cj5716
cj5716
Unfortunately, there are some unintended effects, as noted by @locutus2, whereby the killer and countermove are updated despite being given a negative bonus. We will attempt to fix this, hopefully...
> positions like this could be compiled with the "known" answer, i.e., it is a draw and the data added to the training data for the nets I highly doubt...
By the way, you should remember to replace the tabs with spaces (or the indentation will be off, and formatting will be inconsistent).
oh ok, closing this patch then.
If we're working on cleanups I think there is one available (there are some places in the code where psqtOnly is spelt as psqtOnl) remember to include "No functional change"...
the nw stuff doesn't gain much I believe, so it might be of use removing it and doing a retest for bytes sake (see https://chess.swehosting.se/test/6998/)
I think a non-reg STC would be good, because this has shown to be quite the slowdown for many engines Edit: sorry, I misread and missed the operator overloading Great...
There is no guarantee where a good capture will beat the SEE threshold (and vice versa), though, because the way we determine good captures in movepicker is not a fixed...
ah, I see what you mean. that could be a possible speedup
I think a non-reg here is enough imo. Especially when it is to do with obeying `go nodes` more strictly and accurately