about ending up where you are today so if you don’t mind. Take us back even to when you were a kid growing up, what your interests where, were you always curious about things, give us a little bit of color if you don’t mind?  


I think it’s fair to say that from quite a young age I was fascinated with computers and I’m forty-one years old so, I got a computer when I was seven. I was probably one of the first generation to have access to a personal computer from quite a young age. This interest developed to the point where I went to university to study computer science. Which in my naiveté I was assumed would mean that I would be taught how to program computer games, which at the time it was the career that all nerdy boys and my friends wants to do. 

Unfortunately, the course was very dry and scientific and had too much theoretical mathematics. So I didn’t complete that course. I spent a few years working in a completely unrelated industry and then quite by accident came across a copy of Michael Lewis’ famous book, Liar's Poker in the bookshop. I read the book and the interesting thing is that Michael Lewis wrote this book to sort of scare people off from working in finance. But I think it had the opposite effect. I know so many people that said “yeah, I read Liar's Poker and it sounded amazing and I wanted to get into that world and do that.” 


 Yeah, it was a great book. 


So then I became interested in Economics and ended up studying Economics at University, and going back, and this time I managed to complete me course, so obviously I’d found the right thing. I then went to work for Barclays Capital, an Investment Bank, and I spent a couple of years with them trading exotic interest rates options which was an experience, but not one that I would want to repeat. I think it is fair to say that I’m not really cut out for the cut and thrust of a bank trading floor. 


What time period are we talking about now? 


 This was in the early 2000s. So after the boom before things got crazy again. I think also from reading that section of my book you’ll appreciate that I’m not emotionally cut out for making investment decisions. 




 I actually believe that very few people are. Which is why I think systematic trading is the way to go. I’m sure we will discuss that some more. Anyway, after that, rather traumatic experience, I spent a couple of years doing my master’s degree part time whilst also working for an economic think tank which was a more cerebral activity and a bit of a rest from the frantic trading world. Then again, quite by chance (to get ahead in life you have to have a bit of luck), quite by chance, I saw an advertisement in the financial times for AHL. I knew of AHL because I’d actually worked as a summer intern a few years before hand. 

They were looking for someone who had both a financial markets background but also an economics background and also an interest in computer programming. I sat there thinking this offer has been written for me. So I applied for the job and got through the rather tough interview process and I got the job. 


So you joined AHL, when was that, before the crisis? 


 So yes I joined in 2006. So a very different world from the one we live in now. I also feel, I’m not sure if luck is the right word, but I think it’s been an excellent education having spent the period from 2006 to 2013 when I left AHL. Probably the most exciting period to work in the financial markets, obviously the excitement was both good and bad.  


 But the learning was good. 


The learning was excellent. 


 Now, of course you’ve been very busy writing a book and I can only imagine how much effort and time that takes. But when you’re not trading, when you’re not writing, what do you like to do nowadays that sort of takes your mind off those two things? 


I’ve enjoyed doing quite a lot of cycling. One of the reasons why I wanted to stop working in a full time capacity with the two-hour daily commute was also to spend more time with my young family. So those are my other interests. In the summer I also enjoy doing a bit of sailing. So in full disclosure I was actually a world champion in one particular sailing class when I was younger.  


 Wow! There we are, you are my first world champion on the show. 


 That was a long time ago and I think, again in full disclosure, I’m nowhere near as good as I used to be. 


Fantastic, Excellent Rob, now you’ve written a book about Systematic Trading which normally involves a lot of math and equations in the real world. But you’ve managed to write a book with very little math. Being you, why was this important to you to do it that way? 


The credit for this really has to go to my publisher. When I sent the first draft of my first chapter to my publisher and he came back saying, this is great but by the third sentence you lost me. If you want to write a book that only perhaps a few thousand people in the world can understand, then that’s fine, we are prepared to do that. Obviously, there is kind of an equation that publishers use to do with size of audience, cost of book, length of book, publication costs and there are all these things kind of get factored in together.  

So at one extreme you would have a highly specialized option pricing book that might sell, in a good day, a thousand copies over its whole lifetime. That will cost you over three figures to buy. At the other end you might have Investing for Dummies, which is probably going to be ten dollars and could well sell hundreds of thousands or even millions of copies. So he said you have to think about where in the spectrum you want to pitch your book.  

I very much felt that I wanted to speak to as large an audience as possible. Then the publisher came back again and said you are going to struggle I think to bring your ideas down to the level where anyone on the street could understand them. Let’s try and pitch it somewhere in the middle and that’s where I tried to do it.  

I’m not sure how successful I’ve been because there is a couple of reviews in One says, “This is the first book you should buy when you’re thinking about systematic trading.” Another one says, “You should probably read at least ten other books before opening this one.” Opinion differs as to whether it’s as straightforward as you say. I certainly did try to make it as accessible as I could and incorporate maths only when I felt it was absolutely necessary. 


I think you did a great job. By the way, having also put myself into the public light with this podcast, comments and opinions and reviews, they are what they are and completely uncontrollable and whatever people have in their mind when they wrote them, that is how it comes out. I would focus on the good ones. Even though you could have a hundred good ones and one bad one, it’s the one bad one you focus on isn’t it?  


Well that’s true, but in fairness both of those reviews were good reviews. It’s just that they disagreed about how accessible the book was. I am going to just take those as a sampling error issue and assume that on average I’m in the middle which is something that I’m comfortable with. 


 Yeah, it’s a great book, absolutely. Now, I want to start sort of digging into the book a little bit and in the beginning you start out by defining three types of investors, namely the Asset Allocating Investor, the Semi-Automatic Trader and the Staunch Systems Trader. I want you to tell me a little bit about each of them and why it’s important to define them. But I just want to be completely open here and say I actually, I’m not English by background, so I didn’t know what the word staunch meant and so I looked it up. There may be other people on the show today listening and who have no idea. So it comes out as saying “very loyal and committed in attitude.” It has all these things so now at least I learned something new very early on in the book which is what staunch means that you are very committed to your strategies. Anyway, that was a digression here. Let’s go back and you explain a little bit about the three types of investors as you see them. 


 Okay so, again in the interest of making the book more accessible I didn’t want the book to be read about people who just purely wanted to do what was normally thought of as systematic trading. In other words, you have some rules that look at normally technical price patterns, and then in a completely quantitative way decide what positions you should have as a result of looking at those prices and then make the trades, often, automatically. So that is what a Staunch Systems Trader is.  

It is someone that has an end to end process with trading rules at the beginning that are completely systematic, and then a position management framework that translates those into positions and trades and does all the risk management which again which is completely systematic. Then the trading at the end can be automated but the point is that you follow those trades religiously and you never deviate from them. It’s a system that could be completely automated.  

Now that is quite a narrow set of people and not everyone is in a setup where, or a kind of place themselves, where they're comfortable with that. And it depends on whether you’re comfortable with the idea that a relatively simple, hopefully, set of trading rules can actually predict what will happen to prices in the financial markets. Not everyone signs up for that. I kind of identified two other groups of people who are out there. 

The first group are people who think that humans are better than computers at predicting price movements. To be more specific, they think that they personally are better at predicting price movements than simple rules are. This is what I described as the Semi-Automatic Trader.  

The idea behind the Semi-Automatic Trader is someone who still wants the freedom to say, I think that Apple is a good buy, but I should be short Google. But who then wants to take that opinion and put into a systematic position management framework; that will then decide how big their position should be, what size their stops should be, when they should open positions, when they should close positions, how many positions they should have open, how to manage holistically the risk of that.  

I believe that if you have a good position managing framework in place then how good your ability to forecast or how good your systematic trading rules are, if you’re the first kind of investor, is much less important. To the extent that you can actually run the simulations with completely random entries. 

You basically simulate a trader who is no better than flipping a coin. If you then think that into a position management risk framework, and set up correctly, that guy will still make money. That’s partly because in the past we’ve seen trends in markets and if you’re setting up a system where you’ve got fixed stop losses then they will naturally tend to pick up on trends. It’s still an interesting finding. If you then add in trading rules that do a good job of predicting where the market is going to go that does outperformance your system but not as much as you might expect. And so that is the second kind of person. 

Now the third kind of person, is a very miserable cynical kind of person who thinks that no one can predict what will happen to prices in financial markets. I call them the Asset Allocating Investor because of a kind of long term buy and hold mentality. They often say you can’t pick stocks. The best portfolio is to buy a selection of ETFs that give you exposure to different asset classes. Then you basically say, “you know what, I’ve got no idea what these things are going to do I’m just going to buy all of them.” So that’s that perfectly valid point of view. I’ve run part of my own portfolio on that basis. 

Then again, I still think there is value in using a position management framework to say well that’s fine but how should you account for the different risks of the different assets you are buying, how should you account for the correlations? How should you trade that portfolio, how should you rebalance that portfolio given you’ve got a set of costs? So the challenge for me was to create what I call the framework which is the sort of thing in the middle, between either a trading rule, or a qualitative opinion, or a stubborn buy and hold mentality. That takes all of those opinions and kind of processes them the same way and produces sort of positions. That’s what I’ve tried to do. 


Sure, Thanks for doing that. Now Rob many people in sort of the money management business are very focused on explaining how they do what they do in order to convince investors to let them manage part of their money. In my opinion, a more important question is why, so let me ask you why you should start a systematic trading strategy today? 


 It’s because most people are not as good at trading as they think they are. That’s pretty much it. I mean I talk a lot in my book about what behavioral finance people and psychologists call cognitive biases. So essentially our brains are kind of wired in a way that made sense when we were wandering across the plains of Africa a few hundred thousand years ago and trying to hunt whatever we were hunting then.  

When you’re trying to make decisions based on complex information those emotional biases that come through, I mean you end up often doing the wrong thing. Now I know well myself this is true with me the little Barclays anecdote I read out to you. I could read out many, many more of those bad decisions that I’ve made due to emotional problems. I believe these affect nearly all people.  

