Discuss Scratch

#8441Sept. 16, 2025 21:31:45
S_P_A_R_T_Test
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

coooolboy55 wrote:

WhiteDoveOfficial wrote:

coooolboy55 wrote:

I gave my bot the capability of playing Connect 4 at high speeds. I switched from the scratch timer to the “days since 2000” block (which is more accurate). It's quite funny that my bot can beat/tie some Connect 4 bots at the speed of 0.1 seconds.

Nice!

(BTW, have you tried a few games vs GoC or EosZero?)
My Connect 4 AI, at normal computing time, cannot beat the best Connect 4 engines, also at normal computing time. I tried to put GoC on 0.1 computing time (by changing the code), but for some reason, the computing time wasn't very accurate. I was able to find out how to put computing time on EosZero though. Here is a game of EosZero vs Connect4AI (mine) at 0.1 seconds per move, with mine starting: 434444374333377711111177666666222222555555

Suprisingly it drew against my AI.

EosZero vs Connect4AI at 0.1 seconds per move, with yours starting: 455255544442222777711237745111133

A win against my AI.

Thanks for the games!

Technically, Connect 4 is a theoretical win for starting player, so a draw is still quite a good result for going second
#8442Sept. 16, 2025 23:20:32
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

I updated my Connect 4 AI with move bias. It actually stands a chance against GoC.
#8443Sept. 16, 2025 23:21:47
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

coooolboy55 wrote:

I updated my Connect 4 AI with move bias. It actually stands a chance against GoC.
Now compute time actually matters since the move bias, while making it better, also affects the amount of games played to accurately determine the best move.
#8444Sept. 17, 2025 02:05:00
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

I added win rate (which is a bit broken with scratch). Also, I found that the AI, if it can continue the game with 100% win rate, it will.
#8445Sept. 17, 2025 15:44:56
HasiLover
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

SB 3(5 sec, white) loses to Amarillo 1.92(5 sec) in 29 Moves: https://lichess.org/KlBzRbpI#0
WD(5 sec, white) luckily draws with 3fold against Amarillo 1.92(5 sec): https://lichess.org/MGaml3Zq/black 39 Moves

Last edited by HasiLover (Sept. 17, 2025 15:45:13)

#8446Sept. 17, 2025 19:42:21
Chompof78
Scratcher
1 post

Scratch Chess Engine - Game of Kings

I was told this was connect 4
#8447Sept. 17, 2025 20:14:51
grkw2020
Scratcher
500+ posts

Scratch Chess Engine - Game of Kings

Chompof78 wrote:

I was told this was connect 4
Mostly chess, some of us talk a bit about connect 4.
#8448Sept. 17, 2025 23:59:21
coder23ventana
Scratcher
7 posts

Scratch Chess Engine - Game of Kings

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 41)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 29)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (win at ply 25)
GoC              -  OC4 (AKOS)           1  -  0    (win at ply 33)
GoC              -  OC4VC (ABC)          1  -  0    (win at ply 31)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 34)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 31)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 39)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 42)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/
cAN I PLAY
#8449Sept. 18, 2025 01:24:28
grkw2020
Scratcher
500+ posts

Scratch Chess Engine - Game of Kings

coder23ventana wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 41)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 29)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (win at ply 25)
GoC              -  OC4 (AKOS)           1  -  0    (win at ply 33)
GoC              -  OC4VC (ABC)          1  -  0    (win at ply 31)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 34)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 31)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 39)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 42)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/
cAN I PLAY
That was a tournament between AIs… there are links to each of the AIs provided.
#8450Sept. 18, 2025 01:51:27
waabooboo
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

coder23ventana wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 41)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 29)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (win at ply 25)
GoC              -  OC4 (AKOS)           1  -  0    (win at ply 33)
GoC              -  OC4VC (ABC)          1  -  0    (win at ply 31)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 34)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (win at ply 31)
GoC              -  C4AI (Spentine)      1  -  0    (win at ply 39)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (loss at ply 42)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/
cAN I PLAY

No.
#8451Sept. 18, 2025 03:52:27
Nith04
Scratcher
9 posts

Scratch Chess Engine - Game of Kings

Hey guys, I'm making a Chess AI part of my AI studio “StructureBot”, but the thing is I'm trying to use more AI-like processing rather than many algorithms. I'm trying to make a generative AI that can do well in chess, with training data and stuff, and so far it does okay-ish if you input an algebraic notation it'll output one it saw in famous chess grandmaster games, but sometimes it just “runs out” of moves when it hasn't seen that. I thought with enough training data and optimization I could get it to actually understand the current chess board situation and do moves like such. Sometimes, when you input an algebraic notation of a chess move it recognizes that and captures that piece. But that's only occasionally, otherwise it's a bit inefficient as of now.

I would be happy if anyone knew how to improve my in-dev chess AI a bit, so please do reply.
when I receive [Reply]
say [I'll use this to make my AI better.]
#8452Sept. 18, 2025 04:00:10
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

My AI does okay against this now.
#8453Sept. 18, 2025 04:18:03
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

EosZero vs My improved bot, with 10 seconds each:
444447322224266662667335533131511117755
17 perfect moves in a row lol.
#8454Sept. 18, 2025 04:19:14
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

coooolboy55 wrote:

EosZero vs My improved bot, with 10 seconds each:
444447322224266662667335533131511117755
17 perfect moves in a row lol.
(EosZero started)
#8455Sept. 18, 2025 14:32:31
WhiteDoveOfficial
Scratcher
38 posts

Scratch Chess Engine - Game of Kings

Nith04 wrote:

Hey guys, I'm making a Chess AI part of my AI studio “StructureBot”, but the thing is I'm trying to use more AI-like processing rather than many algorithms. I'm trying to make a generative AI that can do well in chess, with training data and stuff, and so far it does okay-ish if you input an algebraic notation it'll output one it saw in famous chess grandmaster games, but sometimes it just “runs out” of moves when it hasn't seen that. I thought with enough training data and optimization I could get it to actually understand the current chess board situation and do moves like such. Sometimes, when you input an algebraic notation of a chess move it recognizes that and captures that piece. But that's only occasionally, otherwise it's a bit inefficient as of now.

I would be happy if anyone knew how to improve my in-dev chess AI a bit, so please do reply.
when I receive [Reply]
say [I'll use this to make my AI better.]

Hello!

It looks like you're trying to incorporate AI into a strong chess engine, and so far, the two most efficient ways found are NNUEs paired with alpha/beta search and deep neural networks paired with MCTS.

NNUE pros:
  • Very very fast on CPU
  • Requires less compute to train compared to deep neural nets
  • Very easily trainable on consumer grade hardware (even a single laptop can suffice)

NNUE cons:
  • Generally less “creative” than deep neural nets
  • Less accurate than deep neural nets with eval / policy ALONE (paired with search is a different story)


Deep Neural Nets pros:
  • Very accurate evaluation
  • Can play at near grandmaster levels with policy(intuition-ish) alone

Deep Neural Nets cons:
  • Needs GPU to run fast and efficiently
  • Very hard to train by yourself on a reasonable budget & timeframe using commercial grade hardware (well, excluding paid online cloud services)
  • Not very applicable to Scratch due to very very slow inference compared to NNUE

Overall, I would highly recommend using NNUEs on Scratch, as they are simply put easier and faster to run on Scratch, their training is easier, and they've already proven themselves on Scratch with engines such as Black Crow, Delta, or GoK NNUE.
#8456Sept. 19, 2025 04:07:32
coooolboy55
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

ArnoHu wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (41 moves)
GoC              -  C4AI (Spentine)      1  -  0    (29 moves)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (25 moves)
GoC              -  OC4 (AKOS)           1  -  0    (33 moves)
GoC              -  OC4VC (ABC)          1  -  0    (31 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (34 moves)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (31 moves)
GoC              -  C4AI (Spentine)      1  -  0    (39 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (42 moves)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/

Connect 4 Bot Game Series continues

Revisit this posting for updates during the next couple of days…

The New Bots

Scratch 3
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (24 moves)
EosZero             -  GoC                 1  -  0    (25 moves)
EosZero             -  GoC (updated)       0  -  1    (34 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (27 moves)
GoC                 -  EosZero             1  -  0    (39 moves)
GoC                 -  Connect 4 PNS (WBB) 1  -  0    (23 moves)


TurboWarp
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (38 moves)
EosZero             -  GoC                 0  -  1    (18 moves)
Connect 4 PNS (WBB) -  GoC (updated)       0  -  1    (36 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (21 moves)
GoC                 -  EosZero             1  -  0    (33 moves)
GoC                 -  Connect 4 PNS (WBB) 0  -  1    (32 moves)
GoC (updated)       -  Connect 4 PNS (WBB) 1  -  0    (33 moves)
Just saving this. Also, my bot has gotten better since it first played in this game series. (It probably could beat the version of GoC that played in this, as it was able to beat Connect 4 PNS (which could beat GoC at the time) in one of my games against it, GoC is stronger now though.)

