#ttCapture in futility pruning vs LMR, intentional asymmetry?

8 messages · Page 1 of 1 (latest)

tidal saddle
#

Hey, I've been reading through search.cpp and noticed something that seemed worth raising.

ttCapture appears in two places with what looks like opposite intent:

Futility pruning (line 916): ttCapture allows pruning to trigger:

!is_loss(beta))```

LMR (line 1241):  ttCapture increases reduction:
```if (ttCapture)
    r += 1054;```

My read: both are saying "if the TT move is a capture, be more aggressive about reducing/pruning other moves", so maybe it's consistent and intentional. But the futility condition treats ttCapture the same as having no TT move at all, which seems like it might be too permissive.

Is there a known reason for this, or is it worth running a test that removes ttCapture from the futility condition to see if being more conservative there helps?
frank grove
#

Is this AI writing?

#

But sure, worth running a test

#

Do you have a fishtest account?

tidal saddle
frank grove
#

One of us will approve it soon :)

solar snow