I think the biggest bias of all is over confidence. People think they are much better than they are and people think they are much better than average than they are. Only someone who thought that they were better, whatever that means, at trading than anyone else, would actually actively trade the financial markets. All the people who are actively trading the markets can’t all be right; there must be some of them that are below average. I believe that if you use the system with relatively simple rules you can actually overcome these biases and even exploit the biases that other people have. 


Now we always have to look at ourselves in and what we do and early on in your book you explain why people should be skeptical about trading systems that you can buy off the shelf or books that you can buy or blog posts that you can read and we all know that there is a lot of that out there. But can you maybe explain why you think your book is different?  


Probably because I spend most of my book explaining how bad I am at trading and how you should be very… you know, for example I say a lot, you really shouldn’t expect a Sharpe ratio of more than - the absolute maximum you should expect is one, and actually for other people you should expect a lot less. And actually you should do your kind of position sizing as if your Sharpe ratio is half what you expected.  

In contrast you know if you’re trying to sell a trading system, the natural human instinct is to probably to do some kind of backtest and fit this thing until you have a Sharpe ratio that looks attractive, and is unlikely to be realizable in real life because of course you over-fitted to get it. It’s very hard, and I have sympathy for people that are in this marketplace, and I’m not saying they are bad people, they are evil people, or they are out to scam everybody, but it’s like being in a market where you selling cars to people and the people that are buying your cars have no way of knowing how good the car is; no way of verifying for example what the top speed of the car is, and you have to say oh my car can go 250 miles an hour because another guy is saying oh my car can do 200 miles an hour. 

When people then buy the car, and then drive it, and it’s a trading system so of course the returns you get are random to an extent. Then the car only does 50 miles an hour or it crashes, and they can’t really complain because of course that’s the way these things work. 

So I think anything that you are reading you have to ask yourself why is the author presenting me this information, why, what’s their motivation for doing it? My motivation for writing my book was obviously to sell books. From a pure economic perspective writing a book and selling it, and getting a percentage royalty, in terms of hourly wage, I’d have to sell quite a lot of copies just to get to the point where I’d be getting the same, as I would be getting working at McDonald’s, for writing. So I’m genuinely interested in educating people and trying to explain to them that they need to be more realistic. That’s a completely different market place from where you are trying to compete with people who are trying to make the most outlandish claims to stand out from the pack of people making similar outlandish claims. 


 I agree with that completely. I think it is evident from reading your book that, that is the fundamental motivation. Now you touched already upon the point about the flawed human brain in your book, and you end up talking a little bit about some of the temptation of taking profits early and letting our losses run, and that’s really how we as human beings are wired. Tell me a little bit more about that and what it really means when it comes to trading if you get this balance the wrong way around so to speak, and what you found when you test this kind of human behavior if I can call it that. 


 So the natural human instinct when you see a position rise in price and your long is to say… is to want to take a profit. And this comes down to essentially that kind of strong feeling and you want to kind of lock that in. And the reason you want to lock that in is that you want to prove that you are right.  

It’s the overwhelming human emotion to prove that you are doing the right thing and it’s called confirmation bias in the literature. Now when it stocks falling, if you sell a loss, then you’re going to be proving that you are wrong and nobody wants to do that. So what you actually then want to do is hang onto that position and hope it goes up in value. And as it keeps falling of course you have the same conversation with yourself until you’re forced to sell and perhaps because you’ve run out of money. 

So it’s really about the way the human brain is treating unrealized losses and unrealized losses differently. We are thinking about them differently. Even though they’re exactly the same. And you know this kind of mindset that it’s not a profit until you’ve sold it and it’s not a loss until you’ve taken the loss. It’s completely wrong.  

Now it’s very easy to think about a sort of pattern of price where it would actually make sense to buy on a small profit, and that would be if the market was trading in a small range. Now the problem is that most of the time markets don’t do that they trend. This isn’t the time for that kind of theological argument about whether trend following is a good thing or a bad thing.  

Certainly in the past people like Winton have done tests over hundreds of years of data where it is available, and markets have in the past exhibited a behavior where they’re trending. So if markets are going to trend and this behavior where you’re going to sell at a small profit and cut only when you’ve got a huge loss is exactly the wrong thing to do. You should do exactly the opposite of that, which is what a trend following system will do. So this is a really good example of where there is a human bias in our brains creating exactly the wrong kind of behavior. You can then write the really simple rule that will not only correct for that bias but will actually exploit it. If other people are doing this, then this trend following system will effectively taking money off them. 


Now of course it kind of goes into the debate there’s also been about different kinds of strategies: convergent strategies versus divergent strategies. We know trend following is a divergent strategy. Then you have a lot of relative value strategies on the convergent side. I guess, I mean I guess part of your conclusion is that you should have a little bit of everything and that’s probably true, but when you did your test… from memory you did a test with these two different rules on 31 futures contracts, well what did you find in that? If you don’t remember I have the finding in front of me. 


