- Admin wrote:
- A material search only is excellent for a/b, it typically comes 2 plies deeper but the moves are horrible.
Target is not beating other engines.
I also would like to see no pruning or extensions.
The target is to measure how many plies do you need to search to win material.
Example for not easy problem for humans (and I do not mean finding the right move but how many plies you need to search).
Certainly engines can solve and give answer how many plies to win material if they are designed to do it(black to move and win):
chess positionI think that black need to calculate 14 plies brute force to prove winning material when detecting that a position is mate is considered as one ply but I am not sure if it is the longest line to calculate to find a forced win of material.
1...Rc1 2.Qb3 Qe1 3.f4 Be4 4.g4 h4 5.Rf7+ Kg8 6.Rf6 Qh1+ 7.Bxh1 Rxh1 mate
Practically when solving this problem I decided that black has probably a decisive attack after 1...Rc1 2.Qb3 Qe1 and did not analyze but I think that exercises that ask humans to analyze and explain how many plies they need to see in order to win material and what are the lines that they see to prove it may be good exercises.