Dirt wrote:elcabesa wrote:I have another question about the root node search.
State of the art is filtering out bad moves in root and let the search choose the move it like the most?
It's all a matter of what you prefer. Now, I like the idea that in KBNR vs K that you should quickly sacrifice the rook to reach a winning KBN vs K position, but many people think that's stupid. It almost certainly won't lead to the shortest checkmate.
To explain this a bit more:
If the engine at the root simply picks the move that leads to the smallest DTZ value, then that will likely result in an attempt to sac a piece as quickly as possible.
This is "DTZ-optimal" play from both sides from a KRBNvK position with highest DTZ:
[pgn][FEN "8/8/8/8/8/B7/N2k4/RK6 b - - 0 1"]
1... Ke3 2. Nc3 Kf3 3. Bc1 Kg4 4. Bf4 Kf3 5. Ra2 Kg4 6. Rg2+ Kh4 7. Nd1 Kh3
8. Bg5 Kxg2[/pgn]
Against a "normal" opponent playing as black, the bishop would have gone off at move 4. Then white would probably have tried to sac its knight (or at least play so carelessly with it that black could win it - because DTZ optimal means that white assumes the opponent tries to avoid a zeroing move into a winning position).
To avoid such things, it is better to let the engine pick a move and use the DTZ information only to make sure that move makes sufficient progress (i.e. avoids a draw by the 50-move rule).
