RE: [Bug-gnubg] Is eval-4ply and 0-ply-rollout-truncated-4ply the same?

[Ian Shaw]

From:  Christian Anthon
Sent: 28 November 2007 10:08

> Anybody know what is actually 
> going on or do I have to go look at the code? :)

Christian,

There's a reasonable explanation in Thomas Hauk's paper,
http://www.cs.ualberta.ca/~mburo/ps/STAR-B.pdf. It was written before
Joseph introduced the mini-net for pruning, when forward pruning simply
depended on the filters. See Sections 4 and 5.2.

-- Ian