what is the best materialistic engine and what is the rating of it?

[Dann Corbit]

I think it is hard to define a materialistic engine precisely.
A pure wood count is very simple. But what about a bishop pair? Is that part of the wood count or not?
Edge pawns are worth more in the endgame and less in the opening. Shall the evaluation account for that?
Will the engine ignore king safety and center control and battery and other simple ideas like that?

I think, in the end, that the best purely materialistic engine (basic would count) is not going to be best at anything.
Ass you add soophistication, is it still just materialistic or are we adding knowledge?

On the other hand, if you are looking for the best ultra simple evaluation, I think Olithink is a good candidate.
It's basically wood + mobility with a little pawn structure thrown in.

[Uri Blass]

I think it is hard to define a materialistic engine precisely.
A pure wood count is very simple. But what about a bishop pair? Is that part of the wood count or not?
Edge pawns are worth more in the endgame and less in the opening. Shall the evaluation account for that?
Will the engine ignore king safety and center control and battery and other simple ideas like that?

I think, in the end, that the best purely materialistic engine (basic would count) is not going to be best at anything.
Ass you add soophistication, is it still just materialistic or are we adding knowledge?

On the other hand, if you are looking for the best ultra simple evaluation, I think Olithink is a good candidate.
It's basically wood + mobility with a little pawn structure thrown in.

I think it may be interesting to define the positional sacrifices in the opening theory.

Here is one example:

[d]r1b1k2r/pp2nppp/2n1p3/2ppP3/q2P3P/P1P5/2PB1PP1/1R1QKBNR w Kkq - 3 10

in this position white has only 2 moves that do not lose material based on texel materialistic.
After the theoretic 10.h5 black has exactly one move that wins a pawn that is probably also a losing move..

I wonder if strong humans can find the best move to win material based on materialistic chess engines.
[d]r1b1k2r/pp2nppp/2n1p3/2ppP3/q2P3P/P1P5/2PB1PP1/1R1QKBNR w Kkq - 3 10

[petero2]

I think that one problem is with the value of the pieces 1,3,3,5 for pawn knight bishop and rook.

The value of 2 minor pieces is closer to 7 pawn relative to 6 pawns and I can consider losing bishop and knight for rook and pawn as a tactical mistake.

Yes, the standard piece values are quite crude. I once used 1,4,4,6,12 in texel, but I don't know if that would be better. It would also be possible to use fractional values, but then there will be less room between different material configurations for the positional part of the evaluation. It would also be more difficult for a human to see what the material part of the score is.

My guess is that the new materialistic texel should be relatively weaker at longer time control because of weaknesses in evaluation of material but it is only a guess and I did not do testing.

I think so too. I tried playing at 10 times slower time control (60 seconds per game + 0.6 seconds per move) and the elo difference compared to crafty dropped from roughly 200 to 115, but the elo difference of "non-crippled" texel was roughly the same (> 700).

[Uri Blass]

I think that one problem of it is tactics because null move pruning does not work well and it prune moves that have a threat but not threat based on mainly material evaluation. maybe verifying that there is no threat based on normal evaluation before null move pruning can help to make it better at longer time control.

Here is an example from my analysis:
[d]2kr1b1r/pp2ppp1/2n4p/3p4/3P1B1P/1RqB1b2/P1P2PP1/3Q1K1R w - - 0 14 bm Qb1

[Uri Blass]

It seems texel materialistic without null move pruning also does not see it so I guess it is related to different pruning based on evaluation and not null move pruning.