Scratch Chess Engine - Game of Kings

HasiLover

Running Round 2 of my current Tournament now: https://lichess.org/broadcast/supersonic-scratch-2025/round-2/lf4rnuHF/IjqOFdwe

ArnoHu

coooolboy55 wrote:
What is the effective branching factor of your engines? I want to know what I am aiming for when I get to making my chess bot (currently working on the chess program itself).

Quick stats for GoK Classic, white, starting board, full search nodes (not counting q-nodes):

Depth	 Nodes	   BF
 1	    20	
 2	    64 	 3.20
 3	   231 	 3.61
 4	  1129 	 4.89
 5	  2089	 1.85
 6	  7006	 3.35
 7	 13294	 1.90
 8	 29367	 2.21
 9	 74750	 2.55
10	143771	 1.92

Side-notes:

Nodes are counted based on make_move() invocations, not evaluations. This includes make_move() invoked by SEE.
GoK uses check extensions, this might increase numbers, but probably not that much on the starting board.

Last edited by ArnoHu (Nov. 2, 2025 19:41:45)

ArnoHu

ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
GoK Classic vs. Wolverine 2, 94% vs. 88%:
https://lichess.org/study/owu9JlSd/FJaNkCuO

GoK Classic vs. Wolverine 2, 87% vs. 85%:
https://lichess.org/study/owu9JlSd/MlgJ9fJL

GoK NNUE vs. Black Crow, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/BxCtMfBt

GoK Classic vs. Wolverine 2, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/WRM9Acxf

GoK NNUE vs. Black Crow, 92% vs. 88%:
https://lichess.org/study/owu9JlSd/4PyIlAsZ

This is certainly a pinnacle point for Scratch HCE engines (at 5sec think time):

GoK Classic vs. White Dove, 99% vs. 99%:
https://lichess.org/study/owu9JlSd/3I15lM22

GoK NNUE vs. Black Crow, 94% vs. 89%:
https://lichess.org/study/owu9JlSd/zNmeWci6

GoK Classic vs. Wolverine 2, 90% vs. 86% (alternate endgame):
https://lichess.org/study/owu9JlSd/QJYZ4Bjw

GoK Classic vs. Wolverine 2, 93% vs. 90% (king shelter eval bugfix):
https://lichess.org/study/owu9JlSd/SX0LS842

Last edited by ArnoHu (Nov. 3, 2025 17:52:47)

waabooboo

Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it

Destructor_chess

waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it

What do you think about the draw between tk and w2?

Destructor_chess

So, I was working on TK5, I made a lot of progress, but, onviously, the sb3 file broke. Arnohu proposed me to try to fix my .sb3 file. Thanks to him. IF he doesn’t get it to work, I will stop the development of TK5 and resume TK4's one. Obviously (yeah, I learned this word less than a werk ago, thanks to fastskript and internet44

), I think that he wont be able - what am I saying? Obviously, sb3 format is sh** so he wont be able to repair it…

waabooboo

Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?

ArnoHu

Destructor_chess wrote:
So, I was working on TK5, I made a lot of progress, but, onviously, the sb3 file broke. Arnohu proposed me to try to fix my .sb3 file. Thanks to him. IF he doesn’t get it to work, I will stop the development of TK5 and resume TK4's one. Obviously (yeah, I learned this word less than a werk ago, thanks to fastskript and internet44 ), I think that he wont be able - what am I saying? Obviously, sb3 format is … so he wont be able to repair it…

Fixed

Destructor_chess

waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?

I sent it some times ago;

Destructor_chess wrote:
Wolverine 2 vs. TK, 86% vs 88%, 1/2 - 1/2
https://lichess.org/study/5IpBFQUb/y8HxGgyu
I hate this bug

TK vs. SB3, 96% vs. 91%, 1-0
https://lichess.org/study/5IpBFQUb/HWNgERmY

Amarillo vs. TK, 87% vs. 95%, 0-1
https://lichess.org/study/5IpBFQUb/IK0YwWHz

btw, it was on my laptop, so please don't say “blunder prevented in 3 sec on my system…” because on my fast system too for TK…

Last edited by Destructor_chess (Nov. 3, 2025 21:33:13)

Destructor_chess

ArnoHu wrote:
Destructor_chess wrote:
So, I was working on TK5, I made a lot of progress, but, onviously, the sb3 file broke. Arnohu proposed me to try to fix my .sb3 file. Thanks to him. IF he doesn’t get it to work, I will stop the development of TK5 and resume TK4's one. Obviously (yeah, I learned this word less than a werk ago, thanks to fastskript and internet44 ), I think that he wont be able - what am I saying? Obviously, sb3 format is … so he wont be able to repair it…

Fixed

Thanks you really much. I finally found and eradicated the bug I had

(that's why I wanted to texport it in txt, you know, sometimes AI can really helps). So I resumed TK5 development!

waabooboo

Destructor_chess wrote:
waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?
I sent it some times ago;

Destructor_chess wrote:
Wolverine 2 vs. TK, 86% vs 88%, 1/2 - 1/2
https://lichess.org/study/5IpBFQUb/y8HxGgyu
I hate this bug

TK vs. SB3, 96% vs. 91%, 1-0
https://lichess.org/study/5IpBFQUb/HWNgERmY

Amarillo vs. TK, 87% vs. 95%, 0-1
https://lichess.org/study/5IpBFQUb/IK0YwWHz

btw, it was on my laptop, so please don't say “blunder prevented in 3 sec on my system…” because on my fast system too for TK…

Blunder was prevented in 3 sec on my system.

ArnoHu

ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
GoK Classic vs. Wolverine 2, 94% vs. 88%:
https://lichess.org/study/owu9JlSd/FJaNkCuO

GoK Classic vs. Wolverine 2, 87% vs. 85%:
https://lichess.org/study/owu9JlSd/MlgJ9fJL

GoK NNUE vs. Black Crow, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/BxCtMfBt

GoK Classic vs. Wolverine 2, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/WRM9Acxf

GoK NNUE vs. Black Crow, 92% vs. 88%:
https://lichess.org/study/owu9JlSd/4PyIlAsZ

This is certainly a pinnacle point for Scratch HCE engines (at 5sec think time):

GoK Classic vs. White Dove, 99% vs. 99%:
https://lichess.org/study/owu9JlSd/3I15lM22

GoK NNUE vs. Black Crow, 94% vs. 89%:
https://lichess.org/study/owu9JlSd/zNmeWci6

GoK Classic vs. Wolverine 2, 90% vs. 86% (alternate endgame):
https://lichess.org/study/owu9JlSd/QJYZ4Bjw

GoK Classic vs. Wolverine 2, 93% vs. 90% (king shelter eval bugfix):
https://lichess.org/study/owu9JlSd/SX0LS842

Tough GoK Classic vs. Wolverine 2 match ends in a draw, 90% vs. 91%:
https://lichess.org/study/owu9JlSd/JhHWzFzz

GoK NNUE vs. Black Crow, 94% vs. 93%:
https://lichess.org/study/owu9JlSd/WRi8i6Hv

Last edited by ArnoHu (Nov. 4, 2025 12:02:06)

internet44

waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?
I sent it some times ago;

Destructor_chess wrote:
Wolverine 2 vs. TK, 86% vs 88%, 1/2 - 1/2
https://lichess.org/study/5IpBFQUb/y8HxGgyu
I hate this bug

TK vs. SB3, 96% vs. 91%, 1-0
https://lichess.org/study/5IpBFQUb/HWNgERmY

Amarillo vs. TK, 87% vs. 95%, 0-1
https://lichess.org/study/5IpBFQUb/IK0YwWHz

btw, it was on my laptop, so please don't say “blunder prevented in 3 sec on my system…” because on my fast system too for TK…

Blunder was prevented in 3 sec on my system.

