Chess.com's Komodo: 1 second (for the last one).
Mates that are difficult for engines
Sort:
Apr 22, 2023
0
#3
Arisktotle wrote:
Chess.com's Komodo: 1 second (for the last one).
Oops, you're right, that was the wrong mate in 6 that I posted 
I accidentally copied and pasted the wrong one.
Here is the correct mate in 6:
Ah, that's better. Or worse, for me 
Apr 22, 2023
0
#6
DesperateKingWalk wrote:
ungimmicked wrote:
White to play and mate in 11
Mate is 11 was solved instantly for Dragon.
Wow! impressive. You must have a monster of a computer 
Apr 22, 2023
0
#8
DesperateKingWalk wrote:
ungimmicked wrote:
@DesperateKingWalk what are your computer specs?
AMD 2950x Threadripper, 64GB Ram, 2 TB SSD NVMe, RTX 2080 GPU
And built and designed to test chess engines for my YouTube Channel.
Jezuz!! what a monster. No wonder your engine found that mate instantly
Apr 22, 2023
0
#9
drdos7 wrote:
DesperateKingWalk wrote:
The mate in 6 was easy for Stockfish and Threadripper.
yep, easy for me and you because we are using the local version of Stockfish and of course the latest development versions, but not on the engines they use on chess.com.
And what about the mate in 9?
Apr 22, 2023
0
#11
DesperateKingWalk wrote:
drdos7 wrote:
DesperateKingWalk wrote:
The mate in 6 was easy for Stockfish and Threadripper.
What was the time on the solution? My comp which is slower than yours (about 75% of your KNS on the latest Stockfish development version) doesn't solve it very quickly at all, I suspect it may be because I'm not using tablebases as I see you are.
And what about the mate in 9?
It found a mate in 22 for the mate in 9 and did not solve it correctly.
It solved the mate in 6 in about 5 to 7 seconds if I remember. The photo has 9 seconds but it was already solved and it took a few seconds to get a screen shot.
And the TB are most likely your issue. I am using 7 man TB.
It was the TBs, I enabled them and got 10 seconds. by the way what is the link to your YouTube channel so I can check it out.
Apr 22, 2023
0
#12
Here is a mate in 5 once again not too easy for engines, unless you are using a mate-solver or something like Crystal 5 KWK:
Apr 22, 2023
0
#13
DesperateKingWalk wrote:
Pretty easy for Threadripper.
I can see I'm going to have to step up my game for Threadripper.
Mate in 21:
I had to edit this as I had it incorrectly down as a mate in 31.
Apr 22, 2023
0
#14
How about a mate in 58:
This one MIGHT be able to be solved in fewer moves, but I did get it down to 58
Apr 23, 2023
0
#17
DesperateKingWalk wrote:
drdos7 wrote:
Here's a nice mate in 12 that trips up a lot of engines:
Simple.
Were you able to get that crazy mate in 20?
I'll post some more when I get home from work.
Apr 23, 2023
0
#18
Here is a puzzle known as "Plaskett's Puzzle" that is very well known for computer chess, it is actually a chess endgame study created by the Dutch endgame composer Gijs van Breukelen around 1970, it was origanally supposed to be a "white to play and win study", but with the advent of computer engines I found out that there is actually a forced mate in this position. If you would like to read more about this study here is a nice link: https://en.wikipedia.org/wiki/Plaskett%27s_Puzzle
White mates in 32 moves OR Less!
I should also mention that a version of Rybka from around 2007 was always able to solve this study and show a winning score, but not find the mate in 32 or less. If you would like to download that version of Rybka (2.3.2a) you still can here: http://www.rybkachess.com/index.php?auswahl=Demo+version
Also I should mention that if you do download the Rybka 2.3.2a UCI engine that you CANNOT use a hash table size larger than 2048MB or 2GB, if you do the engine will crash (remember it's from 2007 so 2GB hash was a lot back then).
Apr 23, 2023
0
#19
DesperateKingWalk wrote:
Stockfish see a mate! And finds the solution.
Ok, post the mate when you get it, ESPECIALLY if it is less than 32 moves!
White to move and mate in 9 moves