Is Strategy Tester a reliable tool?

Hi, All

I made dozen of tests today, seems having a lot of trouble. the test result was not consistent through many efforts.

I am using same EA, same set of data and its replica ( using export function in the historical center), the copy of data was applied to another platform which is a copy of the original one again.

then, I tested the same setting of the same EA in the platform and its replica using same data/ same timeframe, but, the results are totally different.

another observation is:
I shut down the platform, and then, re-open it, then continue the back testing with same EA and same data under same timeframe. but, the result could be 1mio difference between the before and after of platform shutdown.
Is there anyone who had similar encounter to share with?

In my humble opinion, it is fine for testing daily strategies (or H4 strategies that are not extremely reliant on indicators) ; for very short timeframes the tick interpolation causes weirdness and unreliability.

Pipero

Thanks for sharing. do you have any alternative solution?

forward testing, or using other products like tradestation or neuroshell with tick data if thats not too expensive for you.

You don't need to go down to tick level to get indicative results, in fact if you do you are probably curve fitting noise. I have found the MT4 back tester is accurate, however it is very easy to make mistakes in EA coding which exploit loopholes and give erratic results (e.g. trading on a close in the middle of a candle). There are however many other back testing tools, for instance, Wealthlab & TradingBlox.