You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To see the problem, play PyChess (level 8) vs. itself from the attached position. It will hang.
Direct cause: k4it.de changed its promotion flags. I'd just updated the numbers in my dev branch, but this shouldn't be fatal anyway.
This time I tried to get it right: handle exceptions while parsing, cache a "parse error" result as "no moves", and then fix the numbers.
Aside: Pychess still plays like an idiot in the attached position if the TB probes fail. They slow down the search so much that it can't finish a ply, so it plays a8=B?? basically at random. This suggests 2 ideas to try later:
Unless we're at the root or the depth is large (so searching the subtree would be slow anyway), don't probe. (Checking the cache is OK.)
Consider promotions in move ordering. Promotion to Q and to N (only with check?) can go with captures: they "win material". Underpromotions can go last because they're usually stupid.
The text was updated successfully, but these errors were encountered:
Original issue 653 reported by UncombedCoconut 2011-03-12
To see the problem, play PyChess (level 8) vs. itself from the attached position. It will hang.
Direct cause: k4it.de changed its promotion flags. I'd just updated the numbers in my dev branch, but this shouldn't be fatal anyway.
This time I tried to get it right: handle exceptions while parsing, cache a "parse error" result as "no moves", and then fix the numbers.
Aside: Pychess still plays like an idiot in the attached position if the TB probes fail. They slow down the search so much that it can't finish a ply, so it plays a8=B?? basically at random. This suggests 2 ideas to try later:
The text was updated successfully, but these errors were encountered: