I am not sure where to put this question but as it is to do with an automated system I figured here would be a good spot.
When backtesting multiple systems you end up with a bunch of data regarding how well the systems performed. You have the number of orders opened, the number of times it made a profit, the number of times it made a loss and the number of times it broke even. You can calculate a win percentage (100 * wins/orders) etc…
What would be the best way to rank them? After all a win percentage of 100% is less impressive when you see that the number of orders it opened was 1. Does anybody use a formula that takes into account win percentages as well as number of opened orders etc…?
hi, what kind of software you are use?
I would like to ask a question (apologise for being off thread) regarding EAs.
I am not familiar with programming, but I am hoping to perhaps get an EA developed.
My question - Does an EA have to work on ‘close of candle’ or can it work on, for example - candle passing through an envelope like Keltner channel?
Thanks for any answer.
I wrote it myself using C#.
I pay attention to, among other things, the number of transactions, drawdown, the percentage of winning transactions
I was wondering if there was some sort of formula.
IMO only profits exceeding losses. Everything else is tinsel.
theres only a formula if you create one to determine a variable based on an objective. for example ur asking about ranking and EA. well first off its your EA so u would be ranking it but u have to determine wat is is that your trying to measure. like are you traying to measure the performance based on maximum account RoI or are you tryiing to measure the performance based on accuracy and drawdown or like you said win ratio? the way i rank mined is i combined to or mor factors that i want the performance to measure up to. i like verry low drawdown and account growth. i dont care about win ratio win im focused on maxing out the account growth for the strategy. but if you are concerned with leverage or maxing out margin usage you woul want to measure amount of trades or frequency or how many are opened at one time.
1 Like
it does not need close of candle factor to work. i have developed EAs using the actively counting down candle before it closes. if u want help let me know what the strategy is
1 Like
@YoungTrepTrader Many thanks! 
I’m such an old dumwit when it comes to any thing techy. I do have a programme developer on the case, but I like to try and give some technical input! Basically my system requires entry as the candle exits the Keltner channel and close when the candle re-enters. It’s using the Keltner as trend following system.
Thanks for your help.
I rank the backtests based on best final profit with the least max drawdown over the life of the backtest. Have built several profitable EA’s using this. I wouldn’t get bogged down too much on all the other metrics.
I’m not a bot guy at all. However…
In my experience, unless the candle is closing soon, don’t trust it.
If the candle is half way done, and is bullish, it could still turn bearish. You never know until it’s closed or at least near closing. You never know what could happen.
@dushimes
Thanks - my bot development has been held up by the dreadful war in Ukraine. 4XDev were working on it and they are (or were) based in Kherson and I haven’t heard from them since just after the start of the war. They said they would be ok and will be in touch. My bot can wait,
I just pray for them and Ukraine.
2 Likes
So instead of looking at single backtest strategy - check the strategy based on multiple backtests, if it has a real edge or just “edge of parameter luck”