Secrets From A Profitable Systematic Trader - Robert Carver | Trader Interview

0:03 welcome back everyone to with Rob Carver and we'll talk about systematic trading is a systematic trading expert and someone who is really knowledable on that topic also the author of several books on that topic as well so Rob on the podcast how's it going today uh fine good nice to meet you good to have you here on the podcast of course looking forward to discuss that topic and kind of your journey as well in trading so

0:24 actually tell people first about who you are and kind of what you do these days uh so nowadays I I describe myself as a an independent systematic Futures Trader um which people think means that I spend all my time sitting in front of a screen clicking buttons but the the keyword there is systematic uh so my system is actually fully automated so um I spend most of my days actually writing code for that system or doing research

0:47 on various aspects of of um sort of financial stuff um and I also write a few books as you've kindly mentioned um and uh I also teach a master's course in systematic Trading at the University of London so um I'm pretty busy but I don't I don't have a proper job so to speak interesting definitely not that we we can talk about and focus on so let's go back in time tell me more about kind of

1:11 what made you get into trading in the first place I guess let's go a long a way a long way back to when I was seven years old um which is when actually I got my first computer um so I got into computer programming then uh and actually until I was about 18 I was convinced that that that's what I wanted to do in my life so I got pretty good at programing computers which obviously comes in handy

1:32 nowadays um and but I did I did sort of start a degree in computer science that didn't work out so I switched to economics a few years after a few years Gap um did a couple degrees in economics and got really interested in the financial markets um so just over um 20 years ago now I I sort of started working as a Trader for an investment Bank uh which I was trading exotic exotic interest rate derivatives so kind of still quite mathy

2:01 quany um but not systematic um important difference there um and actually I didn't enjoy the sort of non-systematic world of trading I didn't enjoy sitting in front of a big screen and having you know having to make decisions myself I didn't enjoy that at all uh so I took a couple years off if you like and worked in economic research a couple years and then I went back into um into the sort of world of Finance working for a hedge

2:24 fund called AHL um which is one of the largest systematic feutures Traders in the world um I did a couple of things them the first thing I did was kind of bringing in my economics knowledge I built a a systematic Global macro cross asset strategy for them uh did that for a few years and then there was a rorg and I got well basically promoted to run the the whole of the the fixed income portfolio for them so that's anything

2:49 betting on interest rates um so bonds interest rate Futures uh cash bonds credit default swaps you know interest rate swaps you name it um so that was a lot of fun and that that was sort of for context you know I was working for them between 2006 and 2013 so some pretty spicy years there in the financial markets for those of you that remember U then in 2013 I decided to um kind of take a break again and um

3:15 so I left I left there and and since then yeah I've been doing what I described to you I've been been trading my own money with with a system I wrote myself uh and doing lots of other side projects and uh uh the way I like to describe it is I kind of do what I used to do for work um but I don't do any of the boring stuff so I don't have to to do kind of

3:33 tedious meetings and meet clients and and all that stuff um I could just focus on the research which is what I find interesting sadly it doesn't pay as well as it used to do as well because um but but you know you can't have everything I have a much better work life balance now let's put it that way so one you think of any have a master in in economic which is definitely something interesting there when we think of

3:57 economics and training I'm guessing you'll be most looking at like fundamentals the bigger picture The mcro View and not too much on like chart patterns and like patterns andly on on graphs so how does that play off with systemizing your training kind having everything corded and systematized yeah so that first project I did when I worked for AHL was very much about the idea of saying well okay we think there are certain there yeah as you said there

4:20 are roughly two ways to look at data that might predict markets systematically right there's what you what price data um and that's kind of called technical analysis although a lot what a lot of people call technical analysis I don't do so I don't I don't believe that looking at charts and seeing weird patterns or Fibonacci levels or any of that stuff I don't believe in any of that stuff and I don't use it um but when I say technical

4:43 analysis I'm referring to any systematic strategy that uses price as an input and nothing else maybe volume but but definitely price um or alternatively you can look at fundamentals so um simplest way of think about that is inequities uh because you can use things like you know P ratios dividend yield return on Capital employed cash flow that kind of stuff and you can combine these in various interesting ways to form Equity factors and that's how a lot

5:07 of um you know hedge funds work that trade equities um and what I was employed to do initially was to say well let's take this idea of fundamentals but let's bring it to different asset classes so we do the stuff in equities but we also do we also look at FX and say well what what kind of fundamental variables predict FX things like inflation interest rates you know what kind of things um would predict potentially um bonds again it would be

5:33 things like Central Bank policy what do central banks look at well they look at things like GDP growth unemployment rates inflation rates and so on and so forth so yeah it was about taking a series of of numbers and form and using them to form a a forecast of prices so in principle exactly the same as the idea of taking prices as your input but obviously just with a different data set so a similar approach but a different

5:56 data set is the best way of describing it um for what it's worth I no longer do that kind of trading I focus purely on price and that's not because I don't believe that that fundamental stuff doesn't work it's purely kind of a question of time efficiency um because by just using a single kind of data input I can run my system fully automated with almost no work involved and uh I'm a fundamentally lazy person and that's why I trade systematically so

6:20 you know um the more the more data and more kinds of data you bring into your system the more work you need to do in kind of cleaning and monitoring them and I have no interest in doing them and that that's the only reason I don't currently trade fundamentals it's not that I don't believe in interesting for people who are very kind of like hardcore in in technical analysis and like Shar patterns and like you mentioned Fibonacci and maybe channels

6:41 trend lines they don't imagine there's like a different way to look at price and the way you look at it which is probably a lot more simple than this uh so tell us more about kind of what are is to look at Price within that systematic trading aspect without these like patterns and and tools that have out there so a very a very simple um probably the simplest trading role I use is is what would look to a lot of people

7:02 like a breakout system so what I do is say okay let's look at the series of price over the last n days um and um we you know the value of n will depend on how sort of length of Trades of trends that you want to predict so if you want to predict slow Trends you'd have a big end if you want to predict trends that have only been around for a few days you'd have a small

7:20 end um and then I just take the range of prices in that um period so I have a high price and a low price maximum and minimum and also Al the the mean the mean r in that range of prices and then I look at where the current price sits in that in that range uh and if the current price is exactly on the average I would have no position so I'm I'm saying the price is currently you know

7:43 not doing anything interesting um if on the other hand the price was right at the top of the maximum range in other words the current price is the maximum so we've reached the new high I would have a maximum long position on um and similarly if um the price was right at the bottom of the of that range I'd have a maximum short position on um and my position would vary between those two maximums and Minima depending on exactly

8:08 where I was in that range so that to me that's a pretty simple um way of describing it and the nice thing about that sort of system is it's purely objective so it doesn't involve you staring at a chart and trying to find a pattern um it's not using any kind of insample fitted or kind of made up numbers like Fibonacci levels or levels that you found kind of in the data set that seem to prices um it's just purely using a

8:32 maximum minimum um and yeah it's very simple um and it's intuitive and it's very easy to understand what it's doing um and there's only one parameter you need to fit which is the number of days and as I said that that really depends on the length of Trends you're looking for so you wouldn't necessarily go about kind of fitting that parameter per se you just choose a set of different ends that suit the cost of the market you're

8:54 trading in particular um so yeah I have about 50 of these well probably about 30 35 of these things that try and predict Trends in prices and then I have some things that do other things as well but they're all I could describe all of them to you in a minute or so like I've just described that breakout system they're all fairly simple um but the system as a whole because it's made up of all these

9:13 little simple things is rather looks relatively complicated but ultimately it's based on very simple um rules that that I believe are robust out of sample and know a lot of indicators are fairly easy to use when you code are you a fan of using indicators kind of help with your strategies or is it only B and price right so um okay so what you mean by an indicator I think it's a function that takes price and produces an exactly

9:36 yeah yeah come back to the same thing but yeah yeah so so I I um so if you the system I've just described to you is in fact an indicator effectively because it's taking a series of prices and producing forecast so in fact all of you could describe everything I do as an indicator um having said so I and so for example I do use moving averages um in fact um exponentially weighted moving average crossovers are one of the

10:00 systems I actually trade uh I use them quite a lot um I don't use Ballinger bands I don't use RSI um but but I you know everything I do you could describe as an indicator um but my my my kind of functions of price if you like are usually relatively simple and without you know lots of parameterization um and um without any overfitting which isn't necessarily true of well certainly Fibonacci sequences but um you know um

10:27 ultimately with a lot of these things I'm just looking for a way to find a trend um and to be honest it doesn't really matter to an extent which way you've look to find your Trend none of these um functions is necessarily better than any other but by using a lot of them I can gain a diversification benefit um and what that means in practices there might be a point when particular breakout breakout signals working really well and the moving

10:50 averages aren't working really well there might be another time when those things are reversed by trading both of them I can I can still make a little bit of money in both those scenarios because if I just pick one indicator my performance would be much less even I'm also very kind of systematic with my own trading but I know people who would look at this would say oh systematic is so boring I prefer to kind of be in the

11:09 moment be in the action kind of pick the tra myself why we de someone like this do they have to kind of be more systematic themselves or can they just kind of stay with their own ways like a personality thing that they can just do that differently or do you think systematic is better for everyone well okay let's just first of all the first question to ask is well why are you trading um if you're trading

11:28 for fun um then absolutely don't do not trade systematically because yeah it's tediously boring particularly if you're doing it in a non-automated way um so if I had to run my system manually I I would die of boredom very quickly um because I'm trading you know I'm picking up price signals from a couple of hundred markets and trading a hundred of them um so you know you can imagine how much work that would be manually every

11:52 single day for no way um if you're trading if you're trading for fun and you're trading a small number of instruments and and you're you know you're enjoying what you're doing then that's brilliant but if you're trading to make money then you need to consider whether your discretionary trading would be better than a trading system in terms of its outright performance and I would say for the vast majority of people not everybody there are I do believe there

12:17 are people who can trade in a discretionary fashion and be profitable but it's a much smaller percentage of people than than you think um and the average person off the street is unlikely to be someone who can trade profitably on a discretion basis I know I can't for example um and um and so if if you can't if you're not going to be more profitable in a system then yeah you should be trading with a system if

12:39 your objective is to make money if your objective is to have fun then yeah don't trade systematically just put a small amount of money in your account and have fun that's fine um I'm not going to tell other people what they should do for entertainment um personally I don't find trading fun and I don't do it for for my amusement I do it to make money um what I'm what I find fun actually is the kind

12:58 of Arch and the understanding of the markets and the theory and the stuff behind it but the actual practice of sitting and looking at a chart and clicking on things I I find tediously incredibly boring to be honest I think what's scary for a lot of people is that if they have a system and they just trade a system very systematic in the market then that system could lose money at some time they won't they wouldn't be able to influence it or

13:19 change the results that system if they are there they feel like if they are in the market to them Place trades they can sort of flip around change things up to kind of make money all the time basically it's a belief so again you can probably solve this with like a good portfolio of of strategies which you definitely have how much uh room do you leave for draw downs and losses on one particular strategy is it like very

13:41 strict very tight or can the strategy be losing for for a while and come back to a profit how do you kind of look at that I mean if I look if I look at my live training performance so that which is you know it's kind of true out sample so there's no back test you know no back test kind of influence there um over the last eight years and I've been trading my own money and kind of following it

14:02 very closely I think I've had like two or maybe three losing years um not not big like 5% losses 10% maybe um so to answer your question is can you lose money for a period of time the answer is yes you know a year 18 months um and unless your trading strategy is extremely profitable that's to be expected um in fact one of the nice things about trading systematically is you can run a back test and look at

14:26 your likely performance and your likely maximum draw down and that will then give you an indication of whether when you're running it for real and you're losing money whether that's kind of within your sort of window of expectation um you know so if in your back test your maximum draw down is 40% and your current draw down is 10% you shouldn't be worried or de or deviated from the system or changing or trying to improve it in any way you should just

14:49 let the thing run if on the other hand your Maxum draw Downs 50% then yeah that that's a concern um and this there may be you can know a few things could have happened you may have over 50 your back test it's too optimistic the markets may have changed in some way and the effect that you're hoping was there isn't there anymore um maybe trading costs have gone up maybe there's a bug in your code you

15:08 know there could be all kinds of explanations but it's definitely worth further investigation um but but um for me personally you know knowing the kind of my expected loss over a given time period from a back test gives me confidence the fact that I know and understand the system in detail and because it's a simple system and not like a blackbox fitted system with parameters and you know it's not like I've taken some machine learning algorithm and discovered a set of

15:32 parameters I don't really understand I haven't done that um so I understand the system find it intuitive I know how it's working and I have an understanding of what I expected to make or lose in a given time period putting all that together means uh I can sort of trust the system and the other important component is the amount of risk you're taking um so you know for me I kind of I think the worst draw down I've had in

15:53 live trading has been maybe 15% if I was taking on a lot more risk um then you know if I say taking on double or triple the risk I have then that worst draw down be two or three times worse you know to be down 30 or 45% of my Capital would you know would personally I'd find psychologically very difficult and I could understand why anyone would to be honest um so I've deliberately set my risk tolerance my

16:16 risk leverage a level where I'm comfortable with the you my expected risk tolerance um otherwise yeah if I was down 50% I can understand why I'd want to make changes to my system um but that's something likely to happen um because of you know where I set my leverage Target um yeah so ultimately this if you're trading systematically but messing around with your system that's pretty much the worst of Both Worlds because you don't get the advantage of

16:42 systematic trading but you're also not getting any any advantage you might have as a discretionary Trader um there there is there is room for combining these two things together and I know it's not the question you asked but I'm going to going to give you this anyway because I think it's an interesting thing um so actually I think it's probably okay to make your calls on the market in a discretionary fashion the words to say well I think you know GBP USD is going

17:03 to go up next week and and and I I kind of feel that on a scale of n to five as about three I think that Corn's going to go down next week and I think that's like a two out of five uh and then what you can do and I actually illustrate this in in my first and my third book is to take those kind of Quantified gut feelings or whatever you want to call them discretionary forecasts and then

17:25 plug them into a a systematic position and risk management system system um and that will then and then to actually make you know the trades that come out of that um and that that's quite a good you know that if you do believe you can make Mar calls on the market which I believe there's I believe there is a bit of an art to trading in that sense but I believe that and I believe that the part of trading that

17:46 should definitely be left to science is position sizing and risk management and and management of trading costs um so that should definitely be systematic um so yeah you know if you just have a system but then just play around with it then you're you got the worst of Both Worlds you're not really you're not really a systematic Trader you're not really discretionary Trader if things go wrong you don't know whether it's a systemate fault or whether it's your at fault so you know

18:09 what you what are you going to do so so that would mean if you have a lot of confidence in the trade like it's a 515 you would have more risk on it and if it's a one5 you have less less risk on it is that right uh yeah that's correct and that's actually built into my system so if you take the the break the breakout system I was talking about earlier so imagine that the the price is

18:27 just above that that mean line in the range I have a very small bullish forecast and therefore a very small long position on but if it's all the way at the top of the range I have a much bigger forecast and therefore for a much bigger position on so I'm not I'm not doing what I would call discrete trading which means the moment that thing crosses the center line I buy like a fullsize position and vice versa when it

18:49 goes below it I'm gradually changing my position over time which reduces trading costs and also kind of reduces the the sort of outliers in my returns as well interesting at the same time I'm guessing it's very tempting for a systematic Trader who has something coded to just like go and look at the the settings and the things you could tweak to make it just a little better I'm sure your students kind of go through that too where they want to make

19:13 it the best possible so they tweak everything a little more how does that kind of play out with your training is it like very simple no tweaking or is it more like tweaking a few parameters bit more to get something better out of it I mean I I haven't changed the basic parameters of my system for eight years so you know that that answers your question um you know it's a mistake so I know I've got a 40 a 40 Old 50-year back

19:37 test now for my system because my data goes back to about 1970 um what extra information have I got in the next in the last year for example the last month that would allow me to say that the system should be fundamentally very different the answer is almost none um so you know this chasing of recent performance and trying to you know so a classic example would be if you go back a three years now to

20:00 the covid crash in March 2020 you know fast Trend following systems did much better than slow Trend following systems and the reason was the market just basically did that right pretty much um so a very far system would have ridden that on the way down and then ridden that on the way up we've been very happy a slower system would have basically got it almost exactly wrong it would have got out of the position of the bottom

20:23 gone short and then being dragged all the way back up on the long side so it would have been hurt so a reaction to that might be to say well gosh this isn't very good I'm going to change the parameters of my system so I allocate more to fast Trends and less to slow Trends but that's one episode that's literally like one or two months out of 50 years um you know it shouldn't influence the way that that that the

20:44 parameters are same in the trading system most people um you know overfit their systems far too much and they also you know re recalibrate rep priorize the systems far too much the only people it makes sense for to kind of fit their systems on a really regular basis are people who are trading very quickly high frequency Traders because they're getting an awful lot of data which means that they can actually make kind of statistically robust decisions about

21:09 what they should be trading firstly and secondly they are in a world in which the the market is changing like literally every few months so a high frequency system that worked you know six months ago is probably not going to work anymore I'm in a completely different world I'm trading Rel slowly the stuff I'm trading worked 50 years ago and will hopefully work for another 50 years so you know reacting to a something that happened over a couple of

21:31 months is a complete waste of time definitely good point there and in terms of Market do kind of trade all markets you can put your hands on or is it just a few markets where you know your stat would be good and you know it's going to work on so 100% I do not cherry pick markets on the basis of performance that is again is a form of overfitting that that a mistake that people make um so I

21:52 I do I in one of my books I've got a nice plot I plot the performance of all the different markets I trade on a on a chart and it looks like there are some markets that are brilliant and some markets that are terrible and then I add error bars onto that plot which shows the amount of statistical uncertainty in each of those estimates and once you do that you can see there's no difference between these

22:11 markets all of the the difference in performance is entirely down to noise it's not significant at all so what I do is I want to trade as many markets as I can because the the benefits of diversification particularly for Trend falling systems are extremely high um so for example if I go from One Market to 100 markets then my risk adjusted performance um in my back test goes from like up by a factor of nearly five you

22:38 know there aren't many ways many places in finance you can get a fivefold Improvement in performance that's for sure um so um I want as many markets as possible um so I actually have in my data set that's that's all the markets I'm I could potentially trade about 200 instruments and that's pretty much all the Futures in across World apart from a small number I can't trade because the either it's illegal for me to trade them

23:03 or the data costs are prohibitive there might be another 30 40 or 50 Futures I'm not trading and I also don't trade um single stock futures as well um that's just a preference um and then what I do is I say well out of those that 200 or so markets how many are actually liquid and cheap enough for me to trade um and I also exclude markets with very low risk as well because that would lead to

23:26 excessive leverage after doing all that I end up with maybe a 100 markets that I can actually trade and put positions on in um and then on any given day I'm probably holding positions in somewhere between 10 30 of those markets so um you know I'm I'm really looking for a lot of diversification um because I don't believe that any one market is better than another as you implied in your question I really don't believe that at

23:50 all um I want to trade everything I can subject to you know the fact that it needs to be liquid needs to be cheap and The Leverage is too high and then how much work goes into the the review process of like going back and looking at your stats and your results is it just like like results quickly and then you're good with it you don't do anything with those results or it's a matter of kind of finding out what works

24:11 what doesn't work kind of tweaking it a little bit or so on I don't tweak I don't tweak I don't tweak don't tweak don't tweak don't tweak um I mean I review my trading performance every year I have to do it anyway to like fill in my tax return right um and I usually write a blog posting a bit of a n assist and saying you know well this there's some interesting things that are happening but absolutely you know the

24:33 the only the only changes I make are related to things like for example liquidity and trading cost changing um so you know it might be for example that a given instrument suddenly becomes less liquid and therefore I exclude it from my portfolio it was there before alternatively an instrument becomes suddenly much cheaper to trade so maybe now it's in my portfolio wasn't before or maybe it becomes cheaper to trade and that also means that I can actually

24:58 trade it a bit quicker because the the only difference between instruments as far as I'm concerned is trading costs so things that have got lower trading costs actually allocate more to faster Trend following systems than other fast systems and therefore trade them quicker uh if they're expensive to trade then I I trade them slower and slower until at some point I just stop trading them because it's it's not worth it anymore um so that that's the only change that's

25:21 happening um with with with my system it's um you know because there's no there's no reason other re there's no new information for me telling me that for example yeah 10y year Italian bonds are now a much better Market to trade than than euro dollar because as I said if you look at the statistically significance of that result it's going to be negligible and that's over the whole back test over the last year it's going to be even less than negligible if

25:47 you can say such a thing um and similar you know there's no there's no statistical evidence that for example putting cost to one side I should trade say S&P 500 more quickly than say us 2-year Bond Futures you know there's no evidence of that in the back test and there's certainly no evidence based over a relatively short period of the time um so you know I'm because I'm looking for things that have worked for 50 years um

26:12 it makes no sense at all to be tweaking and changing things the whole time at best that will just lead to higher trading costs because every time you make a tweak the positions the strategy wants to take will change and you'll trade for no reason um and at worst you'll be overfitting and completely destroying your performance interesting that that that once year is a big takeway because lot people would want to see St quickly right now make

26:34 sure it works it's profitable but giv it giving it a year to kind of let it play out and not be touching it or changing it too much and that's a I know a lot of wouldn't be able to do it but it's still a good thing that they they should be able to do one one thing to bear in mind as I wasn't born like this right so when you know when I when I started in this

26:51 business um I think I I probably wanted to change things a lot more um because that is a natural human Instinct um and actually when you're managing other people's money um I my perception at least for me personally was the was actually more pressure to change things because when the system is losing money you know you're getting calls from clients complaining and saying well what are you going to do about this the correct answer is nothing you know

27:17 clients don't like that they want they will like no no no no no so there must be something you should you should be doing about this that the because you know most clients don't have a a real understanding of this particular issue as I've just described to you um so even if you personally like know the theory and know the maths and appreciate that when you got people calling you up and complaining you you know you're going to

27:36 feel Under Pressure to do something because I'm now in a place where it's just my own money you know I don't have anyone ringing up complaining I I can do what I believe is correct what I believe is correct for me personally and for the sort of systems I'm trading is to change things as little as possible uh and to yeah look at performance occasionally but not to you know make changes as a result of that we would talk about so

27:57 far prob going to ask someone in the chat about like we talk about all these Market these tests you've done this like this data You' collected and stuff and people might feel like well then systematic training is not good for beginners people who are just starting out do you feel like that's an issue where it's kind of tougher to be a systematic trailer than just kind of going and become a discretionary trailer or is it possible for anyone to

28:18 systematize thing and kind of do it similar to the way you're doing it so I actually believe that that everyone should start as a systematic Trader um and the way I like the analogy I like to use is like when you learn to ride a a bike a pedal you know a pedal bike you have training wheels on normally and those Wheels protect you from kind of falling over until you get your balance right and then you're

28:40 you're cycling okay and then when you're kind of confident enough you can take the wheels off um so if you start trading with a trading system if it's a well-designed system then it's going to protect you from doing anything stupid like you know overtrading putting on too much leverage and so on and so forth um the system can be really simple so I'm in in my third book leverage trading I give an example of what I call the

29:01 starter system which consists of just one trading rule um and you know the breakout rule I've described for example is a good one you could use because actually you don't need you could literally just pull up a chart on your monitor and change the time range to the last 20 or 30 or 50 or 60 or 70 days whatever it is you're looking at um and you can see you can see what know what position you need to have on you can

29:26 just get a ruler on the screen don't need don't need to spreadsheet you don't to learn how to code or anything like that um and then you you basically got a number that's telling you how long or short you can be um and then you go to another chart another you know any charting package that will tell you the amount of say the ATR some measure of risk in the in the instrument um and then there's a spreadsheet that maybe

29:46 has 10 rows you put those two numbers in it will tell you what position you should have on and what trade you should do um and that that's it you know um obviously that that's one extreme you know my own system runs to I think 100,000 lines of code um a lot of that code is because it's working on an automated basis which means that there has to be you know an automation requires at least if you're doing it

30:08 properly an awful lot of checks and balances and and you know safety built in and all this kind of stuff um but the the actual bit of code that's kind of generating the the trades if you like is actually relatively small it's maybe only a few hundred lines but clearly that's still more complicated than just a simple spreadsheet but but you know you can people I know I have friends who run systems of equal sophistication to

30:30 me just on spreadsheets um so you know it's possible um so yeah I actually believe that as a as a someone who's starting off trading you know pick a simple trading rule find a system that that's safe I'd say you know obviously I have one in my third book but there are others um and you can literally just even without any almost anything apart from just a very very basic charting package or just even free data off the

30:55 internet you can just literally trade um and learn learn what you're doing and understand what you're doing and learn what it feels like to put trades on and not change them and and and kind of and then and only then if you then start to think well you know I've got a feel for the markets then you do what I say earlier which is you you can replace the the simple breakout rule with your own kind of discretionary yeah I think the

31:19 Market's going I think it's going to go down um and actually one thing I include again in my third book is basically a simple table that does the statistical test for you to determine whether your kind of qualitative calls that you're then making are actually profitable or not because it's very easy in trading to confuse lock over skill you know so you do four trades and they all make money and you think you wor and buff it right

31:43 practice you need a lot more trades in that for statistical significance and that's not appreciated so yeah I I think for me personally I I I think everyone should learn to trade with a really simple system and then go kind of one of two or three ways you know either they have got some genuine feel for the market then they could go that way you know carry on putting their kind of gut decision making feeling feels into a um a

32:08 systematic position and risk management framework because that protects you from doing this stupid um and then test whe you they can test whether that's actually a good thing to do or not or they can do what id do which is make the system kind of more and more complicated and add more things to it um and um you know one of the the easiest ways to make extra money as I already said is just diversify across Lots of instruments you

32:29 know that will give you a huge benefit performance so even a very simple system run over enough instruments will give you you know very very acceptable performance so yeah I believe exactly the opposite I think I think um you know if I was sitting down as a beginner trading with no knowledge about about how to trade systematically or about risk manager or any of this stuff I mean you might as well just write a check to

32:48 the broker and give it to them and save time to be honest because you're going to end up probably in the you know unless you're very lucky you're just going to end up losing all your money right how do you get new ideas for strategies is it based on like books or is it just random ideas you get like this on a on a daily basis for different strategies you want to kind of test or create where do these come from um I

33:10 mean obviously comes from various places um I'd say probably the majority come from like stuff floating around on the internet so people will send me a research piece done by someone like say aqr which is a you know hedge fund that does some very good research in in this area uh so we'll have you look to this and I'll test it and maybe it works maybe it doesn't work um I think um if you're if you're spending more time

33:35 looking at the market so I don't spend much time actually looking at the markets like if you asked me where the foot the S&P was now I couldn't tell you um even though I have a position on in it I can't actually tell you what my position is is in it either to be honest um I can find out if you want but but it's not something that I have in my head um but if I think when I was

33:55 working at AHL I spent a more lot more time actually looking at the markets because you know I had a balloon bow terminal in front of me and so on and so forth um and I was act you know more actively monitoring the positions because as someone looking after other people's money with a fiary duty that's what you have to do and there actually more of my ideas might actually come from kind of Market Behavior if you like

34:16 like oh look this this looks like there might be something here going on in the markets is this something I can test and reproduce um but now you know because I'm not doing that the the vast majority of my ideas come from yeah occasionally from books but more often and not from just reading stuff on the internet yeah way to go yeah I mean I'm being honest just and saying you know I have almost no original ideas to be honest so it's

34:37 good to hear because a lot of people think that oh it's like you got to be so clever to kind of find the these ideas sometimes or it has to come from you directly and then they they don't yeah I think it's misconception that that to be a successful Trader and make money you need to have come up with a basically very original and secret ways to to forecast prices which you then know and no one else does uh and in practice

34:58 that's pretty much nonsense uh I mean none of none of the stuff I do is a secret you can actually download it off the internet or buy a book and read about it um and um you know my my system still reads me profitable um and and and that's because there are there are plenty of ways to to kind of collect it's called know economists would call it collecting risk Premier so basically you the the res sources of risk out

35:21 there in markets that people aren't prepared to take uh if you are prepared to take those um and that's why a systematic strategy is so good because it has no emotions behind it it doesn't get scared when it has a certain position it doesn't get greedy it just does what it thinks is best it's been programmed to do um so you know people said me why does momentum work as a trading strategy it's because it's very

35:41 uncomfortable as a human being to trade momentum without a system without automation um my system doesn't isn't a human being has no uncomfortableness about it just goes ahead and does it um so so yeah I don't I don't I think the role of kind of SE what I call secret source is is is massively overblown and even quite successful hedge funds you know 90% of their returns is coming from stuff that anyone could do and everyone

36:06 knows about uh and maybe they're making an extra 10% because they do have a little bit of secret Source um but but most a lot of the time that extra 10% is coming from just not doing anything stupid um so you know my my my trading returns are a combination of earning risk Premier plus not doing anything stupid um and that doesn't require any special skill if requires experience to know that that's what you should be

36:30 doing and not meddling and searching for all these other things out there in The Ether um but it doesn't require skill per se and it's not it's certainly not a secret how how I trade at the same time when you test something new what makes you decide if it's working or not is it that you test it out then you tweak it see if it could work a different way or just test it out quickly and then drop

36:48 it if it doesn't work the first time well yeah I mean so the first thing is I wouldn't just test some random idea I test something that has a kind of some kind of in itive or explanation so there's a nice a nice saying which is if you're trading you're making money who's on the other side of that trade who's losing money if you're making money uh and then you ask yourself the question I know do they know they're losing money

37:10 and are they prepared to carry on losing money so this example that's why you know behavioral kind of risk premum strategies are so good because you know that on the other side of that trade is someone who is uncomfortable with this risk and as long as they don't suddenly become comfortable with it I know and human Behavior's been the same for hundreds of thousands of years so it's unlikely that that will happen them very quickly um then then they you know you

37:33 can just carry on taking their money effectively U if you find some weird anomaly that that's possibly being caused just because and it's just a kind of data mining effect if you like you can't really identify why it's happening it just seems to work well I would never touch a strategy like that because that's almost certainly overfitted um and will not actually make money out sample um so I'm I'm my first step is you know does this idea make sense then

37:57 yes I test it um and I I test it robustly so um if I do fit parameters I fit them out of sample um in a robust way so there's no danger of me making the thing look amazing just by as you say tweaking it um don't tweak um and um I like this word tweak in my first book I use the word fiddle don't fiddle with it but tweaking sounds more socially acceptable um and then

38:23 yeah if if it's something that adds value to my portfolio and and I believe in it um I'll I'll put it in and in fact this may surprise you but I I've got things in my portfolio that actually in the back test don't add value um but they don't detract from it either and because they're a source of Diversified returns I'm happy to put them in um which some people may find surprising um and there's other things

38:44 which I've tested and worked and I've not put in because I I don't believe in the idea enough um so the statistical test is important but it's certainly not to be all than end all there's a bit of Background by the the strategy just the p&l or the results of it but more like the log behind it and the reason why it works in the market yeah definitely so I'm I'm basically looking for things that I know I think will work and I I

39:07 could identify someone on the other side of the trade who's having to give me money um and then I I basically want to see a p&l stream that is not terrible so if if I if it if it like loses like 100% a year in the back test even if I believe in it I clearly I'm not going to put it in right because that that would be insane um but um as I said I'm happy to put things in that are

39:29 at best marginal and maybe even small losers because I believe that you know that that first of all I believe in intuitive idea and secondly they're diversifying um and um to get technical for a second the sampling distribution of correlations is lower than that for means or Shar ratios and what that means in layman's terms is if the if something you add something to a back test that has a a relative low correlation um and it doesn't improve

39:56 your performance well that that just might just mean that you've been unlucky with the the kind of performance over back tested period um and actually it's quite likely that because the correlation is low it will add diversification benefit and therefore improve returns out of sample um whereas something that maybe is very highly correlated but has a very high return um I wouldn't necessarily put in because I'm like well it's not really adding any value um and that higher

40:22 return may just be luck you know so um so I'm I'm I'm more likely to add things that are diversifi diversifying than necessarily improved performance outright I'm more likely to add things that that I think will work and I understand why they work than the things that are overp parameterized and data line I talk about this stuff for hours for sure but I want to SP your time here and what you findun they're going to could you see your work and what can

40:44 they reach out to see what you're doing uh yeah so there's a few places so if you're on Twitter my handle is investing idy or one word and um if you're um going to go and look at my website the main my main website is systematic money.org and from there you can get links to uh my blog which I log on fairly frequently and you can if you want buy all my books as well so um so

41:12 yeah those are the main and also go go go download my trading system if you want to use it for yourself if you know how to use Python obviously otherwise it won't be much used to you awesome I'll definitely leave a link below for your website and your books drop you can check it out as in the P note so they can see your book what you're doing there and also connect with you on social media as well Twitter and and

41:31 everything around so I appreciate you Rob it's been good to hear you talk about his topic it's a topic that I think very few Traders still talk about because it's not the most sexy Topic in terms of systematic trading it's more like a a hard work or or like like a more complete work process there but I love talk about love to hear you talk about for sure so thank you for the advice you gave here and look forward to

41:52 catch up with you pretty soon yeah it's been a pleasure thank you very much