-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Multi stage movegen #49
Conversation
A step towrads multi-stage move-generation
- We were not using it for ordering, as we were always returning 60000 fooling move picker to think every move has a history - We were adding it in the alpha cut-off not beta, prioritizing bad moves, which made Zahak in times play funny moves (due to late move pruning/multi-cut)
aececb1
to
0dddbd0
Compare
A 1m+1s match ended, and this PR one wonderfully:
|
Another match, with longer time control (15m / 40 moves) is going on with the zahak_{master,dev, and this PR} |
The match with the longer time control didn't exactly go well (well, not as good as I want it anyways), my history heuristic change didn't play nice in longer time controls, hence I pushed this commit: 81f00b0 Now, I am running two more matches with both short and long time controls, and see |
The result of the long time-control match:
|
Match is finished with great success:
|
and this is as per bayeselo:
|
This includes #50 and true multi-stage move generation
alongside various bug fixes, which were uncovered thanks to this PR
A match is going on, and this PR is having a lead even over #50, let alone the others