In partnership with:
CME Group
Eurex

15 Model Decay and How Best to Handle It with Karsten Schroeder of Amplitude Capital – 1of2

"When you build short term trading models, it's all green-field research." - Karsten Schroeder (Tweet)

Through courage and vision, Karsten Schroeder co-founded Amplitude Capital as a pioneer in the CTA industry.

Why pioneers? Because they focused on short term trading.

In this episode of Top Traders Unplugged, Karsten and Niels discuss Amplitude Capital's scientific approach to Amplitude Dynamic and Amplitude Klassik. These are the two short-term rule based trading programs that Karsten and his team run to invest billions of dollars on behalf on a small group of institutional investors.

Thank you very much for listening to this episode with the Executive Chairman of Amplitude Capital, Karsten Schroeder.

In This Episode, You'll Learn:

  • The founding story of Amplitude Capital
  • About the motivation source for choosing short term trading
  • On the essence of a great companyThe Team

"There is a negative notion to the word 'change', that's why we prefer the word upgrade. Upgrade means you do, more or less, the same that you were doing before but in a better way." - Karsten Schroeder (Tweet)

  • Describing Amplitude Capitals' Trading Programs:
  1. Amplitude Dynamic - A two day holding period, short term CTA program that trades all asset classes (currently trading app. $1 billion.)
  2. Amplitude Klassik - 8 day holding period program, with less reversion models than Dynamic (currently trading app. $600 million.)
  • The scientific processes guiding Amplitude's perception of the markets
  • How Amplitude manages their in-house and outsourced business processes

"As important for an investor to choose the right fund, as important it is for the fund to choose the right investors." - Karsten Schroeder (Tweet)

  • Where the point of optimal capital under management is for Amplitude Capital
  • A bird's eye view of their historical track record and their reaction to the market shift in 2009
  • On the effects of quantitative easing and other government interventions in market health

"When you evaluate businesses today, I think it's so much more down to the team than to the original idea." - Karsten Schroeder (Tweet)

  • Model decay and how to best deal with it
  • The design structure of Amplitude Capital's Programs
  • Market dynamics and where the Amplitude programs trade

"Having markets that face interventions is not healthy." - Karsten Schroeder (Tweet)

  • Differing philosophies: market specific models vs. models for all markets
  • Comparing mean reversion models (counter trend models) vs. trend following models
  • Plus much, much more...

Resources & Links Mentioned in this Episode:


Karsten's also participated in the Battle of the Quants

 

Sponsored by Swiss Financial Services and Saxo Bank:

Saxo Bank Sponsor of Top Traders Unplugged

Connect with Amplitude Capital:

Visit the Website: www.ampcap.com

Call Amplitude Capital: ++41-41-747 15 00

E-Mail Amplitude Capital: z@ampcap.com

Follow Karsten Schroeder on Linkedin

"The first couple of years were really, really tough because I remember I was doing so many meetings pitching this to all types of investors around the globe." - Karsten Schroeder (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!

Niels

Welcome to another episode of Top Traders Unplugged. Thanks so much for tuning in today. I really do appreciate it. On today's show I'm talking to Karsten Schroeder, co-founder and Executive Chairman of Amplitude Capital. Amplitude Capital is one of the most successful short term CTAs in the world, and has received many international awards for top performance and excellence in their field. Karsten talks in detail of the amazing journey over the past 10 years, and about the unique philosophy and culture that Amplitude has been built upon. For those of you who are new to the show, I just want to let you know that you can find all of the show notes, including a full transcript of today's episode on the TOPTRADERSUNPLUGGED.COM web site. Now let's get started with part one of my conversation. I hope you will enjoy it. 

Karsten, thank you so much for being with us today. I really appreciate you taking the time. 

Karsten

Thank you. Thank you for having me Niels. 

Niels

No problem, I think it's fair to say that when you founded Amplitude in 2004, you founded not only a new CTA firm, but probably one of the first short term CTAs in Europe, and one of only very view short term oriented CTAs in the world, which in my mind takes a lot of courage and vision. You also came to the CTA industry, if we call it that, with a deep academic, or should I say, scientific approach to research and product design, which compared to many of the US based legends, was somewhat different. Now, ten years later, the vision has been rewarded with many international awards for top performance and excellence in your field. So your vision clearly became a worldwide success, but before we go into the details about the company today, I would really appreciate it if you would take us all the way back to the beginning, telling us your story, and what led you to having the vision, the courage, to pursue this career path, and feel free to go back as far as you want since I know you have an interesting background. 

