Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Houdini20c - Houdini2Baracuda 1000 games started
- - By mare1975 (**) Date 2012-07-07 07:12
Parent - - By Pia (****) Date 2012-07-07 09:12
Please confirm

Houdini 2.0 baracuda values
5F6E82747A3021212220222020201E1E1E1E20202020202222202222221E1E2328161C1E20
Contempt set to 0
Parent - - By mare1975 (**) Date 2012-07-07 09:34
Confirmed!:lol:

Regards
Parent - - By mare1975 (**) Date 2012-07-07 11:36
Parent - - By mare1975 (**) Date 2012-07-08 07:51


So far I m satisfied with performance of Baracuda version. Original version is not kicking a... hm.... behind   :lol:  of my tuned version...

When the test reach 200 games I will change value of N and Q.. Talking to my friend, and I decide so, because there is too big value difference between N and B so I will increase value of N to 317. Q value will be decreased to 965.

Regards...
Parent - By mare1975 (**) Date 2012-07-08 15:32


Value of pieces changed as I said in previous post

Regards
Parent - - By Garvin Gray (****) Date 2012-07-08 18:09

> Please confirm


>Houdini 2.0 baracuda values: 5F6E82747A3021212220222020201E1E1E1E20202020202222202222221E1E2328161C1E20: Contempt set to 0


What does this actually mean? How would I change my houdini setting to match this and tune the engine?
Parent - - By keoki010 (Silver) Date 2012-07-08 18:35
You need to use the program HoudiniConfig to change those values.
Parent - - By mare1975 (**) Date 2012-07-08 21:02
Parent - - By mare1975 (**) Date 2012-07-09 09:13
Parent - - By mare1975 (**) Date 2012-07-09 10:56
H20c Pro - H20c Pro Baracuda 6CPU test

284 games, 1min per engine

Games
Parent - - By mare1975 (**) Date 2012-07-09 16:10


And some screenshots





Parent - By mare1975 (**) Date 2012-07-10 05:55
- - By Garvin Gray (****) Date 2012-07-10 07:13
Seems like these results are within the margin of error, so barracuda is not much of an improvement, if any at all, at this stage.
Parent - By mare1975 (**) Date 2012-07-10 12:33
Agree... But, still,I will run all 1000 games...

Regards
Parent - - By Lukas Cimiotti (Bronze) Date 2012-07-10 13:06
True. The question is: do these settings have any influence on Houdini's playing strength?
Parent - - By mare1975 (**) Date 2012-07-10 13:51
Well, seems not big influence at this time control... Just for fun run 10 min per engine game and I think I will test that time control ( maybe even longer time control )... Interesting game:

[Event "10min, Blitz 10m"]
[Site "M"]
[Date "2012.07.10"]
[Round "1"]
[White "Houdini 2.0c w32"]
[Black "Houdini 2.0c w32Baracuda"]
[Result "0-1"]
[ECO "D37"]
[Annotator "0.21;0.16"]
[PlyCount "108"]
[TimeControl "600"]

{Intel(R) Core(TM) i3 CPU       M 370  @ 2.40GHz 2397 MHz  W=18.0 plies; 2.
181kN/s  B=18.5 plies; 2.199kN/s} 1. d4 {[%eval 21,19] [%emt 0:00:13] Both
last book move} Nf6 {[%eval 16,19] [%emt 0:00:15]} 2. Nf3 {[%eval 24,19] [%emt
0:00:21] (Nc3)} d5 {[%eval 11,20] [%emt 0:00:18]} 3. c4 {[%eval 29,18] [%emt 0:
00:12] (Bf4)} e6 {[%eval 16,18] [%emt 0:00:09]} 4. Nc3 {[%eval 21,18] [%emt 0:
00:09]} Be7 {[%eval 11,18] [%emt 0:00:39]} 5. Bf4 {[%eval 30,18] [%emt 0:00:14]
(e3)} O-O {[%eval 24,18] [%emt 0:00:08] (Nc6)} 6. e3 {[%eval 34,19] [%emt 0:00:
14]} Nc6 {[%eval 16,18] [%emt 0:00:14]} 7. a3 {[%eval 32,19] [%emt 0:00:27]
(h3)} Nh5 {[%eval 17,19] [%emt 0:00:08]} 8. Be5 {[%eval 35,19] [%emt 0:00:11]}
g6 {[%eval 19,18] [%emt 0:00:12]} 9. Rc1 {[%eval 36,19] [%emt 0:00:14] (Be2)}
a6 {[%eval 16,19] [%emt 0:00:10]} 10. h3 {[%eval 31,19] [%emt 0:00:09]} Nxe5 {
[%eval 19,19] [%emt 0:00:09] (dxc4)} 11. Nxe5 {[%eval 36,19] [%emt 0:00:10]}
Nf6 {[%eval 18,19] [%emt 0:00:07]} 12. Be2 {[%eval 28,20] [%emt 0:00:29]} c6 {
[%eval 18,20] [%emt 0:00:17] (c5)} 13. O-O {[%eval 29,19] [%emt 0:00:11]} Nd7 {
[%eval 12,19] [%emt 0:00:13]} 14. Nxd7 {[%eval 35,19] [%emt 0:00:08] (Nf3)}
Bxd7 {[%eval 18,20] [%emt 0:00:07] (Qxd7)} 15. c5 {[%eval 29,20] [%emt 0:00:06]
} Qc7 {[%eval 16,20] [%emt 0:00:10]} 16. f4 {[%eval 22,19] [%emt 0:00:27]} Rae8
{[%eval 16,19] [%emt 0:00:07] (f6)} 17. Qd3 {[%eval 29,18] [%emt 0:00:06] (Qd2)
} f6 {[%eval 18,17] [%emt 0:00:08]} 18. e4 {[%eval 23,17] [%emt 0:00:09]} Kg7 {
[%eval 17,17] [%emt 0:00:09] (dxe4)} 19. Qe3 {[%eval 39,17] [%emt 0:00:11]} Kh8
{[%eval 22,16] [%emt 0:00:15] (Bc8)} 20. Rcd1 {[%eval 40,17] [%emt 0:00:06]}
Rg8 {[%eval 13,18] [%emt 0:00:06] (a5)} 21. Na4 {[%eval 27,17] [%emt 0:00:07]}
f5 {[%eval 14,19] [%emt 0:00:06] (Rg7)} 22. e5 {[%eval 25,19] [%emt 0:00:07]}
Ref8 {[%eval 7,19] [%emt 0:00:07] (g5)} 23. Qb3 {[%eval 17,19] [%emt 0:00:12]
(Kh1)} g5 {[%eval -2,20] [%emt 0:00:09]} 24. Qb6 {[%eval 18,20] [%emt 0:00:05]}
Qb8 {[%eval -2,19] [%emt 0:00:07] (Rc8)} 25. fxg5 {[%eval 11,18] [%emt 0:00:05]
(Rd3)} Bxg5 {[%eval -16,19] [%emt 0:00:11]} 26. Kh1 {[%eval 3,18] [%emt 0:00:
09] (Rd3)} Be3 {[%eval -29,18] [%emt 0:00:14] (Be8)} 27. Qb3 {[%eval -14,16]
[%emt 0:00:04]} f4 {[%eval -29,16] [%emt 0:00:00]} 28. Nc3 {[%eval -72,18]
[%emt 0:00:47]} Qd8 {[%eval -138,18] [%emt 0:00:36] (Be8)} 29. Bf3 {[%eval -99,
16] [%emt 0:00:10] (Rd3)} Qh4 {[%eval -164,16] [%emt 0:00:00]} 30. Ne2 {[%eval
-149,17] [%emt 0:00:13]} Be8 {[%eval -172,18] [%emt 0:00:05]} 31. Rd3 {[%eval
-171,18] [%emt 0:00:04]} Qg5 {[%eval -172,16] [%emt 0:00:00]} 32. Nc3 {[%eval
-240,19] [%emt 0:00:35]} Bh5 {[%eval -226,17] [%emt 0:00:06]} 33. Qc2 {[%eval
-245,19] [%emt 0:00:04] (Qd1)} Bg6 {[%eval -227,20] [%emt 0:00:05]} 34. Qe2 {
[%eval -338,20] [%emt 0:00:29]} Bxd3 {[%eval -259,20] [%emt 0:00:06]} 35. Qxd3
{[%eval -224,19] [%emt 0:00:00]} Qh4 {[%eval -245,20] [%emt 0:00:04]} 36. Ne2 {
[%eval -247,20] [%emt 0:00:12] (Rd1)} Rg7 {[%eval -298,20] [%emt 0:00:13]} 37.
Qb3 {[%eval -247,19] [%emt 0:00:02] (Qc3)} Rfg8 {[%eval -342,17] [%emt 0:00:08]
} 38. Kh2 {[%eval -355,17] [%emt 0:00:22] (Qd1)} Qe7 {[%eval -344,15] [%emt 0:
00:00]} 39. Qc3 {[%eval -364,16] [%emt 0:00:09] (Kh1)} Rg6 {[%eval -359,19]
[%emt 0:00:04]} 40. Qe1 {[%eval -367,17] [%emt 0:00:02]} Qg7 {[%eval -359,21]
[%emt 0:00:05]} 41. Rg1 {[%eval -370,18] [%emt 0:00:01]} Bxg1+ {[%eval -359,21]
[%emt 0:00:04]} 42. Qxg1 {[%eval -352,16] [%emt 0:00:00]} Rh6 {[%eval -339,22]
[%emt 0:00:05]} 43. Kh1 {[%eval -350,19] [%emt 0:00:02] (Qe1)} Qg5 {[%eval
-343,21] [%emt 0:00:06]} 44. Qc1 {[%eval -355,19] [%emt 0:00:07]} Rh4 {[%eval
-347,21] [%emt 0:00:04] (Rf8)} 45. Qd1 {[%eval -360,19] [%emt 0:00:02] (Qe1)}
Rg7 {[%eval -347,20] [%emt 0:00:03]} 46. Qe1 {[%eval -360,20] [%emt 0:00:02]
(Kh2)} Kg8 {[%eval -347,20] [%emt 0:00:04]} 47. Qa5 {[%eval -359,19] [%emt 0:
00:02]} Kf8 {[%eval -350,19] [%emt 0:00:06]} 48. Kh2 {[%eval -359,18] [%emt 0:
00:02] (a4)} Rxh3+ {[%eval -1356,19] [%emt 0:00:18] (Rd7)} 49. gxh3 {[%eval
-253,14] [%emt 0:00:01]} Qh4 {[%eval -1298,18] [%emt 0:00:00]} 50. Qc3 {[%eval
-242,13] [%emt 0:00:00] (Qd8+)} Qf2+ {[%eval -1386,18] [%emt 0:00:01]} 51. Kh1
{[%eval -242,12] [%emt 0:00:00]} h5 {[%eval -1350,17] [%emt 0:00:00] (h6)} 52.
b3 {[%eval -755,17] [%emt 0:00:01] (Qd3)} Qf1+ {[%eval -1407,16] [%emt 0:00:00]
(a5)} 53. Kh2 {[%eval -1380,16] [%emt 0:00:00]} h4 {[%eval -1354,15] [%emt 0:
00:00]} 54. Bg4 {[%eval -1792,19] [%emt 0:00:03]} Rxg4 {[%eval -1352,14] [%emt
0:00:00]} 0-1

