Hello all,

I'm working on an implementation of Rémi Coulom's Elo-based move
prediction algorithm[1], and I seem to be running into efficiency
issues. With my current implementation, I'm running at around 8 or 10
playouts per second. I can't imagine that the playouts would be so
accurate that I would be able to accurately evaluate my next move at
that speed. It seems that simply visiting each vacant point on the
board and checking whether that move would be a legal move for the
current player (without even computing which gammas are present at a
point) takes me down to about 60-75 playouts per second.

My question to the list is this: How has this problem been solved in
the past? I feel compelled to to use the Elo-system because it
generates a nice probability distribution across the board, but it
currently looks infeasible. One thought I've come up with is that I
might be abusing the cache (as there's no guarantee of order in the
vacant point collection I iterate over), but my understanding is that
the points tend to be in a stride-one order even if they aren't
certain to be.

Thank you all for your time, I would appreciate any help you can give
on this issue.

Seth Pellegrino
Student, Lewis & Clark College


[1] http://remi.coulom.free.fr/Amsterdam2007/
_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/

Reply via email to