list of material combinations requiring specialized eval ?

[AlvaroBegue]

Maybe I'll add syzygy to Demolito one day. But it's a lot pf woro, for little reward. SF implementation is unusable, because heavily intertwined with SF, and heavy on C++. Fathom is a mess, and I dont want to add illegible and messy code to my engine, because I will have to maintain this mess… So I'm not in a hurry to look at TB.

"Fathom is a mess"?? Fathom has a very sensible interface and it's quite easy to use. If it's a mess inside or not is kind of irrelevant if it's good enough that it works every time. I didn't even look inside the Fathom code, other than the header file.

[Dann Corbit]

Gull has a beautiful material imbalance table computed from first principles.

[hgm]

Syzygy doesn't tell you that KBPPPPKBPP with unlike Bishops is more drawish than KNPPPKBPP. Or that KBBNPKBNN has virtually no winning chances. So your advice is kind of suspect, for something presented as a 'full solution'.

[lucasart]

Syzygy doesn't tell you that KBPPPPKBPP with unlike Bishops is more drawish than KNPPPKBPP. Or that KBBNPKBNN has virtually no winning chances. So your advice is kind of suspect, for something presented as a 'full solution'.

Those are taken care of by scaling rules. Opposite color bishop is one. And winning soxe has only 1 pawn with material difference less than a rook, is another one I already mentioned.

[hgm]

OK, so scaling is important. But the rule you give seems to be off very much, both in terms of magnitude and accuracy. The "less than a Rook" rule should be applied after the projected sac of piece for the last Pawn. That means that in terms of the current material balance you don't have to be much ahead to make that plan a failure. E.g. with KRPKB (or KRNPKBN to stay outside EGTs) you cannot afford to sac B or N for the Pawn (which you eventually must), as KRNKN or KRNKB (or KRK) are very won. So there is no risk to the Pawn. Even KBBNPKRR (roughly equal) holds little extra danger compared to KBBNPPKRRP, because here the defending side has nothing cheaper to sac then a Rook. And KBBNKR is a win.

OTOH, when the Pawn is in jeopardy, winning chances are very slim indeed, and a multiplier of 0.75 still hugely overestimates the score. KBBPKBN has really bleak prospects, even KBBPPKNNPP could be preferable. To catch that in a score the extra passer would have to be reduced to below the value of the Bishop pair. A factor 0.75 won't do that.

[emadsen]

Could you provide me with a list of all material combinations that require specialized evaluation functions , and the general idea/method to correctly evaluate them ?

Are you incapable of answering any of your questions? Have you not done any homework and have nothing to share to give context to your questions? In other words, what have you tried, what were the results, what have you ruled out as ineffective, what do you suspect may be promising?

If you're unwilling to make an attempt to answer these questions, why are you writing a chess engine?

Do it yourself. Find ideas about material imbalances in chess books or, I don't know, the Internet, then enjoy the mental challenge of translating abstract ideas into concrete code, testing the code, refining the code, and then finally, watching the code play a beautiful game of chess.

Seeing all your questions on this programming forum, usually of the form, "Give me the answer," I'm really struggling to understand what you're getting out of this Q&A exercise? Wouldn't it be more direct to simply read open source code rather than asking TalkChess members to spoon feed it to you?

Sorry to be critical. Let me be positive and offer what I enjoy about my chess programming hobby: I love the intellectual challenge of translating ideas into working code. The difficulty of focusing my mind to think rigorously and with discipline is rewarded with a sense of accomplishment and a feeling of satisfaction as I watch my chess engine progress up the ELO rankings.

What do you enjoy about the hobby? What are you looking to get out of it? Ask yourself this: Are all these "give me the answer" questions the most effective way to pursue your goals?

I suspect you are young and have not yet learned this lesson. If so, now's the time to get cracking.

Cheers,
Erik

[Ras]

Reading some elementary endgame chess book would not only answer your question, but also make yourself a better chess player.