Okay well I did find, I think it was 26 out of 31 markets, or 27 out of 31 markets, a very simple rule which took losses early and let profits run. So it wasn’t actually a classical trend following rule. It was something much simpler than that. And it did better in 27 out of 31 markets. You know that’s not a huge surprise because firms that have been trend following futures have been profitable for many decades so it’s not a big surprise. As you say I’m not saying that trend following is the only way to trade. It’s just that this is a really nice example of where a human bias produces a behavior in the market. Which can be exploited by a simple trading rule. There are others on the convergent side as well.  


Yeah, true. Actually you use the phrase simple trading rules. There are two things you highlight also early on in the book, and that’s the importance of having simple trading rules but also the importance of sticking to a plan. Tell me why this is crucial in your opinion. 


Well if you’re not sticking to a plan then you aren’t really trading systematically. The whole point of having a system whether it be the full on what I call the staunch system trader where your running with systematic trading rules and a position management framework or the more qualitative semi-automatic way of your making your own forecasts then putting them kind of binding yourself into this systematic framework to actually trade and manage those positions. The whole point of that is that you gain the benefits you can get from doing that. Which you know you are going to lose if you start meddling with your system and making changes, and this is something that everyone does.  

From the guys, the retail trader who is using an off the shelf charting package; he’s looking at the signals that are coming off of it, saying well I don’t really like that signal, I’ll ignore that one, I’ll do this one, I won’t do that one. But even in the large institutions like AHL, I’m not criticizing them specifically because it happens in all institutions that trade systematically, you still have debates about whether we should override the system or cut the systems risk because of something that is going on in the world. 

Now the key point, of course is, if you’ve got a purely systematic trading system, that you’ve backtested. In the back test all kinds of stuff happened, and no one was there in the backtest to override it, the system just ran and did what it did. Assuming you are comfortable with the backtest and comfortable with its behavior and what trading it was doing in the backtest, then you should really be comfortable with letting that thing run now without interfering with it. Because that’s exactly what happened in the backtest.  

So there are actually a very, very small number of circumstances in which I believe it’s right to meddle with a trading system and to override it. Unfortunately, this is something comes about with long experience. I think the danger is if you are ever in a situation where you are spending too much time looking at what your system is doing and following the financial news and all this kind of stuff, all these things feed into an environment in which you are more likely to try and second guess the system and override it.  

That’s why I don’t sit on my computer all day and watching it trade. I spend a lot of time setting it up so that it’s fully automated and just reports to me when things look like they might be going wrong. But in an institutional setting it’s much harder when you’re trading with other people’s money because you have this fiduciary duty to look after their money, and as you know yourself, and if something happens and you don’t override the system then there’s always that question of what was the right thing to do. So it’s a culture in which it is much hard to stick exactly to what the system is doing. 


 True, I mean it is interesting when you or I were talking in November of 2015, and actually I would say the things you just touched upon is very real right now in the debate of investors, which obviously I spend a lot of time talking to. People are worried that the coming changes in interest rate environment - meaning we’ve gone from a bull market in bonds to a bear market, at some point, when the interest rate cycle turns - which of course we know the U.S. Central Bank has alluded to now a few times this year already and at some point it probably will come.  

There is definitely fear out there that all these track records that we have been able to produce and can document and show, they’re not going to be worth a lot when interest rates suddenly start going up because not many CTAs have traded through a rising interest rate environment. So there is this fear, that oh it’s going to stop working and of course there is a point to it because most testing will have been done on data from the last thirty years, three decades. Not a lot of people and not a lot of data is available going further back. I want to talk about something related to this in a second, but I just want to hear your initial reaction to this kind of concern that investors clearly have. 


So my last job at AHL was managing the fixed income portfolio. I spent an awful lot of time thinking about exactly this problem. And I came to a number of conclusions. The first conclusion is people often forget that what makes… we say we are looking at prices and we assume that if a bond price falls that we are going to lose money, but actually what we are exposed to is total return. 

So if you’re owning bonds, then your total return is going to come (actually this applies to all futures,) is going to come from both the movement in the spot price and also any carry or roll down that you’re getting. On the carrier roll down, essentially, is sort of telling you what the market expects will happen to the spot price over whatever period it is. So what that means in practice, if you are in an environment where interest rates are very low but expected to go up then the interest rate curve and yield curve will be quite steeply upward sloping. And that means that the carry that you’d get on earning Bonds further out in the maturity space will be relatively high. 

So in a nutshell if the interest rate moves in the way that the forward price is expected it will move you won’t actually make or lose any money. It’s only if the rates change unexpectedly, so if they rise too early or too fast, that you will lose money. So that’s the first thing to say. Certain people kind of think, a lot of people miss that. We did a lot of simulations and tests and looking at different interest rate environments and came to the conclusion that there wasn’t as much of a problem as you might think. 

The second thing is diversify - diversification - If you’re running a CTA and 40% of your assets are in U.S. bond futures then you are some kind of crazy guy. This is true regardless of what you think Janet Yellen is going to do. You should have a diversified portfolio. So probably, when I look at my own portfolio, perhaps 20-25% is in bond futures. And if it was more than 30% I’d be sort of thinking well that seems a bit high. Regardless of what I think is going to happen to interest rates, it just seems quite high given all the asset classes that are out there in the CTA space. Why do you have so much money in just one county? Of course, all bond prices will react to what happens in the U.S. What happens in the U.S. will be the most significant thing, so if you have a reasonably diverse portfolio then your exposure to anything unexpected happening in the U.S. should be relative small. 

