New Cerebellum Library

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
clumma
Posts: 177
Joined: Fri Oct 10, 2014 8:05 pm
Location: Berkeley, CA

Re: New Cerebellum Library

Post by clumma » Wed Mar 27, 2019 12:14 am

Dann Corbit wrote:
Wed Mar 27, 2019 12:06 am
It is trivial to turn it into PGN.
Some of us already have all the positions deeply analyzed.
Ok, how?

-Carl

Dann Corbit
Posts: 10187
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: New Cerebellum Library

Post by Dann Corbit » Wed Mar 27, 2019 12:32 am

You can turn it into PGN with bin2pgn.
That PGN is crappy PGN, because it uses algebraic.
So I feed it to pgn-extract to turn it into more suitable PGN.
When you finish those two steps you will have something that looks a bit like the attached file.
Attachments
cerebellum.7z
(5.75 KiB) Downloaded 41 times
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Dann Corbit
Posts: 10187
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: New Cerebellum Library

Post by Dann Corbit » Wed Mar 27, 2019 12:45 am

It contains the following major ECO codes:
"B51", "B90", "C50", "C53", "C54", "C58", "C67", "C80", "C83", "C88", "C89", "D11", "D30", "D37", "D43", "D45", "D46", "E39"
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

jdart
Posts: 3835
Joined: Fri Mar 10, 2006 4:23 am
Location: http://www.arasanchess.org

Re: New Cerebellum Library

Post by jdart » Wed Mar 27, 2019 1:04 am

Eelco de Groot wrote:
Tue Mar 26, 2019 9:39 pm
I'm hoping for a robot that can update the great NCO or some similar work. In textform. With all the computer analyses and human style comments, pages to flip through. Somebody would make spreadsheets from it I think. I would pay for that even if it was 10 volumes. But it would not make the publisher or authors rich I'm afraid, or would still be terribly expensive for us buyers even with robot help to make.
Back in the old days of ECO it was edited by grandmasters, and it was nice to be able to browse the physical books. I still have some of the early editions. But I am afraid the era for that is mostly over. There are still some exceptions: NIC Yearbooks for example are still print (you can get the games online but minus annotations).

--Jon

clumma
Posts: 177
Joined: Fri Oct 10, 2014 8:05 pm
Location: Berkeley, CA

Re: New Cerebellum Library

Post by clumma » Wed Mar 27, 2019 1:20 am

Dann Corbit wrote:
Wed Mar 27, 2019 12:32 am
You can turn it into PGN with bin2pgn.
That PGN is crappy PGN, because it uses algebraic.
So I feed it to pgn-extract to turn it into more suitable PGN.
When you finish those two steps you will have something that looks a bit like the attached file.
Thanks!

But this pgn doesn't have the scores (?)

-Carl

clumma
Posts: 177
Joined: Fri Oct 10, 2014 8:05 pm
Location: Berkeley, CA

Re: New Cerebellum Library

Post by clumma » Wed Mar 27, 2019 1:25 am

In case anyone finds it useful, by hand I copied these lines from the Cerebellum demo

https://gist.github.com/clumma/f340b632 ... c8cc1cddb6

This was done in 2017. The demo has seemingly been non-functional for some time now.

-Carl

clumma
Posts: 177
Joined: Fri Oct 10, 2014 8:05 pm
Location: Berkeley, CA

Re: New Cerebellum Library

Post by clumma » Wed Mar 27, 2019 1:26 am

Dann Corbit wrote:
Wed Mar 27, 2019 12:32 am
You can turn it into PGN with bin2pgn.
That PGN is crappy PGN, because it uses algebraic.
So I feed it to pgn-extract to turn it into more suitable PGN.
When you finish those two steps you will have something that looks a bit like the attached file.
I found pgn-extract but I can't find bin2pgn. Do you have a link?

-Carl

Ferdy
Posts: 4111
Joined: Sun Aug 10, 2008 1:15 pm
Location: Philippines

Re: New Cerebellum Library

Post by Ferdy » Wed Mar 27, 2019 1:42 am

clumma wrote:
Wed Mar 27, 2019 12:14 am
Dann Corbit wrote:
Wed Mar 27, 2019 12:06 am
It is trivial to turn it into PGN.
Some of us already have all the positions deeply analyzed.
Ok, how?

-Carl
polyglot can dump the bin to text lines.

polyglot.exe dump-book -bin Cerebellum_Light_Poly.bin -color white -out cere_white.txt
polyglot.exe dump-book -bin Cerebellum_Light_Poly.bin -color black -out cere_black.txt

Combine:
copy cere_white.txt + cere_black.txt all_cere.txt

all_cere.txt would look like this.

Code: Select all

Dump of "Cerebellum_Light_Poly.bin" for white.
1: 1. d4{33%} a6 2. e4{100%} a5 3. Nc3{100%} e6 4. Bd3{100%} d6 5. Qe2{100%} 
2: 1. d4{33%} a6 2. e4{100%} b6 3. Nf3{100%} d6 4. Bd3{100%} e6 5. c4{100%} 
[...]
Then process this kind of output using a script, then finally use pgn-extract to reformat, remove duplicates, add eco and correct some error if there is.

Sample final pgn output would be:

For white:

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[ECO "D45"]
[Opening "QGD semi-Slav"]
[Variation "5...Nd7"]

1. d4 { 33% } 1... d5 2. c4 { 67% } 2... e6 3. Nc3 { 33% } 3... c6 4. e3 {
100% } 4... Nd7 5. Nf3 { 67% } 5... Ngf6 6. Be2 { 33% } 6... b6 7. Qc2 {
100% } 7... Bb7 8. O-O { 100% } 8... Bd6 9. a4 { 100% } 9... Bb4 10. cxd5 {
100% } 10... exd5 11. e4 { 100% } 11... dxe4 12. Nxe4 { 100% } 12... O-O
13. Bg5 { 100% } 13... h6 14. Bh4 { 100% } 14... Rc8 15. a5 { 100% } *

For black:

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[ECO "E54"]
[Opening "Nimzo-Indian"]
[Variation "4.e3, Gligoric system with 7...dc"]

1. c4 e6 { 67% } 2. d4 Nf6 { 100% } 3. Nc3 Bb4 { 100% } 4. Nf3 O-O { 100% }
5. e3 d5 { 100% } 6. Be2 dxc4 { 100% } 7. Bxc4 c5 { 100% } 8. O-O cxd4 {
100% } 9. exd4 b6 { 100% } 10. Qb3 Bxc3 { 100% } 11. bxc3 Nc6 { 100% } 12.
Qc2 Bb7 { 100% } 13. Bd3 Rc8 { 100% } 14. Ba3 Re8 { 100% } 15. Rfd1 Qd5 {
100% } 16. Bf1 Na5 { 100% } 17. Ne5 Nd7 { 100% } 18. Nxd7 Qxd7 { 100% } 19.
Bb4 Qd5 { 100% } 20. a4 Red8 { 33% } *

Planning to import it to bookup or cpt, extract the end position into epd and then convert the last move comment say 33% into centipawn eval (or analyze the end position with an engine but takes time), the desired output would be
2rr2k1/pb3ppp/1p2p3/n2q4/PB1P4/2P5/2Q2PPP/R2R1BK1 w - - ce -120;

Import pgn and then import epd with ce, then backsolve/mini-max in bookup/cpt.

Nay Lin Tun
Posts: 529
Joined: Mon Jan 16, 2012 5:34 am

Re: New Cerebellum Library

Post by Nay Lin Tun » Wed Mar 27, 2019 12:36 pm

I cant see cerebellum demo evaluation scores on his website! My browser problem or it is the same to you ?
Attachments
Capture (2).PNG
Capture (2).PNG (115.75 KiB) Viewed 923 times

clumma
Posts: 177
Joined: Fri Oct 10, 2014 8:05 pm
Location: Berkeley, CA

Re: New Cerebellum Library

Post by clumma » Wed Mar 27, 2019 3:58 pm

Nay Lin Tun wrote:
Wed Mar 27, 2019 12:36 pm
I cant see cerebellum demo evaluation scores on his website! My browser problem or it is the same to you ?
It's been like that for a year.

-Carl

Post Reply