>
> I cannot answer for the current version of Mogo but I can for the one
> 1.5 years ago. Maybe it still holds.
> We had a transposition table as it was designed like that from the
> beginning. However the prior value of the node, was initialized when
> the node was created and indeed was depending of the previous move
> played in the current sequence.
> In practice, I think it does not change much.
>

We use the two previous moves in the current version (but it's not a huge
improvement, just a small improvement).

I guess that the following details are not very important, however
for completeness, when two nodes have a same son S, the value of this son
depends on:
- the statistics of S (this value is the same for the two fathers)
- the two previous moves (this depends on which father is considered)
- the RAVE values (this also depends on which father is considered)
- the exploration term (which is some heuristic divided by log(nb Simus)
   depends on the number of simulations of S but also on the heuristic
   value of the move and therefore depends on which father is considered.

Best regards,
Olivier
_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/

Reply via email to