The third thing to say is we were having this debate for the best part of three years, before I left AHL. It’s now 2015 so this is a debate we’ve been having for five years. And if you’d done any kind of meddling in that period like reducing your exposure to fixed income you would have been hurt, because being long fixed income and also trend following fixed income has been one of the greatest traits over the last five years. 

For example, last year was an excellent year for CTA’s and most of them made most of their money in bonds, actually mostly in European bonds. If you cut your exposure in bonds too much to say just 10% you’d have seriously missed a lot of that return. So my message really is don’t panic.  

You can go back to probably the last time we had an interest rate rise that kind of panicked the markets in a similar way was in 1994. For example, Orange County happened; a lot of people really got caught short. So you can kind of go back and look at CTAs trading back then or look at backtests simulations, in as much as they can be trusted and look at what happened then. So it’s not like there isn’t any data at all. You’ll see losses, of course, but they should be too large. It’s obvious you haven’t exposed yourself too much to one asset class in one country. 


And the other thing I would add, I feel all your points are very good and of course one can add one more and that is systematic traders today can be short the bonds as easy as they can be long, so there should be no bias there. What is interesting to me is that there are a few, maybe a handful, of these managers who were around in the last interest rate hiking cycle. From 1977 to 1981, for example, interest rates went up dramatically and I can see in the firm I worked for, which happened to be around back then, it was a very profitable period for this kind of trading. 

So to me at least, it looks to me that when the bigger interest rate cycle is up then I think there should be good opportunities. But when people refer to the period as you may mention, 1994, which was a difficult year for the CTA space, well actually what it was it was a correction in interest rates going up in a much bigger down move. To me it was more of a counter trend situation then it was the fact that the interest rate cycle had turned it hadn’t really turned it was just correcting against the bigger turn. 

Anyway, let that be for a minute. I’m going to give you a little bit of a pause to drink some tea because I want you to… I want to point out something that related to what you mentioned earlier which was the importance of sticking to a plan. I happen just to be sent, the other day, a link to an article where AQRs - for those that don’t know is one of the very big firms in our business. The founder of AQR, Cliff Asness, was recently interviewed by Bloomberg and this is what he said about investment success and also in relation to Warren Buffett. 

He basically said that genius is still good, but more and more I think about doing something reasonable that makes sense and then sticking to it with incredible fortitude through the tough times. Maybe I muddled that up a bit in my reading of it, but basically what he is saying is sticking to plan is possibly more important than just being a genius. 

Now the other thing he goes on to say when talking about “greatest investor in the world” Warren Buffett he says that (this is about a study that was done about him), of course they found he was fantastic, but not quite as fantastic. His track record was phenomenal but human phenomenal what was beyond human was him sticking with it for 35 years and rarely if ever, rarely retreating from it. 

So it goes very much to the point you made before now of course at this stage I have to do a little bit of selfless promotion and that is to put things into perspective a little bit and mention that, the founder of the firm I work for Bill Dunn, who has essentially been running his investment strategy for 41 years with an annual return of more than 15%, which puts him, along with our new ownership of Marty Bergin right up with Warren Buffett. Yet you will never hear these kind of rosy descriptions in the media like they describe Buffett’s achievements. Anyway that’s a little bit of us being sidetracked. The sticking to a plan and being exceptionally disciplined over the long run I think it is such an important point to for people to realize. 


Absolutely and I think I completely agree with Cliff to the extent that I don’t think that I am a genius, and therefore sticking to a simple plan is absolutely the right thing for me to do. 


Now let’s talk a little bit about the difference between a subjective and an objective system and why objective systems in some ways are better. Can you talk a little bit about that because you mentioned it in your book as well? 


Sure, I probably start by telling you that often when I bump into people who are in trading. They’ll say, “Oh you do technical analysis or are you a fundamentalist trader?” I say, “Well mainly I’m doing technical analysis because I’m using a system that just uses price data.” Then they either get very excited and start talking to me about double bottoms and horses heads and golden crosses, or they kind of back away and make the sign of the cross to ward off the evil. Depending on which camp they are in. 

So there is this assumption that the technical analysis is this obscure almost voodoo like profession, where you stare at lines on a screen and then make a judgement as to what is going to happen. Now all these kind of classic technical analysis methods which I would describe as pattern matching, because generally speaking there is a pattern that you have in your head and your looking to see that pattern and then you make a decision on that. To me they're all extremely subjective. Generally speaking, you can’t write an algorithm that will identify those patterns. If you try and do so you usually won’t end up with a system that is profitable in the backtest.  

So they’re subjective. They’re subjective systems, there is no rules you could write down, there is no computer program you can write to identify them. They are purely in the eye of the beholder. There may well be, there probably are actually people who can do this, who can look at charts and see these patterns and make decisions and make money. That’s not the world of systematic trading.  

To be trading systematically you need a purely objective way of identifying, given some data, what you position should be. That means can’t use all these weird and wonderful patterns, but can use things like moving average crossovers. You can use things like break outside of a range, assuming that you can identify the range with a purely objective system. 

An objective system has a number of advantages. First of all, you need to make one assumption, which is that the future will be at least a bit like the past. If that is true, then you can back test them and then if the future is like the past then you will have a system that will definitely make money. There is that repeatability there. You can also analyze the properties of the system - the risk properties, the leverage, how fast it trades and you use all these things in your system design, which you can’t really do with a subjective system. To me, at least, anyone who says they are trading systematically but then couldn’t in theory write a computer program that would essentially replace them isn’t really trading systematically they’re doing something different. 


 And then of course there is the whole element of trust - that you have to be able to trust your system. You have a great quote in your book where you say something like, “A system which is fully automated but not completely trusted is potentially lethal.”  


Yes, that is true. I mean there has been a few high profile cases over the last few years of normally high frequency trading firms who’ve had a problem with their software or something unexpected has happened. Of course, they have then lost a lot of money very quickly, because that’s the problem with high frequency trading. If you do something crazy like buy high and sell low you know a thousand times a second you can get through a lot of money very quickly. 

It’s not just enough to, if you got a black box on your desk and it's running and it’s doing its thing you need to be completely confident in what it’s doing otherwise when it starts doing something that’s slightly unexpected you’re immediately going to want to change it or meddle with it. That’s why I have a huge preference for systems that are as simple as possible.  

If a system is simple, you’ll know that, given something has happened in the market today, it should be buying or selling. If it does that you can kind of be relaxed. Very occasionally it won’t, then you can investigate and hopefully there will be a valid reason for that. You should have something that 99 days out of 100 it’s doing exactly what you would expect. And if you got something that is very complicated, then it’s got a lot none in the out unit and that’s much harder to do.  


Let’s shift gears a little bit and talk about the things you need to avoid when creating a trading system. Because this is actually something that I often meet when I sit down with investors because they want to know how you, as systematic fund manager, avoid these things. The things, the three no-no’s that I think you wrote about are over-fitting, over-trading and over-betting are the buzzwords that you use. Maybe you talk a little bit to each and why these are things you really do need to stay away from.  


 Okay, over-fitting is a term most people are familiar with although some people use the term curve fitting. So this is the situation where you are developing your trading system and you naturally want your trading system to look as good as possible. So you have a couple of ways of achieving that. One way of achieving that is just to try lots of things until you find something that looks good enough. So you know you might think, well I’m going to try a moving average crossover. You put it in and it does OK. You think, well maybe it looks like for certain situations it doesn’t do so well so now I’m going to add a layer to it and change it slightly so it does a bit better in those situations, and you trade this a few time and you end up with something that’s really good.  

And the second one is a more quantitative way where you do that process but essentially an automated way, so you do some kind automated fitting you have a number of parameters that describe how the trading system behaves and then you kind of search automatically for the combination of those parameters that produce the best performance. 

It doesn’t really… There is no real fundamental difference between these two procedures. They’re both incredibly stupid and dangerous. It is probably that one is slightly more respectable than the other. So there are a couple of motivations as to why people do this. One is - it seems harsh to call it greed but essentially it’s a desire to have a higher backtest performance than is perhaps realistic. That might be because you know you are working in an institution and they might say don’t bother coming to me unless you’ve got a backtest with a Sharpe of at least one. And that might not be realistic depending on a combination of assets and the style of trading you are doing. So you get push and push until you get to that.  

The second reason is I guess comes down to overconfidence. I think that is a fundamental human flow that affects most people. Most people think that they should be able to get a good backtest and they should be able to make a lot more money than is perhaps realistic. So for those two reasons, over 15 years, it’s rife. It’s really hard to get away from and there are probably three sorts of ways about the fitting if I can just go on a bit more. 

The first one is what I call explicit overfitting. So that’s where you’ve got parameters varying in some automated way. The other way, which is a bit more insidious, is implicit overfitting and that’s where you’re not doing a formal fitting process, but you at least once look at the results of your backtest and then make some kind of change. It might be something as simple as saying, “I’m not going to run the system on the US ten year, because that doesn’t do as well as the U.S. five year. 

The third way of over-fitting is what I call tacit over-fitting. And that is where you only ever try things that you already know will work. And there is absolutely no way to get away from this. Because you are not going to probably sit down at a computer and develop a trading system without having done some research or having no inkling about the industry.  

What that means if you plopped me down in front of a computer tomorrow and with a completely blank sheet and said, “I want you to build a trading system.” the first thing I would do instinctively is probably test some rules that were trend following. I wouldn’t test rules that weren’t trend following because that weren’t trend with a minus sign in front if you like, because I know instinctively and already that they don’t work. That’s a form of over-fitting even though I’ve done no, I’ve not even looked to an account curve, I’ve done no statistical optimization, it’s just come from inside my head. And so maybe on to over-trading? 


 Sure, le-t’s talk about that.  


So overtrading, I think, is the Cinderella of these three problems. It’s the one that gets the least attention. Most people know about over-fitting, and most people know about over-betting. I think relatively few people have a good handle on how fast their trading system is likely to trade and what proportion of their likely performance is going to be eaten up by costs. And perhaps also, not many people have a realistic idea of how much it actually costs to trade. I mean it’s very each to say, “Oh my commission is going to be whatever, $1.00 per lot,” whatever it is. Obviously, when then you go and execute you’re going have to pay perhaps half the spread, if you just cross the spread straight away or perhaps more or perhaps less depending on how you do your execution - how large your orders are. That is hard to quantify.  

So I’m completely obsessed with trading costs. One reason for that is they are a relatively stable and controllable thing to look at, so you know there is not really a lot of point in getting excited about your day to day performance. That's essentially a rounded number, and you hope over a long period of time that will average a positive number. But your trading costs are much less variable and you can analyze them with relatively small amounts of data, and say, “Well, am I’m paying too much, am I’m paying too little.” This is an area I think most people just don’t even think about. 

If I look at the number of people who are proposing that you should day trade, you know trade multiple times per day and do so with financial instruments that are relatively expensive like Spread Bets, OTC retail FX. I run the numbers and I can’t make those systems work on the basis where you’re trading multiple times per day because they are just so expensive to trade. It’s only really with a few relatively cheap futures markets that I can trade up quickly. So that’s the second mistake that I think is really endemic. 

The third mistake, over-betting, is something that most people kind of know exists, which essentially is just, very simply, taking on too much risk. So we can measure risk in many different ways but I use the expected average annual standard deviation of returns. If you look at the institutional CTA space, that number will be around %15 to %20, %25 perhaps. Now, if you look at the trading systems in a lot of books or web sites, or that people claim to be running, that number can be in the hundreds of percent a year.  

So they are running with 5 or 10 times as much risk as what institutional managers would consider to be prudent. And that’s clearly to me way too high. The only way you can justify that is if you have extremely high expectations of what your returns will be. So again it all comes back to over confidence.  


 Now Rob, we’ve hit the one-hour mark and I hope the listeners are making lots of notes from what you’re saying because it’s very valuable. But I just want to let everyone know that actually right now we’ve only covered chapter one of your book. So that’s 25 pages out of 300, so if we are going at this rate we going to have probably another world record for you and that is going to be the longest podcast episode ever produced.  

So I am sure we’re going to have to focus on some few headlines from the book, because I also want to try and put it into the context of some of the usual questions that we have. This is great stuff and very, very interesting indeed. 

Now in Chapter 2 you go into more of the trading rules and what makes a good rule, and also you talk a little bit about the Sharpe ratio, which you’ve already touched upon - what’s realistic and much, much more, so let’s jump into some of these ideas. You talked a little bit about it already but when you come up with a trading rule there are kind of two approaches I guess. There one where you come up with the idea first, or you look at the data first. Tell me a little bit about that and what people should be aware of in this instance. 


 Yes So I call these two approaches ideas first and data first. So I think a lot of people assume that systematic traders all use data first. So a classic data first approach would be if you used a machine learning algorithm, where you get a huge amount of data and you dump it into this algorithm and out magically comes the trading rules we should be using.  

So that is one approach. The second approach is ideas first and that would be where you’d an idea then you’d create the rule that sort of expresses that idea, and you then back test that rule in a similar way. So these approaches are both kind of equally valid they have their advantages and their disadvantages. I personally prefer the ideas first method and that might be because of my background. I have two degrees in Economics so I’m more used to thinking in terms of how the world works and that there are these kind of underlying economic drivers driving asset turns. That’s what we should be trying to capture.  

It’s also because I’ve seen more kind of misuse of the data first method and all kind of overfitting. So you know you have to be careful of overfitting in both cases. It manifests itself slightly differently in each case. I’ve seen… I think people see data first as a kind of magic bullet. You get people coming along who you say, “Well I’m going to just use this big data math-fit which is the buzz word now, and I’m going to discover something that no one else has discovered before and will be incredibly profitable. I’ve got a big problem with that line of reasoning. I think it’s possible, and there are firms that probably do use network machine learning, big data successfully, but it’s probably much harder than you think it is.  


 I like another quote you have in your book, or another sentence you have in your book where you say, “Profitable trading comes out of careful research done by thoughtful and knowledgeable people who seek to understand where their profits come from.” I think it goes to the point about the ideas first, that you have to know why it does what it does, so to speak. 


For me that provides an actual air of validation because you know I like to know why a rule is profitable because it gives me more of a conviction that it might be profitable in the future. And also because it gives me an idea of what the risks might be around it. So what was it exposed to? Let’s take an example which is probably close to your heart, which is the devaluation of the Swiss Franc in January.  


Oh yeah, I remember that. 


