Which backtester to use?
Hi all,

I would like to use a (bruteforce/) backtesting tool and I have spend the last few days setting up BacktestTool. I ran into some issues, got most of them resolved but there are still a few things that do not work as expected. Upon trying to get some help on xFFFFF's forum, I noticed the project is basically no longer active.

The same seems to be the case for GekkoWarez backtesting tool...

Before I embark on yet another adventure with a backtesting tool which is no longer active: could anyone advice me on which tool is currently the most popular/active/easy/reliable/etc.?

I see Nickz's project, but not much action there it looks like.

Tommiehansen's GAB looks quite active, but installation doesn't seem thát straightforward.

I'm running Gekko v0.6.8 on Ubuntu 18.04 LTS.

Looking forward to your suggestions, thanks!
GekkoGA and Japonicus are the working ones.
For gekkoGA you need to add this PR:

Japonicus works, but it is a bit tricky to add all dependencies.
I personally use GekkoGA
I never heard of "xFFFFF's forum".Can you paste a link to it please?
Thanks Remo, I'll look in to GekkoGA too.

With xFFFFF's forum, I actually meant the Issues @ github: https://github.com/xFFFFF/Gekko-BacktestTool/issues

Should've used the proper term there.. Seems to me that xFFFFF himself hasn't been contributing since april/may last year or so.

Anyone else care to share their favorite backtest tool?
(01-09-2019, 06:46 AM)Remo Wrote: Japonicus works, but it is a bit tricky to add all dependencies.
I personally use GekkoGA

Looking at GekkoGa, the last commit was March 18th 2018. I think I would prefer a tool that is still actively maintained, more future-proof I guess?
Just to follow-up:


I've tested nickz's gekko-batcher, which works quite nice actually. Fairly easy to get it up and running. I found a bug and it was quickly fixed by Nicolay, so with regards to an active project; this looks promising. I've mainly tested the bruteforce option.

Personally I would also like to see the %winning trades, %losing trades, biggest win, biggest loss in the output file, but I guess that's a feature that can be added.

Nice tutorial: https://forum.gekko.wizb.it/thread-1535.html

What I like about this project is the smart-aspect of the genetic algorithm being used! I love to think that it's actually guestimating and trying, instead of just throwing some numbers out there. Setup was also fairly easy, as long as you pull "the 49" when using Gekko 0.6. I do miss the output in a file for me to browse through and compare/draw some conclusions myself. I'm awaiting answers if it's a possibility at all.

So for now, I'm leaning towards a combination of the two: let Gekkoga get me baseline which I can use in gekko-batcher to get a bit more insight.

Anyone care to share what they think about my findings and/or what they use?

Forum Jump:

Users browsing this thread: