"Mate the Royal Couple"

[Uri Blass]

Note that chest shows horrible order of moves in the following position
It search all the pawn moves and even Nh3 and Na3 before searching Nc3
that is clearly more logical.

Note that stupid checks like Qf6+ Qf7+ are searched earlier than Nc3.

I guess that an improvement in the order of moves can help chest to be faster in long mate problems and I guess that today chest may need some weeks to solve the mate in 11.

Uri

Sorting the moves of the attacker may help, or may not.
In most cases there is no solution, all moves must be searched
anyway, and sorting did not help this, but did cost some time.

This changes, when solutions are not so rare, so that a significant fraction of the attacker subjobs can be solved.
But how should Chest know that before?

Also, what is a "good" attacker move? Here we would want something very similar to the "eval" of playing programs, right?
I did not not want to go into that direction.

Clearly, Chest could be improved, here, but I have no idea how often it would pay off.

Cheers, Heiner

Time of sorting can be a significant disadvantage when the remaining depth is small but if you sort better only when you search at least for mate in 5 then sorting does not cost a significant time.

I think that score for sorting the move can be based on

1)normal chess evaluation(including SEE and mobility)
2)controlling squares near the king
3)history information if you already found this move in the past to be good in mating the opponent.

Uri