Karsten

Yeah, thank you Niels. It's been, as you said, it's been an interesting journey without any doubt, and I guess if we put ourselves back ten years ago, we would not have seen that ten years later - a decade later, we would be where we are now. I guess that's probably very common for every entrepreneur who will start out - who will certainly have great inspirations and create vision but, you know at the end of the day, it's so far away at that point in time that you don't even dare to dream about it and you're also so concerned with the tremendous amount of everyday problems.  

The reason why we started Amplitude was because we were all very passionate about trading, we were all very passionate about building models, together with Peter Voss, a good friend of mine who I got to know during my military service time back in Germany, and Steffen Bendel who I went to school with, we started working on Amplitude already a couple of years before we actually set it up in London, back in 2004, building our own trading models and just exploring the markets, and seeing what's possible there. It's interesting that you made the remark that it is somewhat of a difficult venture to build up a short term trading CTA, in that we were certainly one of the first firms to attempt to do so back in 2004. It kind of came naturally to us because the way it all started...back in 2002 Eurex back then DTB Deutcshe Temin Boerse was releasing their first tick data...historical tick data packages, which we bought, on some of the major contracts and started to build models on it. We had software which would allow us to run optimization routines on it.  That was the time when we actually figured that most of the optimal solutions would lie in the short term space.  Of course, back then, we were less concerned with execution costs and capacity and what-have-you because those were not the problems we were facing back then. We were basically all just looking at generating maximum return, and to that extent we were, by no means, pricing slippage numbers that we had to deal with in due course, or at a later stage in our careers. Back then we saw the best solutions were all in the intra-day field and that's why we ended up being a short term trader. So it's not like we said, "alright, we make a market analysis and there are the Winton’s’ and the BlueTrends and the MAN AHL, Transtrend’s, and all those firms that were already around, and we want to be somebody different, and we want to sit in a different spot there, so we'll just go and build a short term CTA". No, we basically built our models and that's what we came out with.  

Then back in 2004...end of 2003 the beginning of 2004, I said to Steffen and to Peter, "I think we have to try to turn this, (and by that I mean our trading ideas and what we had) into a proper business, and we were looking for partners, for people who would support us, and we found Shamil who was just phasing out of Deutsche Bank at that point in time in London, and we presented the strategy to him, and to this day I don't know what gave him the conviction to go with us, because in all fairness, I only had two and half years of working experience, generally speaking, and that was in a consultancy firm, because I was with McKinsey back then. We presented the stuff to him and, I have to say, if I would have been in his shoes, I would have never, never, never done this. At the end of the day, even when you evaluate businesses today, I think it's so much more down to the team than to the original idea. I think if you have a dedicated and enthusiastic team, you can even make a very mediocre idea work. And that's more a general statement. Maybe it's not so true for financial markets, but in the world, if you have a business idea, even if it's not the most brilliant idea, as long as people stick with it and have that passion about it, then I think that most likely they are going to make it work, while if it's the other way around, where people just say, "oh, where can I make money?"  And maybe they come up with something great, but then there are so many difficulties on your way to ultimate success, that people would get frustrated about it. I think because we didn't have that, and also there was not much to lose for all three of us, we being fairly young and coming to London we had no money, no financial resources, and lived in very modest circumstances, so we were basically dedicating all of our time to building up this business.  

The first couple of years were really, really tough because I remember I was doing so many meetings pitching this to all types of investors around the globe, this flying economy, having very basic logistics, and just an incredible high density of meetings back then, and just doing a numbers game, and eventually we found the first couple of investors who entrusted us with their money, and so we could build up a little bit of an asset base. And for the first...if I remember correctly, for the first two years after we launched the fund that the assets were below 40 million dollars, certainly nothing that you could run a profitable business with, and then within a year, within 2007 - basically from mid-2007 to mid-2008 it totally turned around, so we grew assets pretty aggressively and we soft closed and yeah, pretty quickly hot closed our flagship fund, back then Amplitude Dynamic, which is the two day holding period program. That, more or less, came overnight, to be honest, and we'd been dreaming about this for so long, and times had been so uncertain and so difficult and so now, eventually we got there. So that was a very rewarding kind of situation, I would have to say. 

Niels

Sure, sure, it's a fantastic story Karsten. I want to take you a little bit back again, before we leave that, because I'm interested in knowing what it was that you saw in the short term space, in the sense that, computer power at the time obviously wasn't as great as it is today, and so what was it, what kind of intuition did you have that made you decide that here's some really good opportunities, if you remember? 

Karsten

Well, as I said, back then when you were running a data set and a sharp optimizing routine, the optimum solutions would come out in the short term trading...in the short holding period area, and that, of course, can change completely if you would feed in costs, so the more costs you would feed in, which means the more assets you have to manage, then obviously you would arrive at totally different strategies. Don't forget back then, when...you know, it was just about like the super short term game, which was just about to start, the data you were working on, it was not really penetrated by a lot of short term trading activities, so far, so, and the markets were very healthy. By that I mean they were free of any interventions. We had healthy trends. We had good movements. It was a good market environment, so short term trading, back then...the first couple of years were really good, it was a really good market environment. 

Niels

Sure, and just out of curiosity, the trading that you did, and that your partners did before even starting this, was that also in the short term, or was it just a completely different type of trading until you started you research? 

Karsten

So very early on...like the proprietary trading that we were doing, was more on a discretionary basis, and then we built the first models, and that was more, I would say more medium term. Then when we built...when we ultimately built these models that we then pitched, we focused more, actually, on the model building activities, in particular towards the end of 2003 and beginning of 2004, and then when we had Shamil on board, we were really concerned with building up the infrastructure, building up the office, and I think the prop trading before launching the fund, which happened in June 2005, that prop trading, then only started in November or December, so we had a half a year of prop trading activities beforehand. 

Niels

Sure, sure, and could you give me just a brief overview for now, at least, of the programs you run today, and where they are in terms of AUM? 

Karsten

Yeah, we basically have two programs. We have Amplitude Dynamic, which is the one I was describing in this brief history, which is a two day holding period, short term CTA program, that trades all asset classes, and it currently trades around a billion dollars. Then we have Klassik, which is an eight day holding period, and the program has around $600 million in it. We have more reversionary elements in Dynamic than we have in Klassik. With regard to the underlying models, it's pretty hard to describe. When you build short term trading models it's all greenfield research, I mean you may use some components from indicators that you would know from text books, but at the end of the day, it's all in house developed, and it has become, now-a-days, the systems have become really, really complex...but I guess we will touch on that. 

Niels

Absolutely, yeah, we will certainly go into that, because I want to ask a slightly different question, maybe a question that you don't get asked so many times when people come to visit you. I know that, or at least my impression is, that within Amplitude as a company, you have a very strong philosophy underpinning the way you develop your business and the fact that you're not trying really to be everything to everybody, if I can put it like that, can you tell me a little bit about this philosophy and what it means and what it has meant in the development of your business? 

Karsten

Yeah, I guess...first of all, like at the very beginning, I don't think that anybody has a super clear positioning, at that wasn't the case with us because we were probably way too immature to this business, and also the business was just about to form and to develop and client base has changed a lot over the course of the years. So, in particular, in post 2008 when clients had become significantly more institutional, and high net worth family offices kind of have retracted from the space, it became more and more important to become somewhat unique, or somewhat special just to part from being a hedge fund that tries to make money, so we have really made ourselves a name for downside protection, which comes a little bit natural when you are a short term trader, because you will be much quicker in terms of turning your positions around, but also because...the way, how you put your reversionary elements in there and so on and so forth, that can have or will have a pretty significant impact on your performance characteristics. So this is certainly one of the key elements that we have in mind, and beyond that, of course we try as anybody else like us does, to make the most efficient system and to produce the greatest sharp, and just deal with the challenging market environments that you face along the time. What we've really tried to do in our research process, is to be very robust and very quantitative in terms of our approach, so we have a pretty thorough testing and, I guess everybody would say that, but because the vast majority of our research team has really strong academic and scientific background, we really take it a little bit from that angle, so it has a bit of a German over-engineered touch to it, that we try to really understand everything to its smallest possible detail before we are comfortable in terms of putting it live.  

