"Confessions Of A Champion Algo Trader" - Kevin Davey | Trader Interview

people's biggest fears I don't have ideas and a lot of people think they have to come up with their own ideas where they have to look at a chart for example and and develop their own theory of the market or come up with some radical new indicator you don't have to do that so give indeed he will come back on the podcast how are you doing today great great how are you doing it's been a long time I spoke with you on the podcast we did this many many episodes back I think a year or

so ago and you just published a new book which is a witch something I'll talk about for sure but to allow me to kind of our goals trading and stuff like yeah but for people letting you write down I didn't leave this with the past episode tell people who you are just a bit about going back to yourself first of all I'm a full-time trader I've been trading full time probably for about 11 or 12 years I was able to make that leap from part-time trader but you know most of us start out doing the

full-time trader because I had some success I did some trading contests and I finished in first or second place three years in a row and that kind of made me think hey I could do this for a living and you know trading is what I really like to do I have an engineering and business background but I always love trading and so I was able to make that leap to full-time training and then probably about five years ago or so I released the book through Wiley it was called building winning algorithmic trading systems and that

was really part of it was because people had known I had won this contest and said well hey how do you actually trade what do you do and some of the writing of the book was just for me it was to kind of document my whole process and what I found was just writing the book really helped me in my own trading so after that book came out a lot more people started asking hey do you teach can you teach a class and that kind of thing and so I started doing more of that in

addition to my full-time trading and so you know for the past five years or so I've been trading and also teaching a workshop that's based in part on the book and then last year I released a book on algo trading and introduction a pretty basic book but for those you out there who don't know anything about algo trading it kind of just describes the terms and kind of helps you figure out if it's for you and then this past April I released another book it's called the entry and exit confessions of a champion trader so

basically what I did was I went through years and years of all the strategies that developed and documented a little over 50 different entries and exits that I've used in systems that have traded that I've tested and evaluated and put them into this book and so this book isn't meant to give people complete systems that they can just go in you know tomorrow immediately start trading but what it does is it gives you ideas because I talk to traders a lot and one of the biggest things that comes up is people say I have no

ideas to trade I don't know what the trade you area I don't know what to how did how to come up with ideas and you know when to end or when to exit and so this book kind of helps with that because it gives you a whole bunch of ideas that have actually worked for me and then you can take them and actually go and trade on your own you know what works in Japanese yen might not work for crude oil and vice-versa so you have to actually test these things but that helps out a

lot when you're trying to algo trade is having these ideas up front and you know then it's just a matter of testing them and trying them yeah and one thing I like by going to the book and reading it quite heavily is the fact that you provide some ideas like a conservative work and then how you would code them in the script for example which is something that I have not seen done before it's quite nice who can know kind of how they were put it in the code to be able to code not go

I guess so I can only hear about how is your progress over trading was it I'll go first from the start or kind of a transition later to our goals and automation way back when so I've been trade over 25 years or so and in the beginning I was just kind of searching for anything so I was doing some what I would call system trading you know like moving average crossover it's real simple stuff but I also looked at discretionary trading based on price action that kind of thing I looked at something called scale trading

which was basically as the price went down you kept buying and as it went up you'd sell a little bit over the done option selling I've done spread trading you know where you do things like by July we sell December we I've tried tons of different things things based on just big chart patterns you know Head and Shoulders patterns that kind of thing trendline trading and eventually I just got the point where I found a lot of things just didn't work or at least I didn't work for me like for example with discretionary trading I

just was not a very good discretionary trader and you know that would be a trader who'd be looking at a chart during the day and when he or she would see a setup that you might oh this looks like a good time to buy you'd pull the trigger and buy and you know you put a little stop-loss and a profit target that just never worked for me and and part of the reason was I think I the engineering math based business background and so I like numbers and I like understanding that things have worked in

the past I like testing things so for example if I see a doji bar and you know a lot of people say oh that signals a reversal well it might but I like to test it and so that's kind of where my trading evolved is to hey if I can test it and prove that it works you know whatever instances I can then incorporate it into a strategy so once you do that once you find some things you can build a strategy around it and then test it make sure it works and then you can

go and trade it in the markets and you can automate it if you want and just have it trade for you all the time and so that's kind of where I'm at now where I have a bunch of different strategies somewhere on the same market some are in different markets and what they all have in common is they all execute to set rules and then they're automated and so they just take trades automatically so all's I have to do really is just monitor and make sure that trades are being taken and that things are kind

