TalkChess.com

This version of Sting should be called rather Sting SF AI 1. Practically I start from a zero point and it needs my next solid work. This version (28) is rather focused on my "IQ" test examples, but now I will try to open it widely. The alpha-beta search with the "black hole" idea gives huge possibil...

At this time, the "Black Hole" idea in Sting is limited by material.
It is my conscious temporary decision and it means that some positions with an useless big material can be unsolved now.

download: https: https://www.mediafire.com/file/7ad6m6j191ngit5/sting-sf-28.zip/file I removed all my old ideas and now Sting is based only on my new "Black Hole" idea. It makes Sting (alpha-beta) an intelligent engine without "Neural Network" learning. The classic alpha-beta Searching, thanks to th...

53 - correction and 39 - new [Result "1"] [FEN "rkr5/p5p1/Pp4Kp/8/3B4/1B6/8/8 w - - 0 1"] 1. Be5 {win} [Result "2"] [FEN "7k/6rp/6pN/1r4P1/3B4/8/8/K7 b - - 0 1"] 1... Rxg5 [Result "3"] [FEN "1B1k4/3p4/1Kp5/PpP2p2/1PqP1P2/8/1P6/8 w - - 0 1"] 1. a6 {win} [Result "4"] [FEN "3N2r1/3Kp3/4Pk2/8/Bp6/6b1/2P...

Crystal3.1 is not bad on this set. Conditions : 10 min (1 core old i7@4 GHz) per position 79 found but I have to check manually a dozen of solutions to see if it really solves them. A good work Vincent! The task is: to solve all 100/100 positions (half hour per item for 4 000 000 nodes/sec) . It is...

"A new day with a new Sting is a good day." Hi lech, This problem is too old but when I set Sting's thread as 3 it starts analyzing with 4. Can you check this problem? Also, I really wonder how strong will be Sting if you add your path into last Stockfish's sources. I think that engines to control ...

download: https://www.mediafire.com/file/vlrfopa4 ... 7.zip/file

Sting 25 : http://talkchess.com/forum3/viewtopic.php?f=2&t=76091
"Black hole": http://talkchess.com/forum3/viewtopic.php?f=2&t=76122
"IQ" test : http://talkchess.com/forum3/viewtopic.p ... 11#p879242

I changed: 15, 71 are more difficult, in 4, 20, 39, 49 some obvious first moves are skipped. [Result "1"] [FEN "rkr5/p5p1/Pp4Kp/8/3B4/1B6/8/8 w - - 0 1"] 1. Be5 {win} [Result "2"] [FEN "7k/6rp/6pN/1r4P1/3B4/8/8/K7 b - - 0 1"] 1... Rxg5 [Result "3"] [FEN "1B1k4/3p4/1Kp5/PpP2p2/1PqP1P2/8/1P6/8 w - - 0...

The hashtable is not common for two (not one) engines and it can sometimes make a difference. It is possible to make an auto-play(ing) manually. Start a search(ing), stop it, skip one move (it is possible to skip more moves too) and start search(ing) again (don't clear the hash table!), stop it, ski...

The "IQ" test doesn't contain middle-game tactical positions and technical endgames. Here are positions dedicated to test so-called by me "quasi-intelligent" abilities of engines. Don't use such tricks like: rule-50 changing, null-move disabling, backward-forward, special parameters (versions) and ...