Destructor_chess

waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?
I sent it some times ago;

Destructor_chess wrote:
Wolverine 2 vs. TK, 86% vs 88%, 1/2 - 1/2
https://lichess.org/study/5IpBFQUb/y8HxGgyu
I hate this bug

TK vs. SB3, 96% vs. 91%, 1-0
https://lichess.org/study/5IpBFQUb/HWNgERmY

Amarillo vs. TK, 87% vs. 95%, 0-1
https://lichess.org/study/5IpBFQUb/IK0YwWHz

btw, it was on my laptop, so please don't say “blunder prevented in 3 sec on my system…” because on my fast system too for TK…

Blunder was prevented in 3 sec on my system.

It’s a joke or it’s true? Or both?

Last edited by Destructor_chess (Nov. 4, 2025 10:16:08)

waabooboo

Destructor_chess wrote:
waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Destructor_chess wrote:
waabooboo wrote:
Wolverine update:

- FEN castling bug found and fixed
- better passed pawn evaluation
- specialized endgame evaluation is back – I found a cheap way to do it
What do you think about the draw between tk and w2?

I don't think anything, haven't seen it. Game link?
I sent it some times ago;

Destructor_chess wrote:
Wolverine 2 vs. TK, 86% vs 88%, 1/2 - 1/2
https://lichess.org/study/5IpBFQUb/y8HxGgyu
I hate this bug

TK vs. SB3, 96% vs. 91%, 1-0
https://lichess.org/study/5IpBFQUb/HWNgERmY

Amarillo vs. TK, 87% vs. 95%, 0-1
https://lichess.org/study/5IpBFQUb/IK0YwWHz

btw, it was on my laptop, so please don't say “blunder prevented in 3 sec on my system…” because on my fast system too for TK…

Blunder was prevented in 3 sec on my system.
It’s a joke or it’s true? Or both?

No, I was just saying it because you told me not to say it. I didn't check.

Also, while the analysis is useful, the worst moves aren't necessarily the ones Lichess calls “blunders”.

Last edited by waabooboo (Nov. 4, 2025 16:57:07)

ArnoHu

ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
GoK Classic vs. Wolverine 2, 94% vs. 88%:
https://lichess.org/study/owu9JlSd/FJaNkCuO

GoK Classic vs. Wolverine 2, 87% vs. 85%:
https://lichess.org/study/owu9JlSd/MlgJ9fJL

GoK NNUE vs. Black Crow, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/BxCtMfBt

GoK Classic vs. Wolverine 2, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/WRM9Acxf

GoK NNUE vs. Black Crow, 92% vs. 88%:
https://lichess.org/study/owu9JlSd/4PyIlAsZ

This is certainly a pinnacle point for Scratch HCE engines (at 5sec think time):

GoK Classic vs. White Dove, 99% vs. 99%:
https://lichess.org/study/owu9JlSd/3I15lM22

GoK NNUE vs. Black Crow, 94% vs. 89%:
https://lichess.org/study/owu9JlSd/zNmeWci6

GoK Classic vs. Wolverine 2, 90% vs. 86% (alternate endgame):
https://lichess.org/study/owu9JlSd/QJYZ4Bjw

GoK Classic vs. Wolverine 2, 93% vs. 90% (king shelter eval bugfix):
https://lichess.org/study/owu9JlSd/SX0LS842

Tough GoK Classic vs. Wolverine 2 match ends in a draw, 90% vs. 91%:
https://lichess.org/study/owu9JlSd/JhHWzFzz

GoK NNUE vs. Black Crow, 94% vs. 93%:
https://lichess.org/study/owu9JlSd/WRi8i6Hv

GoK Classic vs. Wolverine 2, 89% vs. 90%:
https://lichess.org/study/owu9JlSd/CcWC9aoV

GoK Classic vs. Wolverine 2, 90% vs. 91%:
https://lichess.org/study/auQC6T2K/UFTkmzkA

GoK NNUE vs. Black Crow, 96% vs. 92%:
https://lichess.org/study/auQC6T2K/xcIaf5WD

GoK Classic vs. Wolverine 2, 94% vs. 91%:
https://lichess.org/study/auQC6T2K/9kRxE1y8

Last edited by ArnoHu (Nov. 4, 2025 21:59:18)

ArnoHu

HasiLover wrote:
Running Round 2 of my current Tournament now: https://lichess.org/broadcast/supersonic-scratch-2025/round-2/lf4rnuHF/IjqOFdwe

Cool, when is round 3?

