3-second flash - 6,000 games

[PawnStormZ]

               Hi All.

       Below are the results of 6,000 games at a time control of game in 3 seconds, plus 1/10 second per move.    Players were the 64-bit versions of Critter 1.0, Houdini 1.5, Ivanhoe B47cB, Rybka 4.1, and Stockfish 2.1

       Match conditions: Intel core I7 920 at 2.66GHz using 4 cores HT off; Win 7 64 Home Premimum; 6G RAM; pc doing nothing but the match while a game is being played.

       Each engine uses default settings except for the split-depth of Houdini which is set to 12 (tested best on this pc); no large pages; no end-game tablebases; 512 hash each; 4 cores each; ponder off.

       I used the excellent "ChessGUI" to run the match, set to give each engine a 1/10th second "overstep" margin to help avoid any time losses.

       The openings are played using random choices from a set of epd positions that was downloaded from this forum called 8moves.epd.    Each position is played as Black and White by each engine against all the others.  Each engine played 600 games against each of the others.

       Houdini is still untouchable in super-fast games; only Rybka was able to win more than 100 games in their "head-to-head" match-ups while Critter and Ivanhoe did not even manage 60!    Stockfish did better against Houdini and Ivanhoe than Critter, yet lost its match-up against Critter.
       
   
       Engine                    Pts                Houdini_1.5                    Rybka 4.1                    Ivanhoe B47cB            Stockfish 2.1                Critter 1.0

1 Houdini_1.5             1,723.5          **************        + 249 = 191 - 160        + 335 = 212 - 53        + 387 = 138 - 75        + 410 = 144 - 46

2 Rybka 4.1                1,526.0          + 160 = 191 - 249        **************         + 274 = 231 - 95        + 340 = 187 - 73        + 368 = 159 - 73

3 Ivanhoe B47cB        1,006.5            + 53 = 212 - 335          + 95 = 231 - 274        **************       + 206 = 211 - 183      + 210 = 231 - 159

4 Stockfish 2.1               888.5            + 75 = 138 - 387          + 73 = 187 - 340        + 183 = 211 - 206      **************       + 179 = 221 - 200

5 Critter 1.0                   855.5            + 46 = 144 - 410          + 73 = 159 - 368        + 159 = 231 - 210      + 200 = 221 - 179        **************
   
                   PGN-file here... http://www.datafilehost.com/download-6872b0ee.html
 

[Engin]

who cares about 3 seconds per games ?

even if you let playing 1 million games, such nonsense games are didnt interesting me.

[Dann Corbit]

who cares about 3 seconds per games ?

even if you let playing 1 million games, such nonsense games are didnt interesting me.

It might be interesting for programmers who want to tune parameters for blitz games.

[Engin]

sorry, but not for me.

some engines are need to start some seconds, and they losed on time before starting ?!

maybe i have nothing if you give them minimum at 20 s + 0.1 s / game

[Laskos]

Pal, something is wrong. Rybka cannot have + 274 = 231 - 95 against latest IvanHoe. Did you check NPS, time used, losses on time or illegal moves?

Kai

[Dann Corbit]

Pal, something is wrong. Rybka cannot have + 274 = 231 - 95 against latest IvanHoe. Did you check NPS, time used, losses on time or illegal moves?

Kai

I guess that at least 1/3 of the losses are losses on time.

[Dr.Wael Deeb]

who cares about 3 seconds per games ?

even if you let playing 1 million games, such nonsense games are didnt interesting me.

I agree,a pure sample of nonsense....

[Dr.Wael Deeb]

Pal, something is wrong. Rybka cannot have + 274 = 231 - 95 against latest IvanHoe. Did you check NPS, time used, losses on time or illegal moves?

Kai

Kai,aren't you aware by now that the setup of tournaments run by this guy are so erratic that I ignore them by a blink of the eye....

[Dr.Wael Deeb]

sorry, but not for me.

some engines are need to start some seconds, and they losed on time before starting ?!

maybe i have nothing if you give them minimum at 20 s + 0.1 s / game

Exactly....but how to explain to this guy this simple fact

[Ferdy]

who cares about 3 seconds per games ?

even if you let playing 1 million games, such nonsense games are didnt interesting me.

Ability of a program to see the correct move or close to a correct move in a small amount of time does matter. Strong programs are usually good in short (less than 10 sec) TC.