I read the search algortihm, no hash, nothing special but pure alpha beta, so i
wonderred about speed. But the evaluation functions part needs more
investigation.

In fact, I tried to select only a few moves in MLA, and i was proven wrong
against MLA without the pruning.

So i'm looking forward to see a program play good by picking lony a few moves by
looking 2 moves ahead at node level at any level. This would really make me
question my tests.

David.

Quoting peer_sommerlund <peer.d.sommerlund@...>:

> My French is not really good, but if I understand the discussion
> between David and Stéphane correctly David argues that Tax Abalone must
> be slow. Stéphane argues that this is not the case, due to the limited
> branching that Stéphane added to the alpha-beta algorithm: For non-leaf
> nodes he does a 1 or 2 ply search (depending on distance to leaves),
> and keeps only the 6 best moves for the full depth search.
>
> I am not yet done with merging the source code, but a quick comparison
> of version Tax abalone 1.5.3 (which is the code base I have) and the
> current version of Nacre, shows that Tax abalone is pretty fast: 6 ply
> in 1 sec, whereas Nacre does only 3-4 ply in 1 sec. I played only one
> blitz game of 1 sec/move, and Nacre won this 6-0 despite the large
> difference in search depth. Thus, speed is not always enough.
>
> When I have merged version 2.0 with my version 1.5.3, I will return
> with more test results.
>
> Peer
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>