And also, when you talk about statistics, you are dealing with so much uncertainty here, and you will never have a solution or a component which is like a super clear case where you say, "yeah, that's safe", or safe in terms of it's definitely going to do only this, and there is no circumstance under which it's going to cause a problem. Most things, under particular circumstances will actually cause problems, and we really try to take our time, even more so now, to really understand what those circumstances could be, and it's really interesting. We get pitched so many strategies, and we are... having done all these years of research, when you start questions, how you can sometimes see in particular that young people that have maybe worked only on their own and said, "yeah, I had this trading idea" and they are building that and they don't understand the danger of all the biases you can create, being that selection bias, survival bias, or technology bias even, by fooling yourself, because things look great, but things look great because you have somehow in-sampled things that you may not be conscious of. So that's really important for us, so we're trying to put a great deal of effort in it, into our research process to mitigate these potential dangers as much as we can. 

Niels

Sure, sure, and having this focus, does that (I wouldn't say allow you is the right word) does that mean that you also focus on maybe fewer, larger institution that allows you to get, maybe closer to them, and explain maybe more of all of this complexity that you have inside your program compared to maybe a Klassikal medium term trend following program? Is that an important part of your overall philosophy as well? 

Karsten

Yeah, it certainly is. The product that we have, which I would call somewhat esoteric, it's not super easy to understand. It does require a fair bit of due diligence and a fair bit of understanding from the investor side, and in particular, now-a-days, where performance in this space has been very mediocre, so you want to make sure that you have a very educated client base, and I sincerely hope that the good times will come again, and believe me we will be super selective in terms of the people that we take on board, because we have to give a lot of credits to our clients and I think, as important as it is for an investor to choose the right fund, as important it is for a good fund to choose the right investors, because it becomes like a partnership. We've worked together with the vast majority of our clients for such a long time. They know the programs to very great detail. You want to be very consistent in terms of what you deliver and that builds a working relationship. That truly builds a working relationship.  

I know that a lot of these words, and I hate all these buzz words and all this bull shit stuff when people say, "oh yeah, we're building a relationship." I can understand now what it actually means with some of these bigger clients that you've worked together with for years, and I mean five years, or seven years, or eight years, and have had, I don't know how many hours of strategy calls, that would probably be weeks of strategy calls, where they will really understand what you are doing and get comfortable with what you are doing. I think it's very important. I think it helps you. It may be significantly more difficult to get a client in the first place, but it will pay out in the long run, because no matter what strategy you run, you will eventually face some difficult times, and having clients that understand what they get themselves into, and are not just chasing your most recent performance, is super important from our perspective and, as I said, I can only reiterate at this point we are extremely happy with our client base and we have to give a lot of credits to these institutions and the people who work there in terms of how they make the decisions, how they manage their portfolios, and how they work together with us. 

Niels

I want to shift gears a little bit and talk about just the organization as a whole. I don't want to necessarily go into all the details, but I'd like to know how you have organized the business and, maybe in particular, because obviously a short term manager may have other requirements than a medium or long term manager, but also technology wise, and are there some things that now a days can be outsourced successfully, or is it important to keep everything in-house in your opinion? 

Karsten

Well, we try to keep as much in house as we possibly can because when you are so reliant on technology being that on the research, but even more so that being on the trading you don't necessarily want to rely on too many outside parties. We have...there's only one component really that we have outsourced, which is our CRM system, and I think we've lived through three different providers, and it's just a continuous nightmare. It's just really, really, really, horrible. So, having that experience, I mean its CRM so you kind of take a manual override to that, but imagine if this would be on the trading or on the research side, it would be a disaster. So we are trying to keep as much in house as we can, so you have a pretty strong IT and technology function within the firm, which will support the trading, the execution, and the research, and we have built that up over the course of the years. Then we have our execution research, and we have our strategy research. So that's how it's grouped, and I would say on the other functions in the firm we are trying to be super slim and super efficient. As you may know, we only have one person, really, doing the investor relations, which given that we have a fairly small number of investors, that it is perfectly fine, and he is supported by our two assistants who help him a great deal in terms of organizing everything and keeping all of the documentation in place. Then we have the operational side, and we have our CFO, John Harrison, and that round up the business. 

Niels

Do you have, for the two programs that you are running to date, is there such a number as an optimal size? I know that Dynamic has been soft close or hard close, however you want to put it, for a while, but is there, given the strategies that you have currently, a number that you have in mind where you say that once we get to this number we'll consider whether we need to grow any further with these strategies? 

