Strategy optimisation results
- This topic has 17 replies, 5 voices, and was last updated 4 years ago by
.
-
Hey Thally, i get what youre saying, and less rules = better according to more or less all the experts out there, doing what we do.
Adding a filter that lets you keep the winners but shave off say 30-50% of the losers, would help dramatically, without necessarily curvefitting the whole system. Lets say for example you say “Close > 200 MA”, if that simple rule erases 30% of your loosers but lets you keep 95% of the winners, im guessing that would very much help your system without destroying it with 5 different rules.
You are correct about the “low win rate, many losses” = trend following. But usually the “low win rate” is around 40~%, not 20%. If your winrate is in the 20% range, its either a bad entry, or a bad filter or both.. I would highly recommend trying out 100 other filters too see what might work, or combining the ones you have today with more/other filters. As i pointed out i dont think your going to run this system for very long if you see 80 lost trades and 20 winners, with maybe like 5 really good trades that actually make it worth it. The system is going to demand alot of capital to be able to run for 6 months with just pure bleeding money. Also there is no guarantee that a huge winner will come tomorrow.
From my own experience, without trailing stop losses and too much crazy rules, my trend following systems when developing them, come at around 40-50% win rate. Then adding my money management such as a better stop, target, and trailing stop loss (if it helps the system) usually picks up the win rate to 50-55%
Think about what youre trying to do: Buy a breakout and hold as long as the trend continues. Defining the breakout (your entry) is the easy part. You can use “highest high of X periods”, you can use donchian, bollinger band, rsi crosses above X etc++ The hard part is the “setup” that needs to be in place.
Lets say your systems is “Buy if close crosses above donch channel”. Ok fine, thats going to create tons of false signals and thats gonna cost you every time close crosses above. You need some sort of filter in place. How about Buy if close crosses above donch channel + Price must be above “daily MA 200” or RSI 14 must be above 10MA on the rsi, or anything really that can filter out some of the bad signals.
Or maybe the breakout candle youre looking to enter must be above the daily high of yesterday. Or maybe it should be below the high, or above the low or under the close etc…
Maybe looking at the time of the day, say only trade during peak trading hours, or only during “nighttime”.
I dont know the future and i dont know if your system will be profitable or not. But i can tell you that it will be hell to sit through. Would not recommend running that system because sooner or later you are (most likely) going to stop that system in the midst of a downturn and its gonna cost you money.
1 user thanked author for this post.
Find exclusive trading pro-tools on 
