ProDeo
Would you like to react to this message? Create an account in a few clicks or log in to continue.
ProDeo

Computer Chess
 
HomeHome  CalendarCalendar  Latest imagesLatest images  FAQFAQ  SearchSearch  MemberlistMemberlist  UsergroupsUsergroups  RegisterRegister  Log in  

 

 ProDeo 3.0

Go down 
+5
Chris Whittington
Peter Berger
Brendan
Mclane
Admin
9 posters
Go to page : 1, 2  Next
AuthorMessage
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: ProDeo 3.0   ProDeo 3.0 EmptyTue Nov 17, 2020 9:34 am

Playing strength first, I collected a number of changes during the last 4 years resulting in a ~35 elo improvement. ~20 elo came from a new idea replacing IID with what I called IIR. A number of other engines (among them Stockfish) adopted the idea.

Polyglot opening books, improved the standard ProDeo opening book with +46 elo, see elsewhere. Created additional books with new options, an overview:

......

http://rebel13.nl/prodeo/prodeo-3.0.html

Chris Whittington, Mclane, Brendan, matejst, MikeGL and Charly like this post

Back to top Go down
http://rebel13.nl/
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyTue Nov 17, 2020 10:49 pm

Code:
[Event "I7_7500U"]
[Site "ORION8"]
[Date "2020.11.17"]
[Round "1"]
[White "Prodeo3"]
[Black "Stockfish_nnue.2020-07-19 [t2,sergio]"]
[Result "1/2-1/2"]
[ECO "D44"]
[Opening "Damengambit (Botwinnik-Variante), Hauptvariante, 12.g3 c5 13.d5 Qb6"]
[Time "22:18:07"]
[Variation "Botvinnik (Anti-Meran)"]
[WhiteElo "2200"]
[TimeControl "40/300:40/300:40/300"]
[Termination "adjudication"]
[PlyCount "131"]
[WhiteType "program"]
[BlackType "program"]

1. d4 {0.00/1 0} Nf6 2. c4 {0.00/1 0} e6 3. Nf3 {0.00/1 0} d5 4. Nc3
{0.00/1 0} c6 5. Bg5 {0.00/1 0} dxc4 6. e4 {0.00/1 0} b5 7. e5 {0.00/1 0}
h6 8. Bh4 {0.00/1 0} g5 9. Nxg5 {0.00/1 0} hxg5 10. Bxg5 {0.00/1 0} Nbd7
{-0.17/25 16} 11. exf6 {0.00/1 0} Bb7 {-0.22/24 5} 12. g3 {0.00/1 0} c5
{-0.38/26 29} 13. d5 {0.00/1 0} Nxf6 {-0.68/27 18} 14. Bg2 {0.00/1 0} Bh6
{-0.58/26 3} 15. Bxf6 {0.00/1 0} Qxf6 {-0.34/28 6} 16. O-O {0.00/1 0} Rd8
{-0.60/29 14} 17. Qe2 {+1.26/16 13} O-O {-0.29/29 0} 18. d6 {+0.92/17 19}
b4 {-0.67/30 0} 19. Bxb7 {+1.08/17 5} bxc3 {-0.41/31 5} 20. bxc3 {+1.06/18
15} Rd7 {-0.37/32 0} 21. Be4 {+1.33/18 16} Qxc3 {-0.34/29 7} 22. Qh5
{+1.28/18 22} Qf6 {-0.14/28 7} 23. Rad1 {+1.00/17 9} Rc8 {-0.15/29 0} 24.
Bc2 {+1.13/16 8} Qg5 {-0.13/27 0} 25. Qe2 {+0.82/17 10} Qd8 {-0.33/34 28}
26. Qxc4 {+0.79/18 0} Bg7 {-0.11/28 5} 27. Qe4 {+0.90/16 8} Rxd6 {-0.09/29
11} 28. Rxd6 {+0.55/15 0} Qxd6 {-0.10/30 6} 29. Rb1 {+0.63/16 5} Rb8
{-0.07/33 15} 30. Rxb8+ {+0.50/15 11} Qxb8 {0.00/34 0} 31. Bb3 {+0.21/16
10} Bd4 {0.00/37 7} 32. Kg2 {+0.28/16 5} Qb4 {-0.01/34 11} 33. Qg4+
{+0.26/16 0} Kf8 {0.00/34 8} 34. Qf4 {+0.21/16 1} Qb5 {0.00/38 15} 35. h4
{+0.32/16 16} a5 {0.00/39 0} 36. h5 {+0.39/16 13} a4 {0.00/38 15} 37. Bd1
{+0.46/18 0} Qb1 {0.00/40 13} 38. Qh6+ {+0.42/17 14} Kg8 {0.00/41 0} 39.
Qg5+ {+0.23/17 12} Kh8 {0.00/41 19} 40. Qd2 {+0.23/18 0} Qe4+ {0.00/41 10}
41. Kh2 {+0.21/16 0} Qf5 {0.00/37 10} 42. g4 {+0.27/18 19} Qd5 {0.00/32 6}
43. Qh6+ {+0.28/17 9} Kg8 {0.00/35 0} 44. Bc2 {+0.25/18 16} Be5+ {0.00/37
0} 45. Kg1 {+0.30/18 11} Bg7 {0.00/43 0} 46. Qh7+ {+0.30/17 6} Kf8 {0.00/43
0} 47. Qe4 {+0.14/17 11} Qxa2 {0.00/33 7} 48. Qxa4 {+0.18/16 8} Qxa4
{0.00/40 5} 49. Bxa4 {+0.31/19 1} Bh6 {0.00/44 5} 50. Kf1 {+0.25/20 2} c4
{0.00/42 7} 51. Ke2 {+0.20/21 14} Kg7 {0.00/50 0} 52. Kd1 {+0.20/20 7} Kf6
{0.00/44 7} 53. Kc2 {+0.25/20 9} Kg5 {0.00/54 7} 54. Be8 {+0.13/20 8} Kxg4
{0.00/55 9} 55. Bxf7 {+0.45/22 11} Bg7 {0.00/56 0} 56. Bxe6+ {+0.45/22 7}
Kxh5 {0.00/57 0} 57. Bxc4 {+0.45/23 12} Kg5 {0.00/59 0} 58. Kd3 {+0.45/22
11} Kf4 {0.00/62 0} 59. Bd5 {+0.45/22 7} Be5 {0.00/61 1} 60. f3 {+0.45/22
11} Bf6 {0.00/57 10} 61. Be4 {+0.45/22 0} Bh4 {0.00/57 10} 62. Kd4
{+0.45/21 8} Bg5 {0.00/58 19} 63. Kd5 {+0.45/21 8} Bh6 {0.00/61 10} 64. Kd6
{+0.45/21 9} Kg5 {0.00/56 10} 65. Ke5 {+0.45/21 8} Bg7+ {0.00/60 10} 66.
Kd5 {+0.45/23 0} ... {½-½ Benutzer Entscheidung} 1/2-1/2

MikeGL likes this post

Back to top Go down
http://www.thorstenczub.de
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyTue Nov 17, 2020 11:13 pm

Optimal use of the new book.

Mclane likes this post

Back to top Go down
http://rebel13.nl/
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyTue Nov 17, 2020 11:43 pm

Code:
[Event "I7_7500U"]
[Site "ORION8"]
[Date "2020.11.17"]
[Round "1"]
[White "Prodeo3"]
[Black "Wasp400-x64-modern [t2]"]
[Result "1/2-1/2"]
[BlackElo "2200"]
[ECO "B12"]
[Opening "Caro-Kann"]
[Time "23:22:25"]
[Variation "Vorstoß-Variante, 4.Nf3 e6 5.Be2 c5 6.O-O Ne7"]
[WhiteElo "2200"]
[TimeControl "40/300:40/300:40/300"]
[Termination "adjudication"]
[PlyCount "145"]
[WhiteType "program"]
[BlackType "program"]