Karsten

Yeah, so with Dynamic we are currently...it's only available through Symphony, which is like a feeder, a blended version, where you have to invest into both programs. And there is really a small and limited capacity left. We have conversations with our existing client base, we have received a fair amount of additional allocations this year from our existing clients, and there are some outstanding mandates where the decisions are yet to be made. That could drive...I personally would say very likely, a closure of Symphony which means that there is no more access to Dynamic, neither, it hasn't been as a stand-alone for years and years, but also not through this blended version, and then we will grow Klassik from there onwards, and the ultimate capacity in Klassik, we would currently see at around $3 billion, so I think around $1.2, $1.3 (billion)  in Dynamic, and then $3 maybe $3 1/2 (billion) in Klassik. It would bring the firm just slightly south of $5 billion, which I think is a really good number. That's definitely a number we want to get to. We'll see how long it takes. It probably takes longer than we thought, initially, but I'm very confident that eventually we are going to get there. So we will just really work on our two programs and we will continue to improve them until we get there and when we get there then we think what else to do. 

Niels

Absolutely. Well, let's dive in a little bit. Before we go into the programs themselves, I want to take just a bird's eye view on the track record, not so much specifically, but more also with your experience. Now one observation, of course, would be that for many CTAs the period post 2009 has been somewhat different in terms of the environment, and I wanted to ask you, firstly, whether that's something you have noticed as well in your own trading programs and also when looking at your track records, and I know of course that Klassik started in 2009, but certainly with Dynamic, is there anything one should look at in terms of looking at the track record in stages where perhaps upgrades have taken place, and where it's maybe changed a little bit it's profile from the initial part because, I guess sometimes when you look at track records and they can be 10, 20, 30 years long, you're not really expecting to have the same structure of the program that they had 30 years ago, having that today, so how does it fit in with your programs Karsten? 

Karsten

There's constant evolution in our programs and we, because we're so focused and centered around the research and the innovation pressure that you have on short term trading, which by the way, I think is significantly higher than what you have with medium to long term. I think if you stopped doing stuff in medium to long term you will have a degradation, but it's not going to be as hard as what you are going to be seeing in short term trading, so we have a fair amount of implementations. I would say the average half-life of a component in our systems is around 2 years. It's a constant progression. What we have to make sure, and this comes back to the point of how we're working together with our clients, that we are very, very consistent in terms of what we do, so there is a negative notion to the word change. That's why we prefer the word upgrade, because without being too nitty gritty about it, upgrade means you do more or less the same with what you were doing before, but just in a better way. If you change something, then it means you are doing something different, which we can't, so what that means in very simple laymen terms, is that the good months will stay the good months, and the bad months will stay the bad months. So the difficult periods will continue to be the difficult periods, and the good periods will continue to be the good periods. Now that is important because there are things you can do that will turn it completely around, and in particular, when you act as a particular building block in a portfolio, that can become very dangerous, so assuming that, for example, clients really need us to protect them in an equity market downturn, we have to make sure that with whatever we do in the research to improve the overall return profile, we would keep that functionality. That can sometimes create (I'm not saying conflict of interest) but you have something that would improve your performance, but it's going to have a substantially negative impact on your downside protection ability so that then becomes a no go. So you need to always find a healthy balance there.  

Just one more remark with regard to your initial question as to how things have changed post 2009, and to what extent that's a reflection of our program, I would say that we have always done research and we do implementations along the lines, so it's not just triggered by 2008, 2009, but I think you can certainly make the statement that the world post 2008 is a very different one. For a range of reasons, and it's not being helpful to the CTA space no matter if you are in the short term or long term space. For a starter, having markets that face interventions is not healthy. So there are two factors, there was the fiscal policy from the US Central Bank with all the quantitative easing, which takes place in a counter trend manner, is not helpful. So you've got bad news, then you have activity...they tried to stabilize those markets, so whenever you were going short because it looks like there's a sell off, and the markets going a little bit south, that's going to be stopped through either statements or activity from the central banks, or statements from the politicians. On top of that, which made it actually even worse, on a completely different level, was the way how the sovereign...the European debt crisis was managed, and I would say that out of all possible people, I would say Merkel is probably the person that had cost the CTA industry the vast majority of money. Because the way she has handled this crisis, which was salami tactics, with no clear positioning from the very start, but just in a fire fighting mode. It's been horrible for the markets, from our perspective because it just would clear...it would just kill any clear trend, and it takes the natural dynamics away. The markets have become, at least...like, in 2010, 2011, 2012 - ever since it started, have become so news sensitive to those kind of things so you had a negative news flow because some Troika was visiting this peripheral states and found holes in their budget and the markets were taking a negative view on it, and a couple of days later then the ECB or France and Germany were just committing to another rescue package and then the markets traded up again. So that was just really not helpful. Beyond that, commodity markets have been challenging, which they have always been, in history, so with commodity markets always had periods where they were super attractive and periods where they were challenging. But everything comes together at the same time that obviously creates an environment which has just been tricky.  