Destructor_chess

ArnoHu wrote:
HasiLover wrote:
Running Round 2 of my current Tournament now: https://lichess.org/broadcast/supersonic-scratch-2025/round-2/lf4rnuHF/IjqOFdwe

Cool, when is round 3?

it seems to be alr setupped: https://lichess.org/broadcast/supersonic-scratch-2025/round-3/QvJbxss9#boards

R.I.P. goK classic and delta

Last edited by Destructor_chess (Nov. 4, 2025 21:37:35)

waabooboo

ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
ArnoHu wrote:
GoK Classic vs. Wolverine 2, 94% vs. 88%:
https://lichess.org/study/owu9JlSd/FJaNkCuO

GoK Classic vs. Wolverine 2, 87% vs. 85%:
https://lichess.org/study/owu9JlSd/MlgJ9fJL

GoK NNUE vs. Black Crow, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/BxCtMfBt

GoK Classic vs. Wolverine 2, 93% vs. 86%:
https://lichess.org/study/owu9JlSd/WRM9Acxf

GoK NNUE vs. Black Crow, 92% vs. 88%:
https://lichess.org/study/owu9JlSd/4PyIlAsZ

This is certainly a pinnacle point for Scratch HCE engines (at 5sec think time):

GoK Classic vs. White Dove, 99% vs. 99%:
https://lichess.org/study/owu9JlSd/3I15lM22

GoK NNUE vs. Black Crow, 94% vs. 89%:
https://lichess.org/study/owu9JlSd/zNmeWci6

GoK Classic vs. Wolverine 2, 90% vs. 86% (alternate endgame):
https://lichess.org/study/owu9JlSd/QJYZ4Bjw

GoK Classic vs. Wolverine 2, 93% vs. 90% (king shelter eval bugfix):
https://lichess.org/study/owu9JlSd/SX0LS842

Tough GoK Classic vs. Wolverine 2 match ends in a draw, 90% vs. 91%:
https://lichess.org/study/owu9JlSd/JhHWzFzz

GoK NNUE vs. Black Crow, 94% vs. 93%:
https://lichess.org/study/owu9JlSd/WRi8i6Hv

GoK Classic vs. Wolverine 2, 89% vs. 90%:
https://lichess.org/study/owu9JlSd/CcWC9aoV

GoK Classic vs. Wolverine 2, 90% vs. 91%:
https://lichess.org/study/auQC6T2K/UFTkmzkA

GoK NNUE vs. Black Crow, 96% vs. 92%:
https://lichess.org/study/auQC6T2K/xcIaf5WD

GoK Classic vs. Wolverine 2, 94% vs. 91%:
https://lichess.org/study/auQC6T2K/9kRxE1y8

The Lichess accuracy is a bit mysterious to me. For both draws, only GoK had winning chances, and yet Wolverine was awarded the extra percentage point.

3rk3/3q1p2/1p1p4/p1pP3Q/8/2P1P1Pp/P1P5/2K2R2 w - - 0 28

From the second game. Wolverine finds the forced win for white right around the 5 second mark (evals shoot up from around +0.5 to +4 at depth 13). But until Wolverine sees a way to win by force, he is perfectly happy to leave his king stuck in the center. A healthy dose of fear may not be a bad thing…

HasiLover

ArnoHu wrote:
HasiLover wrote:
Running Round 2 of my current Tournament now: https://lichess.org/broadcast/supersonic-scratch-2025/round-2/lf4rnuHF/IjqOFdwe

Cool, when is round 3?

Friday, if everything goes well

Discuss Scratch