Backtesting (sub) community?

Hi,

I’ve been reading up on forex trading for a year now and have traded demo accounts to understand the system better. I’ve been reading posts in this forum passively and around the web during the time and gotten very curious about trading strategies and algorithmic trading. I’ve implemented my own strategies and ideas and got various results. I’m reading, learning and trying out more things on a daily basis.

I’m wondering what a good place is to find people that backtest trading strategies and algorithms in general?

Just to clarify, I don’t mean backtesting EA or scripts found somewhere on the web in MT4. I mean real backtesting with complete data sets and self-implemented algorithms. I don’t consider testing in MT, clicking a bunch of check boxes for indicators and trying out random rules proper testing without a good visual as well as statistical analysis of the results.

In case the question comes up. My background is in statistics and symbolic systems. Because I was dissatisfied with the existing tools and GUI apps I wrote my own backtesting backend and use large sets of tick data. I validated the correctness of my implementation by running my scripts through Oanda in a demo account for 4 months and after that backtested with my implementation to verify outcomes.

I see a lot of threads where people try out EA they found or post, or trade a strategy in a demo account and try to generalize from a month of trading about the performance of their strategy. To me it seems more than surprising how one could prefer this approach to throwing years worth of data at an algorithm to see if at least for the past it would have generated positive results. (Yes, past data cannot predict the performance in the future, but at least you get to judge more than a month’s worth of trades.) I actually implemented a few of the algorithms posted here, since most of them are around 50LOC if at all to judge their performance faster.

I hope this post resonates with a few people. And I’m hoping to find people that would take a similar approach here.

Chris

1 Like