I would say, in terms of model building, you have three break points. You have 2004, 2005, which is the advent of short term trading. So you have very simple models to work up until then, and then decay afterwards - as relatively easy to work your way. Then you have post 2008, with all the things we have described and then you have post mid 2010, or let's say from the start of 2011 were it got really bad. I would say up until the end of 2010, we have really, really good systems that can handle everything, and then post 2010 it becomes a bit specific. It becomes really hard to work the last three years, for the reasons I mentioned, and increased correlation between the markets, which hopefully now goes the other way, because with Europe going a different way than the US could be very helpful, and then we have, obviously, this government intervention factors, which I think in the US is coming down as well, which is great news, and then Europe going the other way, so we'll see how that plays out. That's where we've been...those are the challenges that we have to deal with. 

Niels

And just before we leave that and we go into the trading programs, do you thinkwell two things, actually...one, have you noticed...I don't know how far you've gone back in your research, but have you noticed something similar back in time, meaning that OK maybe it's been 2, 3, 4 years now, and it's been difficult but it's not unusual...I recently heard a study that Larry Hite and his team at ISAM had done going back up to 800 years, and they were basically saying that this is really not unprecedented, and it's happened before, and it's going to happen again. I was just curious, whether you had noticed anything from the short term point of view in the same direction? And also, whether or not the last few years actually speeds up model decay, or whether model decay is more or less a constant, regardless of environment? 

Karsten

Yeah, first of all, I have to say, I don't know what use there is to go back 800 years. It sounds to me like rather weird, to be totally honest, and certainly from our...I don't know what Larry Hite is doing in his programs, and is none of our business and is none of our competitors neither, so from our perspective it is unprecedented for the reasons I've given, and you also have the factor that obviously technology has advanced significantly. So when you run, like whatever we do now, when you run that on data from say pre 2005, it looks of course great because that's something that we call technology bias. So today, with all the experience you have and all the computer power that you have, you will be able to build strategies and models that you would have never been able to build back then. It's like fighting yesterdays wars with today's weapons - I think that is a very similar analogy. To that extent, I think the challenges are tough. I don't think that the environment will stay as bad, because environment changes along the line, and there are very specific things that drive markets, for the time being, which create a challenging environment for what we are doing, but that does not necessarily mean that those things will be there forever, and besides that, we all know that markets are stable for some time, and then there's either a bubble or a problem that nobody has seen before, or something else destabilizes the world, and then it all goes upside down again. And from an investor’s perspective, I think there are very, very few people in this world who manage to call the markets in the right way. And even those guys can be surprised by the development of a due political risk situation that nobody can forecast.  

So, from an institutional investor’s standpoint, what you have to do is actually build a stable portfolio, like a long term strategic stable portfolio, and therefore you need protection for turbulences in the market. As long as that protection is not going to cost you money in the meantime, I think you have to go for it, unless you say, look, I'm totally happy about these five, six, seven years where I produced great returns. That one year, where I'm going to lose it all, let's just forget about it, it was just a crappy year, let's put it behind us. So, if you have that kind of attitude, yeah, then ignore trading strategies, just go for strategies that collect some sort of premium being at a carry or whatever, then you can do that. But if you are actually concerned about these years as well and want to have a decent performance in those years as well I think you have to have a balanced portfolio. 

Niels

Sure. Well, talking about the strategies, let's dive in a little bit to the trading programs, and obviously you'll decide whether you are going to talk about it from the Klassik point of view, or from the Dynamic point of view, or a mix of both, but tell me a little bit about, in your own words, how you've structured the programs, and why you've designed them in the way that you have - if there's a particular sort of...we talked about that there is a design philosophy that you want to provide protection but, talk us through the overall structure of how you have gone about it. 

Karsten

So, as you will probably appreciate, I can only do that in a very superficial way, and the way the systems are built is that we have lots of components that come in to an indicator, and we bring those things together to achieve a couple of things. It's beneficial for us to run many, many strategies because it will allow us to have a more equalized activity in the markets, so that we are not just trading in one goal, from a slippage perspective. That's important. We are obviously also trying to achieve diversification through many, many different models that we run. We use reversionary elements for efficiency reasons, and to be distinct, so to have a unique return stream. In terms of structuring out the portfolio, we need to keep considerable liquidity as a key factor for weighting the markets, so being a short term trader, you will naturally have more concentration on the deep liquid markets, because otherwise the slippage is just going to go through the roof, and in particular, as you accumulate assets, it becomes a factor. So that's how both programs are structured.  