Last edited by coooolboy55 (Sept. 19, 2025 04:17:11)

#8457Sept. 19, 2025 12:09:46
waabooboo
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

coooolboy55 wrote:

ArnoHu wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (41 moves)
GoC              -  C4AI (Spentine)      1  -  0    (29 moves)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (25 moves)
GoC              -  OC4 (AKOS)           1  -  0    (33 moves)
GoC              -  OC4VC (ABC)          1  -  0    (31 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (34 moves)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (31 moves)
GoC              -  C4AI (Spentine)      1  -  0    (39 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (42 moves)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/

Connect 4 Bot Game Series continues

Revisit this posting for updates during the next couple of days…

The New Bots

Scratch 3
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (24 moves)
EosZero             -  GoC                 1  -  0    (25 moves)
EosZero             -  GoC (updated)       0  -  1    (34 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (27 moves)
GoC                 -  EosZero             1  -  0    (39 moves)
GoC                 -  Connect 4 PNS (WBB) 1  -  0    (23 moves)


TurboWarp
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (38 moves)
EosZero             -  GoC                 0  -  1    (18 moves)
Connect 4 PNS (WBB) -  GoC (updated)       0  -  1    (36 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (21 moves)
GoC                 -  EosZero             1  -  0    (33 moves)
GoC                 -  Connect 4 PNS (WBB) 0  -  1    (32 moves)
GoC (updated)       -  Connect 4 PNS (WBB) 1  -  0    (33 moves)
Just saving this. Also, my bot has gotten better since it first played in this game series. (It probably could beat the version of GoC that played in this, as it was able to beat Connect 4 PNS (which could beat GoC at the time) in one of my games against it, GoC is stronger now though.)

I wonder if conspiracy number search might do better… then I could add a (basic) evaluation function. 4M for proof numbers and 4M for disproof numbers works fine, but I worry about creating additional large lists. I'm not sure how many of those I can make before Turbowarp crashes.
#8458Sept. 19, 2025 13:53:55
Destructor_chess
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

waabooboo wrote:

coooolboy55 wrote:

ArnoHu wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (41 moves)
GoC              -  C4AI (Spentine)      1  -  0    (29 moves)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (25 moves)
GoC              -  OC4 (AKOS)           1  -  0    (33 moves)
GoC              -  OC4VC (ABC)          1  -  0    (31 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (34 moves)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (31 moves)
GoC              -  C4AI (Spentine)      1  -  0    (39 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (42 moves)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/

Connect 4 Bot Game Series continues

Revisit this posting for updates during the next couple of days…

The New Bots

Scratch 3
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (24 moves)
EosZero             -  GoC                 1  -  0    (25 moves)
EosZero             -  GoC (updated)       0  -  1    (34 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (27 moves)
GoC                 -  EosZero             1  -  0    (39 moves)
GoC                 -  Connect 4 PNS (WBB) 1  -  0    (23 moves)


TurboWarp
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (38 moves)
EosZero             -  GoC                 0  -  1    (18 moves)
Connect 4 PNS (WBB) -  GoC (updated)       0  -  1    (36 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (21 moves)
GoC                 -  EosZero             1  -  0    (33 moves)
GoC                 -  Connect 4 PNS (WBB) 0  -  1    (32 moves)
GoC (updated)       -  Connect 4 PNS (WBB) 1  -  0    (33 moves)
Just saving this. Also, my bot has gotten better since it first played in this game series. (It probably could beat the version of GoC that played in this, as it was able to beat Connect 4 PNS (which could beat GoC at the time) in one of my games against it, GoC is stronger now though.)

I wonder if conspiracy number search might do better… then I could add a (basic) evaluation function. 4M for proof numbers and 4M for disproof numbers works fine, but I worry about creating additional large lists. I'm not sure how many of those I can make before Turbowarp crashes.
Oh, 4M is not really high for turbowarp… turbowarp handles (in my computer) 1B items lists… (yeah, idk how too xD)
#8459Sept. 19, 2025 16:32:37
waabooboo
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

Destructor_chess wrote:

waabooboo wrote:

coooolboy55 wrote:

ArnoHu wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (41 moves)
GoC              -  C4AI (Spentine)      1  -  0    (29 moves)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (25 moves)
GoC              -  OC4 (AKOS)           1  -  0    (33 moves)
GoC              -  OC4VC (ABC)          1  -  0    (31 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (34 moves)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (31 moves)
GoC              -  C4AI (Spentine)      1  -  0    (39 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (42 moves)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/

Connect 4 Bot Game Series continues

Revisit this posting for updates during the next couple of days…

The New Bots

Scratch 3
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (24 moves)
EosZero             -  GoC                 1  -  0    (25 moves)
EosZero             -  GoC (updated)       0  -  1    (34 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (27 moves)
GoC                 -  EosZero             1  -  0    (39 moves)
GoC                 -  Connect 4 PNS (WBB) 1  -  0    (23 moves)


TurboWarp
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (38 moves)
EosZero             -  GoC                 0  -  1    (18 moves)
Connect 4 PNS (WBB) -  GoC (updated)       0  -  1    (36 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (21 moves)
GoC                 -  EosZero             1  -  0    (33 moves)
GoC                 -  Connect 4 PNS (WBB) 0  -  1    (32 moves)
GoC (updated)       -  Connect 4 PNS (WBB) 1  -  0    (33 moves)
Just saving this. Also, my bot has gotten better since it first played in this game series. (It probably could beat the version of GoC that played in this, as it was able to beat Connect 4 PNS (which could beat GoC at the time) in one of my games against it, GoC is stronger now though.)

