Backtesting features

You're entering at minimum 4-5 years of constant development. No bar-loop based backtester really has a place in the market anymore.
You could build ML into it but by the time you're done with it, you'd keep it for yourself most probably.
You could also implement news trading (difficult) and various fundamental events.
 
What features in a backtesting program would you want bad enough that you would actually pay money to use?
At my previous place of employment I did spend a few grand on a consultant to build a nice backtest browser UI that allowed smart comparison, graphical selection from iteration results, pretty reporting to pdf etc. Not a backtesting program per se, but it was a tool I used a lot across various strategy classes - it had nifty plug-in risk and analysis modules so I can use it for both delta-1 in futures, options strategies or rates.
 
Man...don't go down this road. The road to "I will build a better backtester" is littered with corpses. Best to solve your own problems and/or write tools that can augment existing platforms.

I like the honest feedback :)
 
You're entering at minimum 4-5 years of constant development. No bar-loop based backtester really has a place in the market anymore.
You could build ML into it but by the time you're done with it, you'd keep it for yourself most probably.
You could also implement news trading (difficult) and various fundamental events.

Thanks for the feedback. Everyone is different and I'm trying to learn what people value.
 
At my previous place of employment I did spend a few grand on a consultant to build a nice backtest browser UI that allowed smart comparison, graphical selection from iteration results, pretty reporting to pdf etc. Not a backtesting program per se, but it was a tool I used a lot across various strategy classes - it had nifty plug-in risk and analysis modules so I can use it for both delta-1 in futures, options strategies or rates.

Interesting. This is helpful. Thank you.
 
You're entering at minimum 4-5 years of constant development. No bar-loop based backtester really has a place in the market anymore.
You could build ML into it but by the time you're done with it, you'd keep it for yourself most probably.
You could also implement news trading (difficult) and various fundamental events.

You know what? Your news trading idea is a damn good one.
 
Last edited:
Back
Top