Hey there, im glad you liked my post. Wanna add a couple of things:
When i first started i belived i needed 50-100 lines of code with crazy custom-made indicators and machine learning and 1000 variables all lining up to give me the holy grail of systems.
Now however i see that its much much more easy to run a bunch of “OK – Good” systems rather than tinkering making “perfect systems”. I didnt even know it but i discovered that this is also the mantra of one of the best traders out there: Kevin Davey. He’s preaching that running 10+ diversified, non correlated algos are far superior than running 1-3 systems that looks “amazing” in backtest, and i couldnt agree more. If you run 10 OK non correlated strategies, you might see 1-2 loose, 1-2 win, 4-5 doing nothing 1 day.. But 1 year later when all of your 10 algos are profitable you will have experienced (hopefully) a much smaller drawdown and a bigger runup because of the diversification of systems and hopefully markets and directions (long and short).
So far i got 12 systems running live, but the majority is Long only in stock indicies. Meaning when times are shitty and volatile in the stock markets, im experiencing drawdown because my systems are way too correlated. Im working on diversification every day now. Trying to short stock indicies, trying to create systems for other markets and so on. I have also done analysis on my current systems and their correlation and its actually surprisingly good considering that the majority is Long only. Because some are breakout and some are mean reverting, some are in the US, some in europe (dax), some are 30 min timeframe, some are on daily timeframe, they do trades on different days and different methods. So im getting some diversification but the stock markets are correlated for sure and 1 huge fake spike up could activate 4-5 algos that go into instant stop loss if the move up just comes crashing down the next minutes. its all part of the game..
Regarding my “amazing profits”, i have 1 algo that is my first profitable algo (the one im talking bout in the post “thoughts on profitable…”, its very much profitable today and doing a great job in the Dax even tho the Dax is down like 20% from peak high now. (long only) so its very possible. But I also have a code that is trying to “buy the dip” in nasdaq and seeing “dips break down” day after day after day, for months, has been rough. it has been bleeding money for the past couple of months, but hopefully whenever the markets return to new highs i will be ready.
my philosophy when it comes to trading systems are: Be 100% sure that even if markets go down -20% you will not bleed to death!!!!! Risk management is priority #1 when making systems. You should not have to manually turn on and off systems. If you do that then there has to be a reason. if you believe that the reason = true, just test it! If the reason is “finance numbers coming out” then just manually check the charts and trades on the days in your backtests where important numbers came out, check the results and do what you need to do accordingly.
So yea some of my strategies have been bleeding as the markets have dropped, some have made amazing profits while markets have dropped. that is diversification and that is “the holy grail” in my opinion. Keep creating more and more systems rather than increasing the amounts of shares/contracts youre buying.
Oh btw my most profitable Dax code consists of 2 conditions for buying and 1 for selling (+ stop loss, profit target, and trailing stop loss) So if youre not counting the trailing stop loss its basicly like 10~ lines of code tops.
Some of Kevin Daveys most profitable systems use no indicators! (Maybe 1 as a filter) but he uses alot of “Higher highs” and “Lower lows”. Maybe that will help you.
Also another tip that alot of the pro’s use is this: “Define the “Time” your trade needs to setup” So what exactly does this mean? Well if your on the daily timeframe and you see a move up testing the area around the highest high for the past 10 candles, you can try something like this:
condtion 1: high[3] > highest high past 10 candles
condition 2: close[3]< highest high past 10 candles
(So far meaning that the high of the candle 3 candles ago was above the highest high, but it closed under it)
Condition 3: close [2] < close [3]
condition 4: close [1] < close[2]
condition 5: close crosses over close [3]
This is honestly just something i wrote now from the top of my head but basicly it means that we see a test of the highest high, then a couple of candles going down before the re-test of the highest high level and it breaks above the last 3 candles.
I have now defined the “time it takes for my setup” and the whole setup. I could go more spesific and say that the difference between close[3] and close [1] has to be such and such++
But yea thats an idea that dosnt require any indicators, and maybe it would work better if we said “Close > daily average[200](close)”, maybe not. test, try and so on.