So it would have been very easy to have run a data first approach and look to the Euro/Suisse FX rate and say, “Oh this thing always stays in this narrow corridor, between 120 and 125,” or whatever it was. All we need to do is create a training rule that essentially does that mean reversion between these two levels and it will make a lot of money. 

And it would have made a lot of money. In about 5 minutes it would have lost thousands of times its profits in one go. Whereas an idea’s first approach would be to say, “OK well, you know that there are situations in which Central Banks want to keep their currency depressed and therefore a carry law should work because… The classic of course the Japanese Central Bank, for decades now, has wanted to keep the Yen as low as possible. That’s why funding in the Yen in a carry trade has generally worked - that’s the reason why that trade works. 

If the Japanese Central Bank then turned around and said, “Well we’re not going to do this anymore,” then you may well want to question whether you should carry on doing that role. The point is at least you can have that conversation. You understand what you are doing, whereas if you just look to some data and said well this thing goes up and down with regularity, without understanding the reason why which is that the Swiss Central bank is desperately trying to keep the rate from breaking that threshold. If you had understood that rather than just looking in the data first and going no further, then you hopefully would have realized that doing this trade was an incredibly stupid thing to do. 


 Sure very true. Now some people may have observed and rightly so that we live in a world of constant change. They may wonder – systems, how can they cope with sort of an ever changing world? Can you explain that a little bit in your usual good educational manner so that people don’t confuse systems with being something that is completely static and needs constant change in order to adapt? How does systematic trading adapt? 


So a key question here is what kind of trading system do you have? So if you are running something that’s trading very quickly then you probably will, and will want to have it adapting on a fairly regular basis. That is because, generally speaking, the short term behavior of markets - what the economists would call a market micro-structure, is something that’s changing fairly constantly and fairly quickly.  

So a high frequency trading rule that worked well a year ago may well not work anymore this year. Now the good news is because you are trading so quickly you have an awful lot of data points that you can look at. You can actually look at six months of data, and six months of data may be terabytes or petabytes of data that you can analyze and come up with the new kind of pattern or system that’s in the markets.  

Now at the other end of the spectrum is the kind trading that I do where I’m essentially trying to pick up on what I think are mainly human behaviors that probably haven’t changed for hundreds of thousands of years. And in that situation you probably aren’t going to want to ever change your system unless you have to. So you are going to want as much data as you can to fit your system So I’ll go back to Winton which is a big CTA, actually the biggest CTA still, tried to get hundreds of years of data to fit or at least to validate that their systems worked. So it depends where you are on those two extremes. One view is to say, “Yes the world is changing but if my trading system is a relatively fast trading system then I can and should adapt it to cover that change.” 

The other extreme, is to say, “Yes the world is changing but there are certain things about the world that I believe will not change, or at least will change incredibly slowly and therefore I do not need to adapt my system.” So it’s a question of where your trading system sits on that continuum. Then you should act appropriately. 

Appropriately means, for example, if you have got a system that is trading relatively slowly you probably shouldn’t be trying to adapt it every single year to the new world that exists, because you don’t have enough data to actually tell you, statistically, that the world appears to have changed. You should be using 30 years of data at least if you can. If you are using 31 instead of the 30 it’s not really going to change the parameters that are coming out of your fitting prices very much at all. 


 Now I have a question, I sense from you that you have a lot of experience in sort of the trend following and you seem to have certainly a liking for trend following, amongst other things. You also talk about trend following and lottery tickets, and in both cases you know that there are small losses waiting, but there are some few much, much bigger winners. Now we as human beings we love to play the lottery. But I’ve found very few investors in my 25-year career that use the word love in the same sentence as trend following. So do you have an explanation as to why that is when these two things something in common yet they seem to be received very differently? 


 Yes it is an interesting problem. You are right, people have a strong preference for trading systems that generally make money every month and then every now and then lose a lot of money which is the exact opposite to trend following. I think I’ve a lot of emotional kind of bonding with these people. I feel that I really understand their pain because when I look at the returns on my own system inevitably, if I’m in a drawdown I feel more unhappy than if it is doing well.  

And if I have a day where lost money even though I know, from an intellectual perspective, that this is just a random number - a draw from some unknown distribution of returns that hope has a positive mean. When I have a down day I feel slightly less happy than if I have an up day. If I had an up day every day for six months, then I’m much happier and I have a lot of sympathy for that point of view I think the reason for the dichotomy between this emotional response and the emotional response to buying lottery tickets is to do with the size the pain and the size of the pay off.  

So if you’ve invested all of your money into a trend following fund, and you lose money for two-thirds (let’s take an extreme example), 2 out of every 3 days you are losing money, you could be losing maybe half a percent of your net worth 2 out of 3 days. No one’s probably going to buy that many lottery tickets. If I was to buy enough lottery tickets that would represent half a percent of my net worth that would be quite a few lottery tickets.… 


Ready to learn more about the world's top traders? Go to TOPTRADERSUNPLUGGED.COM and sign up to receive the full transcripts of the first 10 episodes of the show and visit the show notes where you can find useful links to other amazing resources. Thanks for listening and we'll see you on the next episode of Top Traders Unplugged.