In terms of the research implementation that we do, we have a generic approach, so whatever we look at we try to look at it from all different trading timeframes, and it is unfortunately the case that with many things you do, they only work in particular time frames. Some of the critics may say, "oh yeah, that's like over-fitting", but naturally there are things that expand your time frame otherwise you are working on different price actions in the market, you're working under statistics and certain things do not work anymore. So therefore, in terms of the final implementations that take place in the two programs - into Klassik and into Dynamic, it's quite different, what we do in Dynamic and what we do in Klassik. Initially, and I have to admit that we tried to grandfather a lot of the things that we had in Dynamic into Klassik and that didn't really work very well, so the first 1 1/2 years, performance wise, really was a disappointment, so we did a lot of dedicated development with Klassik, and ever since it became more like what we wanted it to be in terms of performance characteristics and quality.  

Niels

Sure. How many markets do you actually trade in each of the programs? 

Karsten

We have roughly 40 markets in Dynamic and around 70 markets in Klassik. 

Niels

OK. And what does the investment process look like, if we take an example from signal generation in a particular market, what happens then, what kind of factors are you looking at? Do you need intra-day data, which I'm sure you do, and so on and so forth, just to visualize a little bit about what it actually means when we talk about these short term processes? 

Karsten

Yeah, so we mostly work on tick data and what happens in real time, fashion our models, recalculate their positions and their signals as they change, as it creates a delta between the position that we want and that we have, that feeds into our execution engine, which will then work the order-book, so depending on what the order-book looks like, this will lead to a certain type of trading activity to get the most slippage efficient execution. There are risk overlays which are largely directed towards the volatility we are experiencing in a market, or particular intra-day true range breakouts. So if there's massive events taking place, there will be a real time downscaling of our positioning as well for risk management perspectives, so that drives the ultimate positioning in the market. 

Niels

Sure. Is the model a market specific model, or, I guess it could be both - it could be a market specific model, or it could be a model that you run on all of the markets, obviously traditional trend following tends to have fewer models and many markets, but I imagine maybe you have a lot more models, but on maybe a fewer set of markets, so how does that work? Do you have both type of models or are they all really designed for a specific market that you want to trade? 

Karsten

Well there are different philosophies on how you can do that, and that's all about how statistically significant you can be on doing things individually or globally across asset classes and so forth. I would prefer not to comment on how we do it, but it's definitely viable to do both approaches, as long as you can be robust. I would always try to be as specific as you possibly can. As long as somebody feels that I'm not fitting to something particular here, which is not robust, then of course you have to expand your data set. But if you are comfortable with that, then try to be as specific as you can. 

Niels

Sure. Do you disclose roughly ballpark in terms of the number of different models you run in each of the programs, just to get a feel for that? 

Karsten

We do that to investors, but generally we don't disclose it. But of course, if you are an existing client, and in progressing through strategy due diligence, we will talk a lot about these things and we also talk a lot about implementations that we have done the research side, so that people can see progressions and see what we do. But that all comes at a deeper stage of the due diligence process. 

Niels

Sure, sure, and would you be able...because obviously the audience has varied experiences in some sense, is it possible for you to visualize, generally, the difference between a trend following model and a mean reversion model if you were just going to try to explain it as to, not specifically how you approach it, but just generally how you approach the two different types of models? 

Karsten

Well, so trend following takes a position in the direction the price has gone to trigger the signal, so the price goes up will trigger a signal that you go long. With mean reversion, or counter trend models, as like the name alludes to, the position you would take actually goes into the counter direction of the price movement in the market that will actually trigger your signal. So if the market continues to go up you will eventually go short. That's the core difference. 

Niels

Sure, and of course when we...obviously there are more of the medium to long term trend following managers out there, and so it's quite interesting to hear your perspective and that is the debate, I guess, very often comes down to are people using moving averages, or are they using price breakout as their indicators. What are the pros? What are the cons? But actually, my question is more do these traditional trend following type indicators, do they also have validity in the short term space? 

Karsten

I don't think that anything that you can read in a book... 

Ending

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.