Q: Problem with backtesting my EA

Umar Ismail 2014.10.31 00:03 #1

you see the quality of the data, if 90% then the result will not change

sorry i use google translation

LEONG KAH LOON 2014.11.02 06:17 #2

jansb000:

Hi,

I run into some problems when backtesting my EA. The problem is that when I repeat the test several times with the exact same settings and same dataset I get very different results. When I apply my EA to a chart in my demo-account it works fine.

Does anyone have suggestions of what might be wrong?

Thanks in advance...

Usually back testing result cannot use to justofy the successful of the Expert Advisor, as both are totally different. Therefore, most of the ForEx Robot Company disclaim that do not use back testing result to justify their robot.

Twan Jansbeken 2014.11.03 08:45 #3

jansb000:

Hi,

I run into some problems when backtesting my EA. The problem is that when I repeat the test several times with the exact same settings and same dataset I get very different results. When I apply my EA to a chart in my demo-account it works fine.

Does anyone have suggestions of what might be wrong?

Thanks in advance...

When I run the backtest for the first time on data for 2014 it does like 5000 trades. When I repeat the test for a second time it does just 4 trades. That is such a huge difference that I can not explain this by a lack of data quality. There must be something more fundamental.

In my EA I use TimeCurrent() for some checks. According to the documentation this should be the server-time at the last tick. Will this time be correctly updated during backtesting on historical data?

REQ HA Offline Chart Timeseries access and documentation Backtesting/Optimization

Stanislav Korotky 2014.11.03 13:41 #4

MT4 or MT5? MT4 does not have spread history.