I wonder if conspiracy number search might do better… then I could add a (basic) evaluation function. 4M for proof numbers and 4M for disproof numbers works fine, but I worry about creating additional large lists. I'm not sure how many of those I can make before Turbowarp crashes.
Oh, 4M is not really high for turbowarp… turbowarp handles (in my computer) 1B items lists… (yeah, idk how too xD)

Yeah, but I would need a list of 4M items for every possible evaluation. I think that would probably be pushing it.
#8460Sept. 19, 2025 17:35:37
Destructor_chess
Scratcher
100+ posts

Scratch Chess Engine - Game of Kings

waabooboo wrote:

Destructor_chess wrote:

waabooboo wrote:

coooolboy55 wrote:

ArnoHu wrote:

ArnoHu wrote:

Connect 4 Bot Game Series

I am conducting a Connect 4 Bot test game series played by Game of Connect (GoC).

The results will be updated in this posting over time. They will cover the Scratch runtime and TurboWarp (TW only if an engine takes advantage of the additional performance there), with both sides as starting players once (as far as the projects allow). When several playing levels exist, games will be played on highest level available for the runtime.

The Bots


Scratch 3
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (41 moves)
GoC              -  C4AI (Spentine)      1  -  0    (29 moves)
GoC              -  C4AI-GL (PullJosh)   1  -  0    (25 moves)
GoC              -  OC4 (AKOS)           1  -  0    (33 moves)
GoC              -  OC4VC (ABC)          1  -  0    (31 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (34 moves)

TurboWarp
Starting Side       Responding Side

GoC              -  CFAI (BirdRacer)     1  -  0    (31 moves)
GoC              -  C4AI (Spentine)      1  -  0    (39 moves)

CFAI (BirdRacer) -  GoC                0.5  -  0.5
C4AI (Spentine)  -  GoC                  0  -  1    (42 moves)

Note: GoC and BirdRacer's CFAI are think-time bound, hence not deterministic. They will play according to system speed (means, better on faster systems).

If you reply to this post, I suggest to link to it and remove the copied content, otherwise we will have outdated results copied over the forum. Here is the link to the original: https://scratch.mit.edu/discuss/post/8319884/

Connect 4 Bot Game Series continues

Revisit this posting for updates during the next couple of days…

The New Bots

Scratch 3
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (24 moves)
EosZero             -  GoC                 1  -  0    (25 moves)
EosZero             -  GoC (updated)       0  -  1    (34 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (27 moves)
GoC                 -  EosZero             1  -  0    (39 moves)
GoC                 -  Connect 4 PNS (WBB) 1  -  0    (23 moves)


TurboWarp
Starting Side        Responding Side

ConnectAI (CB55)    -  GoC                 0  -  1    (38 moves)
EosZero             -  GoC                 0  -  1    (18 moves)
Connect 4 PNS (WBB) -  GoC (updated)       0  -  1    (36 moves)

GoC                 -  ConnectAI (CB55)    1  -  0    (21 moves)
GoC                 -  EosZero             1  -  0    (33 moves)
GoC                 -  Connect 4 PNS (WBB) 0  -  1    (32 moves)
GoC (updated)       -  Connect 4 PNS (WBB) 1  -  0    (33 moves)
Just saving this. Also, my bot has gotten better since it first played in this game series. (It probably could beat the version of GoC that played in this, as it was able to beat Connect 4 PNS (which could beat GoC at the time) in one of my games against it, GoC is stronger now though.)

I wonder if conspiracy number search might do better… then I could add a (basic) evaluation function. 4M for proof numbers and 4M for disproof numbers works fine, but I worry about creating additional large lists. I'm not sure how many of those I can make before Turbowarp crashes.
Oh, 4M is not really high for turbowarp… turbowarp handles (in my computer) 1B items lists… (yeah, idk how too xD)

Yeah, but I would need a list of 4M items for every possible evaluation. I think that would probably be pushing it.
oh lol, it wont handle this if it means what that i suppose it to means…

Powered by DjangoBB