FCP: TOP-8 tourney, 40 in 40 in LIVE mode

[Frank Quisinsky]

Hi there,

at the moment I don't have many time to redevelop my webpage. Furthermore, I am reading the book from Larry Kaufmann for an review I have interest to create.

But for the moment you can visit a TOP-8 tourney in LIVE mode if you have interest.

4x more time as in SWCR ...

URL:
http://www.amateurschach.de/ftptrigger/test-01.html

Code: Select all

Conditions:
#Time control: 40/40 "repeatedly"
#Game average: 160 minutes
# - 8 games running to the same time
# - 10 games per match, 10 engines = 450 games
# - Tournament time: January 27th, 2012 - February 03th, 2012 
#Resign: OFF
#Ponder: ON
#Learning: OFF
#Endgames: 4-pieces, 32Mb cache
#Opening books: OWN v5.11
#GUI: Shredder Classic 4
#OS: Windows XP Prof. x64 Edition
#Processors: Intel, 4xQ9550 2,83GHz
#Cores: 1 core for each engine
#Hash-Tables: 512Mb

Games are in around one week available for donwload.
Have fun with this little tourney.

Best
Frank

[skalde]

Nice tourney! I guess it will be hard for Chiron ...

[Frank Quisinsky]

Hi,

yes!
But it could be possible that Chiron 1.1a will be better with longer time controls. I am sure Chiron will play a good tourney. Each engine have 90 games to play. 25-30 points are possible because if I am looking in SWCR the results vs. stronger engines are really good (means from Chiron).



Will see Stockfish 2.2.2 JA PHQ ...

Best
Frank

[Frank Quisinsky]

After 122 of 450 games ...

With 40 moves in 40 minutes and a move average from 97 moves around 70 games can be play per day if 4 Quad cores (each engine with 1 core) are working.

Code: Select all

FCP, 40/40, ponder=on, Q9550   2012

                                1   2   3   4   5   6   7   8   9   0   
1   Rybka 4 x64 Exp. 42         *** ½½½ ½½1 ½½  001 1½  ½1  ½1½ 1½½ 11½  15.0/24
2   Stockfish 2.2.2 JA x64      ½½½ *** ½½0 ½½½ 10  01½ ½½0 11  ½1½ 11½  14.0/25
3   Critter 1.4 x64             ½½0 ½½1 *** ½1  ½0  ½½½ ½1½ ½½1 ½½  ½1½  13.5/24  160.25
4   Stockfish 2.2.2 JA x64 PHQ  ½½  ½½½ ½0  *** ½½½ ½11 ½0  ½½½ 1½½ 11½  13.5/24  154.75
5   Houdini 2.0c x64            110 01  ½1  ½½½ *** ½0½ 0½½ 10½ 1½  ½1½  13.0/24  155.25
6   Komodo 4.0 x64              0½  10½ ½½½ ½00 ½1½ *** 1½½ ½½½ 1½½ 1½   13.0/25  152.00
7   RobboLito 0.10 x64          ½0  ½½1 ½0½ ½1  1½½ 0½½ *** 1½0 ½½½ ½½   12.0/24
8   IvanHoe 999946hm x64        ½0½ 00  ½½0 ½½½ 01½ ½½½ 0½1 *** ½½1 ½1   11.5/25
9   Chiron 1.1a x64             0½½ ½0½ ½½  0½½ 0½  0½½ ½½½ ½½0 *** ½0½   9.0/25
10  Rybka 4.1 x64               00½ 00½ ½0½ 00½ ½0½ 0½  ½½  ½0  ½1½ ***   7.5/24

Move average = 97
Tourney is without resign mode.
Move average in SWCR without resign mode after 161.160 games = 86 moves.

1:0 = 28 games = 23,0%
Draw = 76 games = 62,2%
0:1 = 18 games = 14,8%

Lost on time games = 1 (1x Chiron).
Game is replayed!

Best
Frank

[Frank Quisinsky]

If I compare the results with SWCR (only the actual participant engines):

With 4x more time as in SWCR 6% more remis games at the moment. Move average go higher ...
89 moves in SWCR with 40 in 10 (actual engines only)
96 moves with 40 in 40

In the latest weeks I made some experiments with blitz tourneys 40 in 2 games (around the same group of engines)

Move average = 84 !!
remis quote 49% !!

After this stats it should be clear that with more time completly other results are building.

In my opinion it make more sense to search an other way for testing engines as to play eng-eng with different time controls. It make more sense to search weeknesses and strengths by all the engines.

I believe that a rating list with time control x and hardware x is only interesting for people which used the same conditions. To search for an engine ELO x is only interesting if we like to know this one:

Group 1: Engines 2900+
Group 2: Engines 2800+
Group 3: Engines 2700+

An exact ELO for an engine isn't possible with 100 or 10.000 games. What we can do is this one ...

Rating with 40 in 2 =
Rating with 40 in 5 =
Rating with 40 in 10 =
Rating with 40 in 20 =
Rating with 40 in 40 =

And now to build a average result.
Not possible for one person or a normaly group of persons. Only if 100 or more people work on such a project.

Best
Frank

[mcostalba]

Will see Stockfish 2.2.2 JA PHQ ...

HI Frank,

I have incorporated part of your PHQ setting (namely increased mobility) in the 2.2.2 so if you have rised again mobility it is like you have done it twice in respect to 2.1, I am not sure it is good.

Marco

[Frank Quisinsky]

Hi Marco,

at first, the new version is great.
I played 1.600 SWCR games with SWCR conditions in secret mission
SWCR is closed, used it only for myself.

+ 36 to version 2.1.1 default (I used the default settings).
My congratulation, 2.886 SWCR ELO and rating place two !!

To the PHQ setting ...
But what I have in detail to change in PHQ setting now?
I will test a bit with the information you gave next weeks.

If you like, again ...
Stockfish is the engine I like the most because the playing style is that what I like. SWCR ended and if you like I can test beta versions with my Q9550 systems.

Best and good luck ...
SF is one the way!

Best
Frank

[Frank Quisinsky]

Hi Marco,

I checked the PHQ from version 2.1.1 for version 2.2.2. You have absolutley right, after different test position it seems to be clear that this idea was bad.

I changed the engine in the still running tourney.
From Stockfish 2.2.2 JA x64 PHQ to Stockfish 2.1.1 JA x64 PHQ.
This is much more interesting

Thanks again!

Furthermore, I have one lost on time game (after 39 moves) with Stockfish 2.2.2 JA x64 and two to fast remis games. This games will be replayed too. Fast remis games under 20 moves I replayed in my tourneys.

With the result:
WirrWarr in the tournament table for around 12 hours
Not very important.

Best
Frank

[Frank Quisinsky]

Hi there,

I added a replay zone for Test-Tourneys on my webpage. For the moment you can find the first games from my TOP-8 tourney. Will be updated from time to time.

TOP-8 tourney:
http://www.amateurschach.de/ftptrigger/test-01.html

Replay-Zone:
http://www.amateurschach.de/test-tourne ... est-01.htm

Best
Frank

[Frank Quisinsky]

Code: Select all

FCP, 40/40, ponder=on, Q9550   2012

                                1     2     3     4     5     6     7     8     9     0     
1   Komodo 4.0 x64              ***** 0½½11 ½1½½1 ½½½11 101½1 ½01½0 1½½1½ ½½½1½ 1½½01 1½½½1  28.5/46
2   Rybka 4 x64 Exp. 42         1½½00 ***** 001½0 ½½111 ½½½½0 ½½½½0 ½1½11 ½1½01 11½11 1½½1½  27.5/46
3   Houdini 2.0c x64            ½0½½0 110½1 ***** ½1½10 01½1½ ½01½½ 0½½1½ 10½11 ½1½½½ 1½011  27.0/46
4   Critter 1.4 x64             ½½½00 ½½000 ½0½01 ***** ½½111 ½11½1 ½1½1½ ½1½1½ ½1½0½ ½½½½½  25.5/46
5   Stockfish 2.2.2 JA x64      010½0 ½½½½1 10½0½ ½½000 ***** 1½½½1 ½½010 111½0 11½½½ ½1½11  24.0/46
6   Stockfish 2.1.1 JA x64 PHQ  ½10½1 ½½½½1 ½10½½ ½00½0 0½½½0 ***** 01½1½ 00½½½ ½1½1½ 01½1½  22.5/46
7   RobboLito 0.10 x64          0½½0½ ½0½00 1½½0½ ½0½0½ ½½101 10½0½ ***** 1½0½0 ½½½½½ ½½½1½  20.5/47  468.25
8   IvanHoe 999946hm x64        ½½½0½ ½0½10 01½00 ½0½0½ 000½1 11½½½ 0½1½1 ***** ½10½0 ½½1½½  20.5/47  463.25
9   Rybka 4.1 x64               0½½10 00½00 ½0½½½ ½0½1½ 00½½½ ½0½0½ ½½½½½ ½01½1 ***** ½1½½1  19.0/46
10  Chiron 1.1a x64             0½½½0 0½½0½ 0½100 ½½½½½ ½0½00 10½0½ ½½½0½ ½½0½½ ½0½½0 *****  16.0/46

Stats:
Lost on time games: 9 of 231 = 3,9%
3x Stockfish 2.1.1 JA x64 PHQ
2x Stockfish 2.2.2 JA x64
3x Chiron 1.1a x64
1x IvanHoe 999946hm x64

All games are replayed!

Move average without resgin mode = 98

With ELOstat 1.1 ...

Code: Select all

    Program                            Score     %    Av.Op.  Elo    +   -    Draws

  1 Komodo 4.0 x64                 :  28.5/ 46  62.0   2941   3026   73  71   50.0 %
  2 Rybka 4 x64 Exp. 42            :  27.5/ 46  59.8   2943   3012   76  75   45.7 %
  3 Houdini 2.0c x64               :  27.0/ 46  58.7   2943   3004   78  76   43.5 %
  4 Critter 1.4 x64                :  25.5/ 46  55.4   2947   2985   69  68   54.3 %
  5 Stockfish 2.2.2 JA x64         :  24.0/ 46  52.2   2949   2965   77  77   43.5 %
  6 Stockfish 2.1.1 JA x64 PHQ     :  22.5/ 46  48.9   2949   2942   69  69   54.3 %
  7 IvanHoe 999946hm x64           :  20.5/ 47  43.6   2955   2910   72  72   48.9 %
  8 RobboLito 0.10 x64             :  20.5/ 47  43.6   2955   2911   65  66   57.4 %
  9 Rybka 4.1 x64                  :  19.0/ 46  41.3   2957   2896   66  68   56.5 %
 10 Chiron 1.1a x64                :  16.0/ 46  34.8   2961   2852   58  65   60.9 %

Shredder Opening Book v.5.11
available in my download area, tested with around 100.000 SWCR games.

Code: Select all

Games        :    231 (finished)

White Wins   :     65 (28.1 %)
Black Wins   :     47 (20.3 %)
Draws        :    119 (51.5 %)
Unfinished   :      0

White Perf.  : 53.9 %
Black Perf.  : 46.1 %

ECO A =     62 Games (26.8 %)
ECO B =     68 Games (29.4 %)
ECO C =     34 Games (14.7 %)
ECO D =     42 Games (18.2 %)
ECO E =     25 Games (10.8 %)

Hints:
Stockfish 2.2.2 JA x64 PHQ replaced with Stockfish 2.1.1 JA x64 PHQ
Games can be found on my webpage in Replay-Zone