Testing the same EA for exactly the same period for historical data for the same symbol from the Alpari download website vs historical data obtained via History Center (F2) from the terminal, I am getting quite different results.
I understand there will be differences in the data from different brokers, but the differences I am seeing are as if the results are for completely different symbols. Any comments on why this might be?
Given that they are different, which one would you say would be closer to live trading conditions?
One reason i can think of is the modelling quality. If one has 90% compared to 25% there will be a difference. Are both 90% then small differences could be there regarding differnt spreadratios but with both 90% and huge differences imho it smells like manipulation.
Can that also be due to the curve fitting process? I’ve heard due to this process backtesting results are not that accurate
I agree, for me which matters is the spread and especially how brokers support my performance. It seems interesting more tools on the platforms than usual but this could block me. I’ve a plain platform and I don’t really need more than that
Yeah, I have had the same experience getting different results between downloading it from the metatrader history center compared to using the actual tick data of Alpari.
What I have found is that Alpari data is the one you should use because it is actual history tick data, whereas the metatrader download is generated in some way.
Here is a link to Download Alpari M1 data for all 7 major currency pairs 2001 to 2009