1. e4 {0.00/1 0} c6 2. d4 {0.00/1 0} d5 3. e5 {0.00/1 0} Bf5 4. Nf3 {0.00/1
0} e6 5. Be2 {0.00/1 0} Ne7 6. O-O {0.00/1 0} Nd7 7. Nbd2 {0.00/1 0} h6 8.
Nb3 {0.00/1 0} Ng6 {-0.43/22 12} 9. Ne1 {0.00/1 0} h5 {-0.36/20 6} 10. Nd3
{0.00/1 0} h4 {-0.51/20 6} 11. Bg4 {+0.48/15 12} h3 {-0.56/22 11} 12. g3
{+0.41/17 1} Bxg4 {-0.60/22 6} 13. Qxg4 {+0.46/16 6} b6 {-0.60/22 7} 14.
Bg5 {+0.87/17 31} Be7 {-0.89/24 0} 15. Bxe7 {+0.74/18 14} Qxe7 {-0.81/22 7}
16. f4 {+0.83/17 3} c5 {-0.82/22 7} 17. f5 {+0.75/17 8} exf5 {-0.92/23 7}
18. Qxf5 {+0.61/17 7} O-O {-0.87/23 4} 19. Nf2 {+0.48/17 12} cxd4 {-0.04/22
8} 20. Rae1 {+0.29/18 0} Ndxe5 {+0.12/23 8} 21. Nxh3 {+0.07/17 0} Rae8
{0.00/25 11} 22. Qh5 {+0.28/17 3} f6 {0.00/26 5} 23. a3 {+0.17/18 4} Qe6
{+0.02/26 6} 24. Nxd4 {0.00/18 10} Qg4 {0.00/27 0} 25. Qxg4 {0.00/19 9}
Nxg4 {0.00/27 0} 26. Ne6 {0.00/19 8} Rf7 {+0.12/27 11} 27. Nf2 {-0.10/21 0}
Nxf2 {+0.13/26 11} 28. Kxf2 {-0.07/21 0} Ne5 {+0.06/26 11} 29. Nd4
{-0.13/21 4} Rc7 {+0.05/25 7} 30. Kg2 {+0.09/18 12} Ree7 {+0.05/25 1} 31.
b3 {0.00/19 12} Ng4 {0.00/27 3} 32. Rxe7 {0.00/19 4} Rxe7 {0.00/29 9} 33.
Kg1 {0.00/18 0} a6 {-0.01/28 14} 34. h3 {+0.17/20 12} Ne5 {-0.06/28 2} 35.
Re1 {+0.18/19 8} b5 {0.00/29 9} 36. g4 {+0.26/19 15} Rc7 {-0.04/28 19} 37.
Re3 {+0.16/19 11} Kf7 {0.00/30 6} 38. Kh2 {+0.16/19 7} Rc8 {0.00/30 22} 39.
Kg3 {+0.23/18 8} g5 {0.00/31 15} 40. Kf2 {+0.20/21 0} Kf8 {0.00/31 16} 41.
Ke2 {+0.17/20 12} Ke7 {0.00/27 5} 42. Nf5+ {+0.18/19 3} Kd7 {0.00/27 5} 43.
Kd1 {+0.18/18 2} Rh8 {0.00/28 5} 44. Nd4 {+0.21/18 5} Kc7 {0.00/27 5} 45.
Kd2 {+0.13/18 11} Rh6 {0.00/28 0} 46. Kc3 {+0.15/19 12} Rh7 {0.00/29 0} 47.
Nf5 {+0.19/19 10} Kb6 {0.00/30 0} 48. Kd4 {+0.16/20 11} Rd7 {0.00/29 0} 49.
b4 {+0.16/19 9} Nc6+ {0.00/26 5} 50. Kd3 {+0.10/19 2} Ne5+ {0.00/29 2} 51.
Kc3 {+0.13/18 9} a5 {0.00/25 6} 52. bxa5+ {+0.11/19 21} Kxa5 {0.00/26 6}
53. Kd4 {+0.01/19 0} Kb6 {+0.01/27 5} 54. Rb3 {+0.10/18 0} Nc6+ {0.00/27 5}
55. Kd3 {+0.03/19 3} Rh7 {+0.09/27 12} 56. Ke2 {+0.08/20 0} Ra7 {+0.12/26
14} 57. Kf2 {0.00/19 11} Na5 {+0.11/25 12} 58. Rd3 {0.00/22 0} Nc4 {0.00/25
9} 59. Ne3 {+0.09/19 9} d4 {+0.02/26 8} 60. Rxd4 {0.00/20 2} Nxe3 {0.00/28
4} 61. Kxe3 {0.00/19 0} Rxa3+ {0.00/29 7} 62. Ke4 {+0.03/19 1} Rxh3
{+0.10/27 8} 63. Rd6+ {0.00/21 9} Ka5 {+0.06/29 0} 64. Rxf6 {0.00/21 10}
Rg3 {0.00/28 1} 65. Kf5 {0.00/20 9} Rc3 {0.00/31 0} 66. Kxg5 {0.00/18 10}
Rxc2 {0.00/31 0} 67. Rf1 {0.00/19 8} b4 {0.00/31 2} 68. Kf6 {0.00/18 4} b3
{0.00/32 7} 69. g5 {0.00/20 13} Rc6+ {0.00/31 0} 70. Kf7 {0.00/18 9} Kb4
{0.00/30 5} 71. g6 {0.00/18 8} Rc7+ {0.00/31 7} 72. Kf6 {0.00/19 2} Rc6+
{0.00/31 14} 73. Kg5 {0.00/20 0} ... {½-½ Benutzer Entscheidung} 1/2-1/2
Back to top Go down
http://www.thorstenczub.de
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyWed Nov 18, 2020 1:56 am

I get amazing results on all machines I use with this new prodeo3.
Seems you made it Ed.

It’s now 2h and I will go to sleep. But that’s really interesting new “change”.

Admin likes this post

Back to top Go down
http://www.thorstenczub.de
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyWed Nov 18, 2020 9:05 am

In my testing I only measured ~35 elo  What a Face

http://rebel13.nl/dump/pd3.7z

Code:
No. Engine           1     2  Score  Games   Perc   Moves
---------------------------------------------------------
 1 ProDeo 3.0    xxxx 1105.0 1105.0 / 2000 (55.25%)  72.4  
 2 ProDeo 2.2    895.0  xxxx  895.0 / 2000 (44.75%)  72.4  

Code:
Engine                 Depth       Time   Games     Moves  Average Forfeit  Book Depth     MIDG   EARLY    ENDG    LATE
ProDeo 2.2             14.74  115:22:09    2000    143547    2.89     0    15832  7.92    13.23 | 14.06 | 14.76 | 18.35
ProDeo 3.0             16.11  114:25:36    2000    143459    2.87     0    15830  7.91    14.71 | 15.61 | 16.00 | 19.64

Played without books.
Back to top Go down
http://rebel13.nl/
Brendan




Posts : 398
Join date : 2020-11-18
Age : 40

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyWed Nov 18, 2020 2:50 pm

Admin wrote:
In my testing I only measured ~35 elo  What a Face



Code:
No. Engine           1     2  Score  Games   Perc   Moves
---------------------------------------------------------
 1 ProDeo 3.0    xxxx 1105.0 1105.0 / 2000 (55.25%)  72.4  
 2 ProDeo 2.2    895.0  xxxx  895.0 / 2000 (44.75%)  72.4  

Code:
Engine                 Depth       Time   Games     Moves  Average Forfeit  Book Depth     MIDG   EARLY    ENDG    LATE
ProDeo 2.2             14.74  115:22:09    2000    143547    2.89     0    15832  7.92    13.23 | 14.06 | 14.76 | 18.35
ProDeo 3.0             16.11  114:25:36    2000    143459    2.87     0    15830  7.91    14.71 | 15.61 | 16.00 | 19.64

Played without books.

I think I'm missing something with Prodeo 3, Ed.

What were the changes outside of adding polyglot functionality? Changes to the engine itself?
Back to top Go down
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyWed Nov 18, 2020 3:26 pm

Brendan wrote:
Admin wrote:
In my testing I only measured ~35 elo  What a Face



Code:
No. Engine           1     2  Score  Games   Perc   Moves
---------------------------------------------------------
 1 ProDeo 3.0    xxxx 1105.0 1105.0 / 2000 (55.25%)  72.4  
 2 ProDeo 2.2    895.0  xxxx  895.0 / 2000 (44.75%)  72.4  

Code:
Engine                 Depth       Time   Games     Moves  Average Forfeit  Book Depth     MIDG   EARLY    ENDG    LATE
ProDeo 2.2             14.74  115:22:09    2000    143547    2.89     0    15832  7.92    13.23 | 14.06 | 14.76 | 18.35
ProDeo 3.0             16.11  114:25:36    2000    143459    2.87     0    15830  7.91    14.71 | 15.61 | 16.00 | 19.64

Played without books.

I think I'm missing something with Prodeo 3, Ed.

What were the changes outside of adding polyglot functionality? Changes to the engine itself?

Several (five) search changes, it made ProDeo 3 a somewhat better comp-comp fighter, but all that extra pruning at the cost of playing style.

Brendan and matejst like this post

Back to top Go down
http://rebel13.nl/
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 3:43 pm

Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2    854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.
Back to top Go down
http://rebel13.nl/
Peter Berger




Posts : 130
Join date : 2020-11-20

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 4:43 pm

ProDeo is still ProDeo/Rebel.

But the book ideas are pretty cool. This comes +extremely+ close to what Marc Lacrosse and me did in 2004/2005, +manually+ ( investing an +insane+ amount of time) I may add Laughing .

It is cool to see what a programmer can do with such a setup ( minus some of the things book cooks can still do better). I think you default to a WAY too long book depth btw, but I really like the general setup.
Back to top Go down
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 8:11 pm

Admin wrote:
Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2    854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.

I wish Ed you would stop testing against older Rebell versions.
IMO this only measures how different a new version is. The opponents have to be
4-8 different strength and style engines.

Only One opponent is IMO giving wrong data,
Back to top Go down
http://www.thorstenczub.de
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 8:41 pm

Peter Berger wrote:
ProDeo is still ProDeo/Rebel.

But the book ideas are pretty cool. This comes +extremely+ close to what Marc Lacrosse and me did in 2004/2005, +manually+ ( investing an +insane+ amount of time)  I may add Laughing .

It is cool to see what a programmer can do with such a setup ( minus some of the things book cooks can still do better). I think you default to a WAY too long book depth btw, but I really like the general setup.

You remind me I still have to test the new book for blunders, forgot all about it due to the setup of the forum. Although the book is made from elo 2700+ humans players it's not unlikely it contains a number of losing variations. I have started the procedure. Basically, run 2000 games at a decent time control with SF12 and modify the book with the POLY tool. Nowadays everything goes automatic, no more hand tuning.

Mclane likes this post

Back to top Go down
http://rebel13.nl/
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 8:59 pm

Mclane wrote:
Admin wrote:
Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2    854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.

I wish Ed you would stop testing against older Rebell versions.
IMO this only measures how different a new version is. The opponents have to be
4-8 different strength and style engines.

Only One opponent is IMO giving wrong data,

Stockfish exclusively are doing incest testing, seems to work quite well. Besides, I am not very interested in playing strength any longer, just following the progress others make is fascinating enough. Playing style is a different issue, Benjamin 1.0 was a good start.

Mclane likes this post

Back to top Go down
http://rebel13.nl/
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 9:14 pm

Yes you are right, Benjamin 1 was ok.
I liked the style it did.

Ever thought about making prodeo SMP ?
Today any person has a pc with 2-64 cores, even my mobile phone has 8 cores.
Back to top Go down
http://www.thorstenczub.de
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 9:34 pm

There was a time I had to do things I disliked, things that had nothing to do with chess, programming the leds on the board, programming the keys, programming the clock, creating your own character set for the screen, scanning the board for a move. Programming SMP is just like that, nothing to do with chess, I am free now.

Mclane and matejst like this post

Back to top Go down
http://rebel13.nl/
Chris Whittington




Posts : 1254
Join date : 2020-11-17
Location : France

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyFri Nov 20, 2020 9:42 pm

Admin wrote:
There was a time I had to do things I disliked, things that had nothing to do with chess, programming the leds on the board, programming the keys, programming the clock, creating your own character set for the screen, scanning the board for a move. Programming SMP is just like that, nothing to do with chess, I am free now.

That’s what I thought, but actually SMP is just a few lines of code and a header file threads.h or pthreads.h
The basic idea is well known, very simple to implement. And assuming one’s engine isn’t writing to globals all over the place.

Mclane likes this post

Back to top Go down
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySat Nov 21, 2020 10:54 am

It does have globals all over the place, secondly I need to switch compilers, 1+1=NO.
Back to top Go down
http://rebel13.nl/
Peter Berger




Posts : 130
Join date : 2020-11-20

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySat Nov 21, 2020 7:20 pm

Admin wrote:
Peter Berger wrote:
ProDeo is still ProDeo/Rebel.

But the book ideas are pretty cool. This comes +extremely+ close to what Marc Lacrosse and me did in 2004/2005, +manually+ ( investing an +insane+ amount of time)  I may add Laughing .

It is cool to see what a programmer can do with such a setup ( minus some of the things book cooks can still do better). I think you default to a WAY too long book depth btw, but I really like the general setup.

You remind me I still have to test the new book for blunders, forgot all about it due to the setup of the forum. Although the book is made from elo 2700+ humans players it's not unlikely it contains a number of losing variations. I have started the procedure. Basically, run 2000 games at a decent time control with SF12 and modify the book with the POLY tool. Nowadays everything goes automatic, no more hand tuning.

Yes, blunderchecking the book is worthwhile IMHO, and Marc and me did this for hours on end.

For the thing book cooks still can do better:

https://www.chessgames.com/perl/chessgame?gid=1417766

All these lines were entered into the book the day before the game btw. The Crafty version in Turino was actually +way+ weaker than the one that played in Ramat-Gan ( programmers didn't test like they do now) , but once THIS line was on the board I knew Diep was toast ( and Vincent knew, too Wink ) - those were the days Twisted Evil. Last book move was 17.. a5 and Crafty still had to find Re8 on its own , I sweated a little Very Happy

Admin and TheSelfImprover like this post

Back to top Go down
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySat Nov 21, 2020 7:58 pm

Admin wrote:
Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2     854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.

After book tuning +84 elo, 33 elo more.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx 620.0  620.0 / 1000 (62.00%)  75.2  
 2 ProDeo 2.2    380.0 xxxxx  380.0 / 1000 (38.00%)  75.2  

Suspect openings before:  http://rebel13.nl/text/result-1.txt
Suspect openings after:  http://rebel13.nl/text/result-2.txt

Download new book: http://rebel13.nl/dl/ProDeo-1.7z

Disadvantage, book is less varied.

Mclane and Charly like this post

Back to top Go down
http://rebel13.nl/
Brendan




Posts : 398
Join date : 2020-11-18
Age : 40

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySun Nov 22, 2020 11:38 am

Peter Berger wrote:

Yes, blunderchecking the book is worthwhile IMHO, and Marc and me did this for hours on end.

Totally. But I think it would also be cool to keep the original (or an even weakened one) for human play.

Perhaps in UCI options can choose "Strong book" and "Human book" from a dropdown or something?

It is quite common in human play to face people playing slightly inferior lines, but not very common to face people whose entire repertoire is on Caruana's level.

Would be cool imo, but not sure if too much work.
Back to top Go down
Brendan




Posts : 398
Join date : 2020-11-18
Age : 40

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySun Nov 22, 2020 11:41 am

Admin wrote:
Admin wrote:
Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2     854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.

After book tuning +84 elo, 33 elo more.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx 620.0  620.0 / 1000 (62.00%)  75.2  
 2 ProDeo 2.2    380.0 xxxxx  380.0 / 1000 (38.00%)  75.2  

Suspect openings before:  http://rebel13.nl/text/result-1.txt
Suspect openings after:  http://rebel13.nl/text/result-2.txt

Download new book: http://rebel13.nl/dl/ProDeo-1.7z

Disadvantage, book is less varied.

+84 Elo! Nice.

Really like this book version. Even going through lines in analysis mode is cool looking at the info window and seeing that Stockfish NNUE played 13.Rc1 here recently and Carlsen played 13.b3 last year and so on.

Very nice work.
Back to top Go down
Mclane

Mclane


Posts : 3011
Join date : 2020-11-17
Age : 57
Location : United States of Europe, Germany, Ruhr area

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptySun Nov 22, 2020 1:21 pm

As I said before, the results of testing Version x+1 against Version x is not making much sense regarding ELO.
The ELO works only if the opponent is the old engine.
But in the moment other engines enter the battle, the ELO has no base anymore.

Therefore when I test I do usually test against many different engines.
Strong ones, weak ones, old and new ones. Stupid ones , intelligent ones.

The overall ELO is then the real one.

When programmers Mainly test against old versions of their own engine, the reason is often that their testing system can produce large amounts of games in a very fast time. They are used to testing this way. And they eliminated errors.
But the question still remains: is the „progress“ you measure similar if the new engine plays versus others.
And here I don’t think the Stockfish method works.

Back to top Go down
http://www.thorstenczub.de
supersharp77




Posts : 26
Join date : 2020-11-25

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyWed Nov 25, 2020 10:32 pm

Admin wrote:
Playing strength first, I collected a number of changes during the last 4 years resulting in a ~35 elo improvement. ~20 elo came from a new idea replacing IID with what I called IIR. A number of other engines (among them Stockfish) adopted the idea.

Polyglot opening books, improved the standard ProDeo opening book with +46 elo, see elsewhere. Created additional books with new options, an overview:

......

http://rebel13.nl/prodeo/prodeo-3.0.html

Well I went to the Rebel site and was going to download BUT...This engine was HUGE....1/2 Gig or some huge number....I would download just the engine only and the books and personalities..how come its so big?
Thx AR.... Smile pirat bounce
Back to top Go down
Admin
Admin
Admin


Posts : 2571
Join date : 2020-11-17
Location : Netherlands

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyThu Nov 26, 2020 1:48 am

supersharp77 wrote:
Admin wrote:
Playing strength first, I collected a number of changes during the last 4 years resulting in a ~35 elo improvement. ~20 elo came from a new idea replacing IID with what I called IIR. A number of other engines (among them Stockfish) adopted the idea.

Polyglot opening books, improved the standard ProDeo opening book with +46 elo, see elsewhere. Created additional books with new options, an overview:

......

http://rebel13.nl/prodeo/prodeo-3.0.html

Well I went to the Rebel site and was going to download BUT...This engine was HUGE....1/2 Gig or some huge number....I would download just the engine only and the books and personalities..how come its so big?
Thx AR.... Smile pirat bounce

It's called the "Book Edition". The focus is on opening books and some of them are big.

Download size is 290 Mb, is that really a problem nowadays?
Back to top Go down
http://rebel13.nl/
Peter Berger




Posts : 130
Join date : 2020-11-20

ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 EmptyThu Nov 26, 2020 12:41 pm

Admin wrote:
Admin wrote:
Tested with books on, +51 elo.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx1145.5 1145.5 / 2000 (57.28%)  77.6  
 2 ProDeo 2.2     854.5 xxxxx  854.5 / 2000 (42.72%)  77.6  

Expected more.

After book tuning +84 elo, 33 elo more.

Code:
No. Engine       1     2  Score  Games   Perc   Moves
--------------------------------------------------
 1 ProDeo 3.0    xxxxx 620.0  620.0 / 1000 (62.00%)  75.2  
 2 ProDeo 2.2    380.0 xxxxx  380.0 / 1000 (38.00%)  75.2  

Suspect openings before:  http://rebel13.nl/text/result-1.txt
Suspect openings after:  http://rebel13.nl/text/result-2.txt

Download new book: http://rebel13.nl/dl/ProDeo-1.7z

Disadvantage, book is less varied.

Have you made a new download for this one? lol - i should read more carefully Wink forget my post.
Back to top Go down
Sponsored content





ProDeo 3.0 Empty
PostSubject: Re: ProDeo 3.0   ProDeo 3.0 Empty

Back to top Go down
 
ProDeo 3.0
Back to top 
Page 1 of 2Go to page : 1, 2  Next
 Similar topics
-
» ProDeo Knowledge
» ProDeo 3.2 elo 3411 :lol!:
» ProDeo 3.1 - The PESTO version
» Book testing (tuning) with ProDeo
» My first post: A humble request for Mr. Ed Schröder and Prodeo

Permissions in this forum:You cannot reply to topics in this forum
ProDeo :: Computer Chess-
Jump to: