My Algo journey 5 years deep
- This topic has 31 replies, 8 voices, and was last updated 1 month ago by
.
-
So its been 5-7 years with algotrading, most seriously the past 5 year i would say. Its a been a wild ride.. Im going to include a bunch of systems equity curves here so you can see how its been performing “in sample” vs live the past 5 years. All codes have been made between 2017-2022 i would say.
First let me just summarize the past 5 years btw i use IG and obviously prorealtime:
- It has not been easy, and im not rich lol. However i would say my account grew from roughly $20k -> $55k over 5 years, which is a nice number, but its currently at $35k. after taking some huge drawdowns since Trump (and tarriffs) rocked the markets.. My algos does not enjoy fast volatility, they mostly enjoy nice even calm swing trades.
- Ive taken my backtest “max drawdown”, added 20% ish to that number, and allocated enough contracts to each system so that my total max drawdown, if everything went to shits and all algos are at max drawdown at the same time, my total account drawdown would be -50%. This has happened to me twice and let me tell you.. thats painful. Ive listened (and read) many books of great traders (and algotraders) saying that you need yo withstand a -50% drawdown and build your way up again from that. It happens some times. And it has, twice, and its painful.
- Every single time you allocate more contracts to an algo you can just assume its going to be another “new high” for your max drawdown on that algo from day 1 lol.. its crazy, ive never once managed to actually get the timing correct.. Every time my account getsd +10k$ i try to allocate the funds to my top performing algos and wouldnt you know it, add 1 more contract = 10 losses in a row and profits for a year is suddenly gone lol..
- Real life results vs backtest is some % worse.. Slippage and cost makes the results i would say 10-20% worse than what backtest says.. Holding positions overnight eats away so much of your profit its crazy.. Last year i paid $8K in just overnight holding costs.. Ive tried my best to see how i can get it lower but most of my algos demand holding positions over night 1-2 nights or more on some and that hurts your profit really bad..
- Most of my algos are 15m, 30m, 1h and some daily. Markets i trade are Dax, Wall st, US500, US tech. Ive made all the algos myself
- Some of my algos use “dynamic stop loss” (example 5* ATR = Stop loss) but honsetly for many years ive just used fixed “stop Ploss”, same for profit and its been working good. I cant tell you why, when i started Wall st was not as high as it is now, but still it works.. and optimizing for a fixed number stop loss today gives me the same number as for the previous years.
- The Win % has gone down in Out of sample vs In sample, as expected, but as long as its profitable im running the algo.
- All algos are Long only except my daily algos are short as well when in bear markets
- Ive used trailing stop losses alot over the years and it works good for a year or 2, but removing them have ALWAYS given me better long term results.. Worse %winrate but higher avg wins making more then enough up for it.
- I try to never change my algos. I reoptimize every year to see if anything has changed, but making a new version of my algo almost never happens. When i discover something new like a new filter or new stop loss i try it out and if its better i change it, but i dont reoptimize variables ever.. Like i never change RSI 14 to RSI 10 next year because it looks better that spesific year.
6 users thanked author for this post.
Find exclusive trading pro-tools on 
Similar topics:
