lichess.org
Donate

Proof Games wanted to improve accuracy of 4x10^44 estimated legal positions

My Chess Position Ranking project at github.com/tromp/ChessPositionRanking
has produced the first accurate estimation of the number of legal chess positions, at
4.2e44 +- 1.1e44 with a 95% confidence level, based on a sample of 1000 random positions.

I want to improve the accuracy by determining the number of legal positions within a much larger sample of 10000 random positions. With 9081 of them are clearly illegal, this leaves 919 to be analyzed by hand.

I have filed an issue for each one of these 919 under the "10k" milestone at github.com/tromp/ChessPositionRanking/issues

I would like to ask interested chess players to help me prove as legal the positions labeled "proof game wanted",
by composing a game reaching the position and posting the PGN record as a comment.

Thanks in advance!

My Chess Position Ranking project at github.com/tromp/ChessPositionRanking has produced the first accurate estimation of the number of legal chess positions, at 4.2e44 +- 1.1e44 with a 95% confidence level, based on a sample of 1000 random positions. I want to improve the accuracy by determining the number of legal positions within a much larger sample of 10000 random positions. With 9081 of them are clearly illegal, this leaves 919 to be analyzed by hand. I have filed an issue for each one of these 919 under the "10k" milestone at github.com/tromp/ChessPositionRanking/issues I would like to ask interested chess players to help me prove as legal the positions labeled "proof game wanted", by composing a game reaching the position and posting the PGN record as a comment. Thanks in advance!

@tromp1234 I figured a solution to position #846, this was one of the most time consuming things I did! However glad it got through, anyways here is the solution and link to the position on your site!

https://adjva4.dpdns.org/or8DLx9D#0
(solution)

https://github.com/tromp/ChessPositionRanking/issues/846 (site)

@tromp1234 I figured a solution to position #846, this was one of the most time consuming things I did! However glad it got through, anyways here is the solution and link to the position on your site! https://adjva4.dpdns.org/or8DLx9D#0 (solution) https://github.com/tromp/ChessPositionRanking/issues/846 (site)

@tromp1234
That's great!
Being a chess researcher, I would recommend you to use: https://wismuth.com/jacobi/
Follow what it says to run the program and it solves the proof games.
Also, you are only talking about proof games not shortest proof game. I am not sure whether the program does that as well.
In all, it's a great tool to use.

@tromp1234 That's great! Being a chess researcher, I would recommend you to use: https://wismuth.com/jacobi/ Follow what it says to run the program and it solves the proof games. Also, you are only talking about proof games not shortest proof game. I am not sure whether the program does that as well. In all, it's a great tool to use.

Thanks, @Ben2006Tor and @IsaiasZM !
In future, could you please also paste the game record in PGN format?
It's some effort to transcribe it from the links you provide.

Thanks, @Ben2006Tor and @IsaiasZM ! In future, could you please also paste the game record in PGN format? It's some effort to transcribe it from the links you provide.

If you took one of these already solved positions but did it for the other side to move would that count as a new position?

If you took one of these already solved positions but did it for the other side to move would that count as a new position?

Also do people that solve one of the positions get credit somewhere like naming rights given to the scientist that discovers a new thing? If you did that I could see it increasing participation by a lot.

Also do people that solve one of the positions get credit somewhere like naming rights given to the scientist that discovers a new thing? If you did that I could see it increasing participation by a lot.

@Linspiring said in #9:

Also do people that solve one of the positions get credit somewhere like naming rights given to the scientist that discovers a new thing? If you did that I could see it increasing participation by a lot.

If you change side to move then it's a different position.

I encourage people to register a github account so that they can submit their proof game as a comment. That's the best way to get credit as having resolved an issue. For now I will mention their name in a comment when I transcribe the PGN record.
Once I write a paper I will name the top 50 contributors to resolving issues in an Acknowledgement section.

@Linspiring said in #9: > Also do people that solve one of the positions get credit somewhere like naming rights given to the scientist that discovers a new thing? If you did that I could see it increasing participation by a lot. If you change side to move then it's a different position. I encourage people to register a github account so that they can submit their proof game as a comment. That's the best way to get credit as having resolved an issue. For now I will mention their name in a comment when I transcribe the PGN record. Once I write a paper I will name the top 50 contributors to resolving issues in an Acknowledgement section.

This topic has been archived and can no longer be replied to.