The Unofficial World Championship Of Computer Programs 2024

[BertTuyt]

Just say what you want to say. A pity Damage did not win, but i knew this could happen (although not often) and need to correct it.
So your decision for the draw was the right one.

Bert

[Krzysztof Grzelak]

Just say what you want to say. A pity Damage did not win, but i knew this could happen (although not often) and need to correct it.
So your decision for the draw was the right one.

Bert

You need to correct this in the program code so that the program does not make such errors. I looked at the monitor and the rating was 42.50 or 42.70 and I thought the program won. But the program made its moves without any thought (and it had time to think) and the game ended in a draw.

[BertTuyt]

I fully agree with you.
Did you say that the moves were played by Damage without thinking, that is indeed strange.
I will definitely need correct this, as with one point more the match result would be slightly different. But the same is valid for Dragon.

Bert

[Krzysztof Grzelak]

I fully agree with you.
Did you say that the moves were played by Damage without thinking, that is indeed strange.
I will definitely need correct this, as with one point more the match result would be slightly different. But the same is valid for Dragon.

Bert

Just as you write.

[BertTuyt]

A more difficult endgame where I'm not certain if a draw opportunity was missed, was the game Flits - Kingsrow. See the diagram with Flits to move at move 45.



Flits (white) played 38-33 and the game was concluded soon as a win for Kingsrow.
A better defense seems to be 40-35 (9-14) 32-28 (21-27) 37-32 (27-31 or 14-20) with a possible draw.
I'm not sure this sequence is right and that this ends in a draw, Ed might have the final verdict here.

As Flits played without endgame databases (if I'm right) then it would be hard to imagine that the program would have found all critical moves.

Bert

[Krzysztof Grzelak]

A more difficult endgame where I'm not certain if a draw opportunity was missed, was the game Flits - Kingsrow. See the diagram with Flits to move at move 45.



Flits (white) played 38-33 and the game was concluded soon as a win for Kingsrow.
A better defense seems to be 40-35 (9-14) 32-28 (21-27) 37-32 (27-31 or 14-20) with a possible draw.
I'm not sure this sequence is right and that this ends in a draw, Ed might have the final verdict here.

As Flits played without endgame databases (if I'm right) then it would be hard to imagine that the program would have found all critical moves.

Bert

This is how he played without the endings database due to a known bug in the program. And the author of the program does not want to fix the error.

[Krzysztof Grzelak]

I would also like to mention Sjende Blyn who has never lost a game.

[BertTuyt]

Krzysztof, you are right, especially impressive as Sjende Blyn is the only top program (to my knowledge) which uses a HCE (Hand Crafted Evaluation) evaluation, whereas other top-programs use an evaluation based upon machine learning.
Not sure if Edeon Sport uses a pattern based evaluation, but if so, Sjende Blyn is the only program in the top 10, which does not use machine learning.

Bert

[Ed Gilbert]

Flits (white) played 38-33 and the game was concluded soon as a win for Kingsrow.
A better defense seems to be 40-35 (9-14) 32-28 (21-27) 37-32 (27-31 or 14-20) with a possible draw.
I'm not sure this sequence is right and that this ends in a draw, Ed might have the final verdict here.

Kingsrow agrees with your analysis. 38-33 is a db loss, 40-35 is a definite draw.