Screenshot

Parent - By mare1975 (**) Date 2012-07-11 09:23


- - By Garvin Gray (****) Date 2012-07-11 10:28
This is the hardest part of testing engines for meaningful results for corro play. The time control has to be long enough to have any chance of being representative of evals for corro analysis, but allows for enough games to be played to get some trending information at the very least.
Parent - - By mare1975 (**) Date 2012-07-11 10:57
Agree... But any test is better then no test; agree?:lol:   Anyway , this test is better then ultra fast (bullet) games, which shows almost nothing....

Regards...
Parent - - By mare1975 (**) Date 2012-07-11 14:04
.
-
H20c - H20Baracuda
.
6 CPU, 3 min per engine

.
.
Parent - By RFK (Gold) Date 2012-07-11 14:07
no thanks!:wink:
Parent - - By Uly (Gold) Date 2012-07-12 00:19

> This is the hardest part of testing engines for meaningful results for corro play.


Not really, you don't need to play any games to know what engine is better for the current position, just play the moves of the engines into one another and see which one refutes the moves of the other :smile:
Parent - - By mare1975 (**) Date 2012-07-12 07:28
People, I just like chess and enjoy it .... Don t need to be philosophic here... 

Regards

Parent - - By mare1975 (**) Date 2012-07-13 13:40
Parent - - By Barnard (Bronze) Date 2012-07-13 13:49
my friend,are you noticing that when the end of the match arrives,you % gets closer and closer?

i think you should to a few more changes,and increase the agressivity of the engine...also maybe,the B,Q,R and N pieces

and if you are interested to experiment,play with the settings of the 'doubled' and 'isolated' pawns

regards
Parent - - By mare1975 (**) Date 2012-07-13 14:06
Thanks for the tips,friend... I will run this test till end with values as it is....

Also I will run test vs T3 version- also with these settings...

After that I will try to improve Baracuda version...

Regards
Parent - - By Testo (**) Date 2012-07-14 06:19
Houdini 2.0cPro x64 - Baracuda vs Houdini 2.0cPro x64

I will test the new Baracuda version with this settings:

5F758274733021212220222020201E1E1E1E20202020202222202222221E1E2328161C1E20
Con set to 0

Conditions:
Processors: Intel Core i7 980X
Operating System: Win7 Professional x64
Hyper Threading: Disabled
Chess Gui: Fritz 12
Time Control: 3 Minutes/Game
Hushtable Size: 512 MB
Large Page : Off
Permant Brain: Off
Endgame: Gaviota TB
TB Cache: 32MB
Opening Book: Perfect2012b

---------------------------------------------

after 343 games

Parent - - By mare1975 (**) Date 2012-07-14 07:36
Thanks G! :lol:
Parent - - By mare1975 (**) Date 2012-07-14 07:40
Parent - - By mare1975 (**) Date 2012-07-14 17:37
Great result for Baracuda version! 234 - 216

6 CPU 450 games pgn
..
Parent - - By RFK (Gold) Date 2012-07-14 17:48
And I thought I was the only one that talked to himself on this forum! :lol:
Parent - - By mare1975 (**) Date 2012-07-14 18:00
Now You see You are not the only one...:lol::lol:
Parent - - By RFK (Gold) Date 2012-07-14 18:02
I used to get shit for it all the time! :yell:
Parent - - By mare1975 (**) Date 2012-07-15 06:20
Parent - By mare1975 (**) Date 2012-07-15 21:32
Up Topic The Rybka Lounge / Computer Chess / Houdini20c - Houdini2Baracuda 1000 games started

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill