Who can I trust (different results in backtests)?

New comment
 

Hello, I am developing an EA but when I run a backtest, I can see a big different result between Meta Trader 5 and my broker (it got its own software).

Both of them have a 99-100% history quality and the modeling is "Every tick based on real ticks".

Who can I trust regarding my backtests?

How can I see who is wrong?

Testing trading strategies on real ticks
Testing trading strategies on real ticks
  • www.mql5.com
The article provides the results of testing a simple trading strategy in three modes: "1 minute OHLC", "Every tick" and "Every tick based on real ticks" using actual historical data.
 
Please don't post randomly in any section. Your topic has been moved to the section: Expert Advisors and Automated Trading
 
Aurelien Pierre Laval:

Hello, I am developing an EA but when I run a backtest, I can see a big different result between Meta Trader 5 and my broker (it got its own software).

Both of them have a 99-100% history quality and the modeling is "Every tick based on real ticks".

Who can I trust regarding my backtests?

How can I see who is wrong?

Hello, the reason you get different results on different brokers is because of the following reasons ;

  • GMT difference between them (in case your EA is based on time or open trades on specific time of the day)
  • Spread difference
  • Swap and commission charges difference
  • Ping difference
And there are some other factors also which may show you different performance results on different brokers, so what you can do is, do some testing on different brokers and choose the one on which your EA performs the best.
 

Usually you get same backtest results in different brokers if your trading decisions are based on candle closure in higher time frames. It is risky to decide based on ticks or in lower time frames.

 

Ok, so it is better to make backtests with the choosen broker and not with Meta Trader.


I don't use high time frame and not small time frame (between M15 and H1).

 
Aurelien Pierre Laval #: Ok, so it is better to make backtests with the choosen broker and not with Meta Trader. I don't use high time frame and not small time frame (between M15 and H1).

It is always with MetaTrader. You probably mean a MetaQuotes demo account.

It is best to always use broker supplied accounts, be they real or demo accounts.

 

I am not sure but my current broker is [removed by moderator] and they have a software I use to make backtest and to trade with real accounts:  [removed by moderator]

I can remove the link and the name in my comment if needed (I don't want to be in trouble), it is just to be clear.

Is it what you mean?

 
Aurelien Pierre Laval #: I am not sure but my current broker is [removed by moderator] and they have a software I use to make backtest and to trade with real accounts:  [removed by moderator]. I can remove the link and the name in my comment if needed (I don't want to be in trouble), it is just to be clear. Is it what you mean?

The software you linked was MetaTrader 5. That is what this website is about — MetaTrader and it's MQL programming language.

Brokers simply add cosmetic branding to it, but the application is the same for all brokers that support MetaTrader.

 
Aurelien Pierre Laval #: Is it what you mean?

The differences in back-tests can happen when you test against different accounts and brokers, especially for non-centralised markets like Forex, etc.

Different brokers have different liquidity providers and can have different spreads or other trade costs or operate on different time-zones. All these can contribute to different results.

Its not a question of who to trust. It is simply that they are different and it is up to your strategy and your EA to be able to adapt to the differences.

Also, the future is not yet known. If your EA is unable to adapt to these differences, how is it going to be able to adapt to future events?

 
Aurelien Pierre Laval #: Ok, so it is better to make backtests with the choosen broker and not with Meta Trader. I don't use high time frame and not small time frame (between M15 and H1).
Yes that’s right, backtest with the chosen broker, you will be trading on and optimise your EA according to your chosen broker, by reading the above info you provided, if your broker has been removed for some reason then make a list of few best brokers and backtest/optimise your EA on those brokers and in the last choose the one with the best performance. In this way you will be able to get the best outcome of your EA.
New comment
Reason: