Author: blass uri
Date: 11:56:11 11/01/99
Go up one level in this thread
On November 01, 1999 at 12:11:40, Robert Hyatt wrote: >On November 01, 1999 at 10:09:44, blass uri wrote: > >>On November 01, 1999 at 09:17:45, Robert Hyatt wrote: >> >>> >>>Here is the way to find out. Log on to ICC and type "search crafty ban" >> >>I cannot do it because I am not registered >> >>> >>>Here is what I found. Out of the last 40 games, crafty won one on time, while >>>up material. Out of the last 80 games played, that one win on time was the >>>only game that 'search' showed. I assume he didn't lose on time _often_ based >>>on that, wouldn't you agree? >> >>It is interesting to see the games with the times per move. >> > > >Here is the first 25 moves of a recent game: > >Move crafty ban >---- ---------------- ---------------- > 1. e4 (0:01) e5 (0:08) > 2. Nf3 (0:00) Nc6 (0:03) > 3. Bb5 (0:00) Nd4 (0:03) > 4. Nxd4 (0:00) exd4 (0:03) > 5. O-O (0:00) Bc5 (0:04) > 6. Qg4 (0:00) Qf6 (0:14) > 7. d3 (0:00) c6 (0:08) > 8. Qh5 (0:17) d6 (0:04) > 9. Bg5 (0:08) Qg6 (0:07) > 10. Qxg6 (0:03) hxg6 (0:03) > 11. Bc4 (0:08) Kf8 (0:03) > 12. Nd2 (0:12) Rh5 (0:12) > 13. Bf4 (0:00) g5 (0:04) > 14. Bg3 (0:11) b5 (0:03) > 15. Bb3 (0:07) Bd7 (0:10) > 16. a4 (0:11) b4 (0:04) > 17. Bc4 (0:11) Rh6 (0:03) > 18. Nb3 (0:11) Re8 (0:09) > 19. e5 (0:11) Ne7 (0:10) > 20. exd6 (0:00) Bxd6 (0:03) > 21. Nxd4 (0:06) Nd5 (0:04) > 22. Bxd5 (0:10) cxd5 (0:12) > 23. Bxd6+ (0:00) Rxd6 (0:09) > 24. Rfe1 (0:01) Rxe1+ (0:13) > 25. Rxe1 (0:00) Bxa4 (0:08) > >Looks to me like he was out of book after 4 moves, crafty was out after 7. >IE he lost time on the first 3 book moves, then was thinking... > > > >>You can learn about the time managemant of Junior by looking in the time per >>move in the opening. >> >>I forgot that permanent brain does not help on the opening and if crafty is >>using 0 seconds per mopve in the opening when Junior is using 4 seconds per move >>then it is clearly a time advantage for Crafty. >> > >If the games were going 20 moves, perhaps. I just played a manual game as >'guest' to see how I did operating manually. I generally lose 1-2 seconds >being careful. And out of 5 moves, I generally get at least 1 in 0 seconds. >Just for measurement. > > > > > >> >> >>> >>>As far as time odds, your formula holds up for _any_ game on unequal hardware. >>>Which is well-known. If a machine is twice as fast, it is twice as fast. If >>>you want to call it 4.5x faster because it is twice as fast but time dilation >>>makes it look bigger, fine... >> >>I did not say that it is 4.5 to 1 but I believe that the fact that Junior is a >>manual program increase the time odd from 1600:700 to something clearly bigger. >> >>I have no idea if it is 3:1 or 4:1 or 4.5:1 > >It depends on the operator too. If the operator is a "parrot" and _only_ >does what the engine says, there is a definite time advantage of about 2 >seconds per move. If the operator exerts _any_ influence, then there is no >overall handicap at all, because the operator can force obvious recaptures >instantly without having to lose any time, and he can even overrule blunderful >moves. It all depends on the operator. It also depends also on the program. If the program knows to play some obvious recaptures fast then the fact that the operator knows to do it does not help much. In the case of ban it seems clear that the time advantage is more than 3 seconds per move. so it is not 2 seconds per move. I guess that it is usually 3 seconds per moves but it can be sometimes more than it Junior used 8 seconds for the first move. I guess that sometimes the operator has to do other things so I can understand advantage of about 4 seconds per move. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.