Quality of ForexTester historical data vs Alpari
Posted: Sat Jul 20, 2013 7:21 pm
Hello,
I have used your data (from 'download' section) to check my systems on MetaTrader4. I have checked about 20 systems, found about 5 good systems (backtesting on 12 years gives good results and stable growth).
So I have decided to test my best system on historical data from Alpari (as their data are also considered good ones). And the surprise came - the system works terribly on data from Alpari. (I have checked that it is not a problem with symbol parameters, spread, timeshift etc.).
I thought it may be because the system is not good or its parameters are too tight. But I have checked the other 4 good systems and the result is the same - they work well on backtest with last 12 years of history from ForexTester data but they completetly fail on Alpari historical data.
If that was only one system, I could conclude that system or its parameters are bad. But if the same happens for other systems too - I wonder, maybe one of these two sources of historical data is not so good after all?
What could be the reason that not only one, but all systems fail when compared backtesting from these two sources (ForexTester and Alpari)?
Best regards!
PS My previous thread was here: http://forums.babypips.com/expert-advis ... lpari.html
I have used your data (from 'download' section) to check my systems on MetaTrader4. I have checked about 20 systems, found about 5 good systems (backtesting on 12 years gives good results and stable growth).
So I have decided to test my best system on historical data from Alpari (as their data are also considered good ones). And the surprise came - the system works terribly on data from Alpari. (I have checked that it is not a problem with symbol parameters, spread, timeshift etc.).
I thought it may be because the system is not good or its parameters are too tight. But I have checked the other 4 good systems and the result is the same - they work well on backtest with last 12 years of history from ForexTester data but they completetly fail on Alpari historical data.
If that was only one system, I could conclude that system or its parameters are bad. But if the same happens for other systems too - I wonder, maybe one of these two sources of historical data is not so good after all?
What could be the reason that not only one, but all systems fail when compared backtesting from these two sources (ForexTester and Alpari)?
Best regards!
PS My previous thread was here: http://forums.babypips.com/expert-advis ... lpari.html