of lined up the way they should be so that's kind of how my trading is evolved from you know just the the real simple stuff to what I'm doing today which is a little bit more complicated but overall I'd still think it's fairly simple at least for me it is and the question is going to come up of course some point is for people that don't know how to code so did you know how to code first or was it something you had to learn to go through I had to learn to code so having

an engineering background I had learned some computer programming and really if you're gonna follow this path you're probably going to want to learn to code because there's a couple routes you could take obviously you could learn to code in whatever platform I use tradestation which uses something called easy language it's pretty simple to to pick up quickly you can do that or you can hire somebody you know you can hire a programmer and you can tell them hey when there's a breakout of you know twelve bars or I want to leave that open any number

of bars I want to buy the next bar and you give them all your rules they'll program it the tough part about that is one you're going to be working with a programmer it's going to be that turnaround time and - a lot of the ideas you come up with aren't gonna work there's it's amazing you know people think hey they come up with an idea and they expect it to work most ideas don't work especially when you add in slippage and Commission you know trading costs most things end up losing money so you got

to try a lot of different things so if you're gonna program have a program or do it it's gonna cost you a lot of money and then the third way which is starting to come a little more popular some software vendors out there offer some visual tools or you know kind of drag-and-drop of you put a box that will calculate the RSI and you connect that up and you connect it to some rules now I've done some of that with metatrader4 where there's a group out of Australia that have worked with that's created a really

slick visual tool and you know I was able to create some strategies where I can't program the Metatrader but I could create it visually what I wanted and then you hit a button it generates the code now not every piece of software out there every trading platform not every trading platform offers that but that might be an option too but I think eventually for the most part you want to understand how to program and if you really want to get into it I think in the long run you're gonna want to learn the language that

you're dealing with whether it's trade station or multicharts and ninjatrader you know there's a ton of them out there so you know one question people might ask is like wow how do I even know what language you know I might be able to understand because I look at some computer languages and I'm sure you do too where it's just like wow this is confusing so one thing I found one trick that people can use get a copy of technical analysis of stocks and commodities magazine so that's a monthly trading magazine and what they do every

month is there'll be one article they'll provide the code in maybe a dozen different languages so they'll have it in tradestation ninjatrader neuro shall there I can't even think of all the ones that are in there but what you can do is get the magazine and look at all the code and see which one kind of feels good to you you know chances are you gonna see one you're gonna I don't understand what they're doing but maybe another one you'll say oh I kind of understand how they took these rules and made this code that

might tell you what language would be easiest for you to learn so that's one way to kind of see things and to you know get started because there are tons of different languages out there and you know each platform adds its own basically so that's one little trick that's pretty cool chick how do you personally find your ideas of strategies to practice do you find them from as I said magazines or I guess a bunch of different places right yeah and this is something I teach in more depth in my workshop but that as I

mentioned the beginning that's one of the people's biggest fears I don't have ideas and a lot of people think they have to come up with their own ideas where they have to look at a chart for example and and develop their own theory of the market or come up with some radical new indicator you don't have to do that you can go out there there's tons of things whether it's on the internet it's in books magazines trading forums there's tons of things that people have done now though those can all be a starting point that's

the way I kind of look at it I almost never take something I find in a book or magazine and just go with it and you know start trading with it what I'll do is I'll take it and I'll modify it I'll look at it and I'll oh this kind of makes sense or may see I do the opposite of what they're saying and I kind of play around with it but a lot of my core ideas come somewhere else you know somewhere off the internet from for example and there's tons of things out there

and that's a mention monkeys I where I can sometimes I give the background or you know where I think the idea came from I say a lot of times I don't know where the idea came from it might have been me originally but more than likely it was probably somewhere else but I might have taken it and transformed it into something that I liked and that's what I encourage people to do my goal I always tell people that take my workshop is you might come into it think you know I don't have enough ideas to

test but at the end of it with the sources of ideas that I provide you end up with so many ideas that you're not going to have the time to test them all then that becomes your big thing oh man I just don't have time to test all this and and it happens to me I have a list I could probably test non-stop for months and months and probably years and still never test everything I want to test it's just impossible and then the other thing is so once you have an idea that you want

to test and then you test it in the city results are bad are you gonna try to make it work and prove it tweak it or you're gonna kind of drop it and go to something else well yeah that that's a great question because what happens is a lot of people become led to their ideas you know meaning they think it should work and so then they'll start forcing it and the the process I use I ran into the same thing in my own trading I used to force ideas oh I've got to get this

moving average with an RSI combination it's got to work it mate everybody tells me it should work so I'm just not testing it right and and so I'd force it you know had filters add conditions that kind of thing and what ended up happening was I saw that yeah I could eventually force it to work but when it came to real time trading those kind of things almost never worked so based on that I kind of adjusted what I do when I test and so now what I do is I actually incorporate what I call

a preliminary or initial feasibility for a fancier term test where I do a short test on a limited amount of data and I just see does the idea look like it's gonna work and if it doesn't I just move on to the next idea so you know that works obviously if you have a lot of things to test a lot of ideas both strategy code but a lot of markets time frames you know because an idea a certain strategy code might work with one market but not with another so you got to have a test

process that will allow a bunch of different markets a bunch of different time frames so what I'm looking for there is just some indication that maybe there's something to this idea and usually you can tell that pretty quickly and when you do you don't have to waste your time with most ideas because like I said most ideas turn out to be useless or you know that they lose money when you include slippage and commissions what's interesting is most ideas work when there are no trading costs so you can have a moving average crossover for example

which is really simple you can find it works in a lot of markets when you don't have trading costs and you can also find interestingly enough that the reverse of that works you know a reverse moving average crossover that works a lot of the time too but when you put trading costs on either one that kills most trading systems and unfortunately that's just a real-life thing you're going to have trading costs so it's better just to test something a little bit upfront it doesn't work you move to the next idea the people I've seen get

into trouble tend to test and test and test an idea and won't let go of it you know they they for whatever reason they think they must make that idea work and eventually you'll get you'll probably get any idea to work if you add enough conditions and filters and you know if you have like bull markets and bear markets you just specify exactly where you're gonna trade it you can find everything will probably work at some point but more you do that the less likely it is to work going forward and that's that's something most

people don't understand and that's the other thing I was kind of wondering about is the fact that you only have mechanical filters mechanical rules loose trash is so nothing like manual that you look at the chart now why is that because you don't want to be looking in the charts or because you think there's no point or some reason the mixing is I like to use history as my guide you know they always say past performance is not indicative of future results and that you know that's true but at the same time I like trading

what I know has worked in the past so take a simple candlestick pattern you know if I build a system with it and it's produced profit for the next ten years I feel pretty comfortable going with it and trading with it going forward where if I would to put where to put some manual kind of override on it everyone feels comfortable because I don't know if it works or not and it's hard to test some things are hard to test they're important might be impossible to test but there are people out there who do what

you know I I've heard it called grey box you know a black box would be all your rules you never deviate from them you just follow them a grey box would be you have your rules but you can override them based on your judgment now for me personally just knowing my discretionary trading my judgment I don't think would help my trading but some people might find hey you know my judgment is really good I guess the key to that is having it historically prove itself you know so you do that you know for the last

year hey when I override things get better well then you probably could do it for me personally I don't think I can but there probably are people out there who could do that so for example something I see for myself is the fact that I really believe that the market context and the like the way the best for you trade something is important to your trade to your result so I couldn't find a way to replicate this encode that kids are selling microcut XOR which kind of character context in the code you know with this

yourself or do you kind of put this aside and believe it isn't helped at all for your trading well if I can't code it I kind of just ignore it and and that's not to say that it might not work for some people I just don't think it would work for me because you know for one I'm looking at trading dozens of systems at any particular time and so it's hard to know it would be hard for me just to control all that you know oh yeah my crude oil system do I override that today

but not my gold systems you know that's what would make it tough so what I found is there's there's more benefit for the way I trade into just trading all these multiple markets realizing that maybe I'm giving up something by not having that kind of override but I can certainly see the other side of it where you know if you have a way that's mainly mechanical you know meaning rule-based and then you had some kind of judgment override and it works for you you know definitely I'd say stick but my philosophy is more things you

can test in the long run probably the better off you're gonna be mm-hmm interesting so how do you manage the risk with all those systems at the same time because of course someone could start as a beginner in thing that they can have many many systems but they're all current for example and then they lose the money in one thing so how do you make sure that you don't have too big of a drawdown or that they don't kind of interact together your systems yeah that's a great question so a lot of it depends on

this and the systems you pick that you have so for example I last January when in 2018 back when the stock market kind of crashed real quickly and there was all that volatility I realized unfortunately a little too late that I had too many stock indices systems you know and they all kind of became correlated and where even historically they weren't that correlated so correlations a big part of it you definitely want to look at that and see what these systems can do but beyond that what they've done is story Berkeley is you also got

a sure balance as far as sectors so the way I look at the futures market you have your stock indices you have your metals you have energies softs currencies interest rates those are probably you know the big ones a lot of people break up agriculture and softs but if your equal and each kind of sector chances are that's one way to diversify where you're not going to be caught in any real bad correlation events but that you know it doesn't mean it's not going to happen you could have a system that is just one in

crude oil one in gold and something could happen where they both work against you at the same time where maybe normally they don't so I look at a lot of little scenarios like that of just how they correlate with each other over time and so there's a couple different ways that I teach that you can avoid that correlation obviously one is different markets but then different bar sizes to you know maybe 30 a system based on 30-minute bars could be a lot different than a system based on daily bars for example and then the other

thing obviously is just the the code itself or the base idea you know maybe use breakout systems and maybe you have some mean reverting systems that are anti breakout you know instead of going with the breakout they fade the breakout so you try to put all those together and you know you do your correlation analysis and I also run simulations a lot of money Carlo simulations just to see well how bad could the drawdown get if they all go bad and that helps but ultimately there's always some unknown out there that you can't quite get

all the risk out of it you know you can't just by trading say a hundred systems doesn't eliminate the correlation if you say all these systems are different therefore I've never correlated so I don't have to worry about that there's still new cases coming up all the time new situations where you're gonna find some correlation issues eventually yeah yeah so like news events or things that could impact and different weighting before stuffing debt right so you mentioned the length of calcic or bars the other preference of like day trading or string trading with your systems

or do you kind of mix it with different doctrines I like to try to have everything so I'm kind of open as what I'll what I'll do what I've found though is like intraday systems ones that exit by the end of the day those are typically the hardest create where the easiest ones are the easier ones I should say are probably based on daily bars and and those tend to hold overnight you know the problem is with those a lot of times sometimes you have to give away give back some open profit because those tend

to have bigger draw downs inter trade where maybe the intraday trading you know if it doesn't work you just get out right away but those can still have drawdown issues too but the way I kind of look at it I try to leave myself open and whatever tends to give me the best risk adjusted return is what I'll actually trade now I'd love it I'd love it every day if I was out by the end of the day and it was just intraday and I could make a good risk adjusted return with that but what

I found is some swing trading some longer-term trades that last days two weeks is almost always a part of it because you want to catch those big trends you know whether it's drought in the US that leads to soybeans and wheat going up or you know a longer-term situation with the central bank where a currency is being devalued where that's a long-term trend we want to be able to take advantage of those things too which you know might not show up on an intraday basis so I like to do both and I would say the

swing trading parts a little bit easier to develop systems for mm-hmm regards that the risk adjusted return do you have some measures you look for like a minimum of for example like one two three or things like that or is always dependent alike what other things you have already built so what the metric I use is based on some Monte Carlo which were simulation which is just taking all the existing trades and running some simulations of them and generally what I'm looking for overall is a return to drawdown that's at least two so meaning if

during the year I had a ten thousand dollar drawdown let's just say trading at one contract I would want about twenty thousand dollar return in profit yeah and you can look at it percentage-wise you can look at it dollar wise but that's generally return then you want risk and when you look at the pros out there you know the commodity trading advisors and those kind of people they generally hope for over the long term even a return to drawdown of like one so if they want a 10% annual return they're expecting at some point a

10% drawdown and you know if you especially if you do that over a few years that's usually a pretty good performance measure that's interesting so that's an issue people can look at and they could deserts well it's awesome it's any topic we didn't cover you like to touch on or any maybe kind of mistake people tend to make in that field of creating out goals or systems that you would like to point out you know I would just say it was something we touched on earlier the biggest mistake that I see people doing with algos

is just in how they backtest first of all because most software is so most trading software are so simple now that all you have to do is open up a chart insert a strategy that you've either created or a canned strategy that comes with it and immediately what it kind of pushes you to do is to optimize it optimize those parameters and people get caught up in the optimization part and they think that optimizing is the way to go and it's not I mean optimization is important to a point in you can do it and

improve what you're doing but at the same time all optimization really does is just curve fit the past meaning you get a better back test and a lot of people just keep trying and trying to improve a back test I mean I can tell you story after story of people who literally after every losing trade they have they go back to their system and figure out why it occurred and try to eliminate it and so they add a little rule Oh today was a Tuesday and it was a full moon I should have been going

long today you know that was dumb it was so then they put a rule in no full moon Tuesday trades and oh the back test looks so much better and that's what and people then think that back test is going to keep going forward and that's not it at all about building a better back test almost always leads to worse results and for me one of my big breakthroughs was realizing when a back test was good enough and resisting the urge to make it better and I'll tell you what that's a hard thing to do

with fun yeah yeah because you you see it and you see a draw down on your back test and you're like I don't really like that I could get rid of that by doing some financial engineering you know add some rules filters whatever but realizing what's good and that you don't need it to be better that's probably the biggest thing of all the mistakes I see people make creating a better back test is like pretty much number one I would say and on that note all your ideas it's just only four entries or they also

apply to exits they use it to train in my new book most of the ideas are entries probably 44 so our entries and about 10 ten or so little more net our exits but that doesn't mean you can't use the Andes and there's two different ways you can do that you can use every entry as an exit if it was a like a stop and reverse type of situation where first entry happens you just switch and go the opposite direction so that's one way to do it and the other way to do it is kind

of unique you could have an entry but then take another entry from the book and just make that an exit so don't make it an entry but make it an exit you might say a lot that's kind of weird I'll give you a case study a couple of years ago I had a bunch of students come to Cleveland Ohio where I'm at and we we had a weekend of basically some advanced strategy development topics and one of the things we did is a lot of group work and one of the exercises I gave everybody was

to create some simple systems that tended to work and one of the specific things I said is for exits don't use your simple of five hundred dollar stop-loss two thousand dollar profit target you know or breakeven stop those things don't use those just use what you would consider an entry condition but just use it to exit and they did that and what we found is sometimes those things were the best exits they were better in stop losses and profit targets probably because they were kind of price based market based you know they reacted to whatever

the the price was doing rather than being an arbitrary dollar figure for example which most people do with stop losses so that was kind of interesting so that if you think about it every entry you've ever used you could take that and just make an exit out of it and right there now you've pretty much doubled all the things you're gonna test you know you can test all your entries and then you can test your entries as just exits and that's kind of a neat thing that a lot of people don't do right off the

bat this awesome that's a really good tip which I didn't think of before so that's even awesome for me I love it perfect so I regretted what she got new book entry exit condition of a champion trader two ways professional theater gets in and out of more of the stock futures and forex market which is awesome and how can people find you they'll clinic with you or reach other serious podcast well obviously if you're looking for the book it's only on Amazon so let's go to Amazon just type my name Kevin Davey you'll see up

I've written three books so far you'll see them all or you can go to my website cage a trading systems dot-com and you know I have links to my books I also have links to my workshop a lot of free information you know typically I do probably one or two free trading webinars a month and there's also free of a free smaller book that's just some entry and exit conditions different ones and the ones in the book in the confessions book but they're just simple ones all you have to do is just give you give

me your email address and then and I don't spam people I just kind of tell them when I'm doing some other free things that would might be the best way to get a hold of me so KJ trading systems dot-com perfect links in the show notes as well for that so Kevin I thank you for being here on the podcast Rosie Hanson a pleasure and I just want to have anyone that Pfister dress with people what is one piece of acting like one action step you recommend them to do this week or this run to

progress what their goals and trainings that's a great question the first one would be know what your goals are you know start right at the beginning it and make sure you have goals for your trading don't just say hey I want to be rich and want to make a lot of money trading now try to be specific I want 20% annual return or 50% annual return and if you're going to do that also at the risk component but I don't want more than 20% drawdown or add I can only handle X dollar drawdown and at

least then when you have those figures you'll be off to a good start because then you'll be able to everything you look at you'll be able to put in context you know does this help me meet Rita reach my goals or are my goals to driving it I you be amazed how many people say hey I have five hundred dollars I want to trade full-time and draw a salary off it and you know I tell them you're not going to do that it's it's almost impossible yeah there's probably a few people who've done it but

you need to have some realistic goals so that that's probably the first thing everybody should do and if you haven't done that that's the step one it's make sure you have some realistic goals for your trading awesome sounds very good given to you so much pleasure and we'll catch you guys pretty soon all right well thanks for having me you