zamar wrote:If you really want to play with extensions, you should start with the most interesting ones.
{ "Check Extension (PV nodes)", "2", "2", SPIN, 0, 2, {""} },
{ "Check Extension (non-PV nodes)", "1", "1", SPIN, 0, 2, {""} },
If you disable check extensions, program will lose quite many elo points.
Probably, but whether extending by half a ply or a full ply is better is an interesting question. As you can see above, Glaurung extends checks by only half a ply, except in PV nodes (which are very few).
{ "Single Reply Extension (PV nodes)", "2", "2", SPIN, 0, 2, {""} },
{ "Single Reply Extension (non-PV nodes)", "2", "2", SPIN, 0, 2, {""} },
This is quite obvious (disabling it makes no point), but this one can also be quite easy to measure.
I think (based on educated guesses, not on tests, as usual) that checks and single replies shouldn't
both be extended by a full ply at non-PV nodes. In other words, if check extensions are increased to a full ply, single replies should be decreased to half a ply.
{ "Mate Threat Extension (PV nodes)", "0", "0", SPIN, 0, 2, {""} },
{ "Mate Threat Extension (non-PV nodes)", "0", "0", SPIN, 0, 2, {""} },
This can be used to tune engine to more attractive style of play, but my little tests suggest enabling it makes program weaker in general.
I don't even think it has any noticable effect on the style of play. It solves some test positions more quickly, that's all.
{ "Pawn Push to 7th Extension (PV nodes)", "1", "1", SPIN, 0, 2, {""} },
{ "Pawn Push to 7th Extension (non-PV nodes)", "1", "1", SPIN, 0, 2, {""} },
This happens so rarely (or game is already decided when it happens) that measuring value of it is almost impossible. Test this only if you can run say >100k games.
I think this extension is far more important than you believe. Pawn pushes to the 7th rank are not
that unusual, and they often occur at a very critical moment of the game. The outcome of the game is very likely to depend on whether the pawn can promote in a few moves, and extending the move makes a lot of sense.
{ "Passed Pawn Extension (PV nodes)", "1", "1", SPIN, 0, 2, {""} },
{ "Passed Pawn Extension (non-PV nodes)", "0", "0", SPIN, 0, 2, {""} },
Hard to measure, you need a lot of games (30k?)
I think it helps, but as always, I am just guessing.
I agree with others that there is not much (if at all) to gain from extensions, instead there are other search parameters (like LMR & futility stuff) in glaurung which can (and should) be tuned.
I don't think there are any huge improvements to be found by tuning these parameters either. I guess that even after fine-tuning all the many search and evaluation parameters in Glaurung with millions of games, you wouldn't be able to improve the strength by more than 30-40 Elo points. For bigger improvements, you'd have to modify the source code.
Tord
