EA backtest results different on different brokers
Hello
I created my own EA for my personal use
When I back test on one broker the result is profit for example $250 test balance with 0.01 lot end with profit like 10 or 20 percent profit, but when test the same EA on another broker data the result is loss
What makes the backtest results differ? and how to fix this?
Hello
I created my own EA for my personal use
When I back test on one broker the result is profit for example $250 test balance with 0.01 lot end with profit like 10 or 20 percent profit, but when test the same EA on another broker data the result is loss
What makes the backtest results differ? and how to fix this?
search google, or this site. There is hundreds of threads and comments discussing this. seriously? did you do a search?
I know the reason of different broker quotes but I wanted to talk and discuss it myself with someone
You can ignore my post if you just reply like this, thanks anyway
quotes may give the differe
I totally understand
But if results will differ from broker to broker then MQL signals also will be not reliable because it is copied on different brokers, so a signal with profit maybe a loss on another account with another broker
This is not correct, trade book is different from liquidity provider to another but generally the market is correlated, even pairs are correlated and cannot understand why completely different results
There may be several reasons for that:
- First of all, as already said, every broker may have different quotes, so a little discrepancy is normal and should be accepted
- If your strategy is very delicate (little SL,TP, or even trailing stops) discrepancies may become bigger
- If your system make trading decision based on current candle information (Shift 0) or indicators checked on current candle, discrepancies may become even bigger
- If your system is tick sensitive, discrepancies may be even bigger again (every broker have its own quotes, ticks may be even more different from each others than normal HLOC information)
In the end, a system that is profitable with a broker and not with another, is more likely a system that is just overfitted on a specific broker trying to find the perfect equity curve, instead of focusing on the trading approach.
The same discussion may apply when you do massive optimizations using dukascopy quotes/ticks: no broker have same datafeed and (even if dukascopy is considered as good quality data), the real trading results may differ a lot from backtests.
How to avoid, or at least mitigate discrepancies:
- Use not little SL/TP. Generally speaking not less than 1/10 of daily range.
- Trade low spread symbols
- Check candles information (open, high, low, close) or indicator values starting from last closed candle backward. Never check them on the current candle
- Avoid using trailing stops, or at least, code a trailing stop to move SL based on new candle formations instead of each tick change (like 99% of people still doing)
There may be several reasons for that:
- First of all, as already said, every broker may have different quotes, so a little discrepancy is normal and should be accepted
- If your strategy is very delicate (little SL,TP, or even trailing stops) discrepancies may become bigger
- If your system make trading decision based on current candle information (Shift 0) or indicators checked on current candle, discrepancies may become even bigger
- If your system is tick sensitive, discrepancies may be even bigger again (every broker have its own quotes, ticks may be even more different from each others than normal HLOC information)
In the end, a system that is profitable with a broker and not with another, is more likely a system that is just overfitted on a specific broker trying to find the perfect equity curve, instead of focusing on the trading approach.
The same discussion may apply when you do massive optimizations using dukascopy quotes/ticks: no broker have same datafeed and (even if dukascopy is considered as good quality data), the real trading results may differ a lot from backtests.
How to avoid, or at least mitigate discrepancies:
- Use not little SL/TP. Generally speaking not less than 1/10 of daily range.
- Trade low spread symbols
- Check candles information (open, high, low, close) or indicator values starting from last closed candle backward. Never check them on the current candle
- Avoid using trailing stops, or at least, code a trailing stop to move SL based on new candle formations instead of each tick change (like 99% of people still doing)
The EA buy and sell when the conditions apply on the broker data
Yes the data between brokers are different in time but same in general trend on the market so the data maybe different from broker to broker but still trend is same and indiscators will show the same based on the broker data
What I mean is the strategy work on trend and indicators that are almost same on the market even if different brokers and different tick quotes
So the results if different should be little different but not extremely different
The EA buy and sell when the conditions apply on the broker data
Yes the data between brokers are different in time but same in general trend on the market so the data maybe different from broker to broker but still trend is same and indiscators will show the same based on the broker data
What I mean is the strategy work on trend and indicators that are almost same on the market even if different brokers and different tick quotes
So the results if different should be little different but not extremely different
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
Hello
I created my own EA for my personal use
When I back test on one broker the result is profit for example $250 test balance with 0.01 lot end with profit like 10 or 20 percent profit, but when test the same EA on another broker data the result is loss
What makes the backtest results differ? and how to fix this?