vdbergh
vdbergh
The SPRT test in cutechess-cli assumes that the outcome of each game is independently drawn from a trinomial distribution. This assumption is not correct for an unbalanced book when games...
Steps to trigger - Make sure that the default state of the "Finished" button is "Show" (i.e. the content of the corresponding div is hidden). - Press the "Finished" button....
I am creating this issue to report workers with issues. E.g. currently. https://tests.stockfishchess.org/actions?action=failed_task&user=Oakwen&before=1655180683.077&max_actions=100
Occasionally there are requests for running the worker on a RPI. To make this possible we need a cutechess-cli binary for the RPI. Perhaps someone who is familiar with the...
https://tests.stockfishchess.org/actions?count=3&before=1637403710.599 This is an assertion error in the conversion to pentanomial which is a bit scary. However it seems that all the Windows workers were affected and no Linux workers....
cutechess-cli depends on the Qt framework and hence is a very heavy dependency of the worker. c-chess-cli is much more lightweight.
Leela evaluates KRKB as +6 even though it is usually theoretically draw. As a result she throws away won endgames by doing bad conversions. It is interesting to speculate why...
This reduces clutter and avoids duplicate definitions. Also: The run schema has been tightened. Various tweaks in other schemas. Requires upgrading the vtjson package.
This reverts commit 54048a24bfed7795a29bfc06db2992f2e30895cf. See https://tests.stockfishchess.org/actions?action=failed_task&user=Wencey&text=%22too+slow%22 EDIT: The above query currently (March 26, 2024) generates 149 pages!