Can't replicate Strategy Tester results
It's weird again. I was testing my EA and then suddenly I couldn't replicate the results on the same time range with the same parameters. I wasn't editing the code, just changing very few parameters. Even optimization gave at best only 75% of the previous profits.
Why did this happen? I wish I knew. Are there other people who were baffled by the same thing or is it just me ?What spread did you use if i may ask (and you are talking about MT4 right ?)
There are two main things to be considered:
- Spread (in MT4 you backtest with fixed spread, you need at least to choose a real value, consider that if you trade from 23 to 01 having reliable backtest is pretty impossible and all "night scalper" here on the market reports fake results and are scam).
- Tick accuracy (if you system works with only closed candle prices is not a problem and backtest can be same as reality. If you use indicator values from Bar 0 and Ask/Bid prices or any function that are tick influenced and use Ask/Bid prices for reference, like trailing stop you will soffer from tick values rounding and all results become unreal)
What spread did you use if i may ask (and you are talking about MT4 right ?)
There are two main things to be considered:
- Spread (in MT4 you backtest with fixed spread, you need at least to choose a real value, consider that if you trade from 23 to 01 having reliable backtest is pretty impossible and all "night scalper" here on the market reports fake results and are scam).
- Tick accuracy (if you system works with only closed candle prices is not a problem and backtest can be same as reality. If you use indicator values from Bar 0 and Ask/Bid prices or any function that are tick influenced and use Ask/Bid prices for reference, like trailing stop you will soffer from tick values rounding and all results become unreal)
I was not comparing backtest with live trading, but the same backtests a few minutes appart.
I've encountered another problem since: there was an unexplained deadly downward slope in the test. How did I solve the issue?
By downloading the historical data again!
Why ?
Now with more fine tuning I have succeeded in bringing back the profits close to 100% too so I don't worry much about it.
In MT4 if the tester spread is set to Current and the spread changes you can get two different results.
In MT4 if the tester spread is set to Current and the spread changes you can get two different results.
Yeah I know but like I said the spread was the same.
It is surprising how many people consider that when the spread is set to Current for each test, that is IS the same. So lippmaje's point was valid.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
It's weird again. I was testing my EA and then suddenly I couldn't replicate the results on the same time range with the same parameters. I wasn't editing the code, just changing very few parameters. Even optimization gave at best only 75% of the previous profits.
Why did this happen? I wish I knew. Are there other people who were baffled by the same thing or is it just me ?