Author: martin fierz
Date: 23:31:58 04/12/02
Go up one level in this thread
On April 12, 2002 at 23:59:22, Joe McCarro wrote: >On April 12, 2002 at 21:58:25, martin fierz wrote: > >>On April 12, 2002 at 21:39:00, Joe McCarro wrote: >> >>>On April 12, 2002 at 20:44:38, martin fierz wrote: >>> >>>>On April 12, 2002 at 17:05:45, Mark Young wrote: >>>> >>>>>On April 12, 2002 at 14:28:13, martin fierz wrote: >>>>> >>>>>>>On April 12, 2002 at 03:43:16, Mark Young wrote: >>>>>>> >>>>>>>Sorry,Fritz 7 could be stronger,NPS means nothing between 2 programs >>>>>> >>>>>>NPS does not mean everything, but it does not mean nothing either. >>>>> >>>>>I did not say that it means nothing, I said NPS means nothing between 2 programs >>>>>that are not the same. >>>>> >>>>> >>>>> typically, >>>>>>NPS is a "design choice" of the programmer - smart but slow or less smart but >>>>>>faster. >>>>> >>>>>This is clearly wrong here, are you selling us that you can tell how smart a >>>>>program is by NPS? Your understanding is limited, if this is your bias. >>>> >>>>my understanding is quite all right, thank you :-) >>>>NPS is very much a function of how smart a program is. of course, if you write a >>>>program in basic, and get low NPS, this does not mean that your program is >>>>smart! i never said that. if you compare well-optimized engines (take some of >>>>the top commercials, fritz, junior, shredder, hiarcs, whatever), THEN NPS is a >>>>measure of how much work the program does on every node. in my checkers program, >>>>for example, i can turn on or off several features like enhanced transposition >>>>cutoffs, evaluation-based forward pruning and arbitrary parts of my move >>>>ordering. i can also add or remove parts of the evaluation function. everything >>>>i turn off makes my program faster in NPS, but it plays worse. the difference >>>>between smart and dumb is a factor 2 in NPS for my program. >>>>this is the design choice i was talking about, and it is the reason why people >>>>here always say "NPS means nothing". if you are talking about this kind of >>>>difference in NPS, it *does* mean nothing. talking about a factor 50, without >>>>the penalty of being dumb in return is another matter entirely. >>>> >>>> >>>>>How strong was Deep Blue 1997? How strong are todays micros? If you look at the >>>>>facts based on games played....It is clear your argument does not hold water. >>>>>What is Clear based on the data. Todays top micros are playing better then 2700 >>>>>elo chess. And playing very close if not better then DEEP BLUE of 1997. >>>> >>>>facts based on what? like rebel-van wely 2-2 or huebner-fritz 3-3? i don't think >>>>that these results are "very close" to defeating kasparov by 3.5-2.5! look at >>>>what kasparov does to these 2 guys: >>>>searching the big2000 database for kasparov - van wely gives me 4 games at >>>>standard time control - 4-0 for kasparov. in a match against huebner in '85 >>>>kasparov won 4.5-1.5. in matches at STC, where humans prepare and learn, i have >>>>yet to see "better than 2700"... i know no other serious results than the 2 >>>>above. >>> >>> >>>What about Gulko?? >> >>you should not compare apples and oranges! gulko had to play at the much faster >>G/60 time control which favors the computer. additionally, the match was against >>4 different programs, so he did not have the chance to repeat an opening (or at >>least try...) against the same opponent. >> >>cheers >> martin > >I agree this does not boad well for fritz. Just to clarify I think the games >were 60 plus 10 sec per move. Also I don't believe Glko had to play those time >controls but instead chose to. >I admit those are quibling points. Nonetheless I think there are other serious >problems with saying Clearly this shows Deeper blue was alot better than this >new fritz computer combo. yes, gulko chose the time control, but probably he just wanted to make as much money per hour as he could :-) i don't know about the 10s/move, you may be right, but the time control was definitely faster than in the deep blue match. >1) You are comparing oranges to apples when you say Kasparov plays the best >against humans therefore he plays the best against computers. It wouldn't >suprise me if we find many circumstances amoung human v. human players were >player A did well agains B then B did well against C yet c did well against A. >I often wonder if karpov in 1997 would have beat Deeper Blue. Just because he >seems to have a more positional approach whereas kasparov is better at tactics. right, also kramnik should be much better against the computer because he has an even drier approach to chess. >2) The fact that kasparov played Deeper blue cold I think means alot. I wonder >if Kasparov was given deeper blue for a month to play with or even had Deep >blue(the 1996 version) to play against before the match what would have >happened. (I'm assuming that both Huebner and Van wely had these programs or at >least a earlier program to practice against before they played them.) This must >give them a huge advantage. Many people say Kasparov seemed unnerved when >playing Deeper blue. Familiarity could have eliminated this. (Of course if >kaparov had deeper blue before the match and found he really couldn't beat it >there might never have been a match :) right again... >3)Although there is alot of talk about how deeper blue had specially designed >hardware and ran at 200m nps that same hardware design,(that appearantly adds >knowledge without losing speed etc.) this was also in Deep blue 1996. Deep Blue >1996 won the first game against Gary. Gary won the next. He drew the next 2. He >then won the next two back to back. It seemed very much like there was a wake >up call then the match was easilly under his control. How could this be with >this and that hardware design and 100m nps?? I don't think Gary thought Deeper >blue would be to tough of a challenge. I also think he tried to force a win in >the last game and mentally could not accept settling for a draw. > >Yes the nps doubled with Deeper blue but that didn't add that many moves to the >brute search. Accordingly we are stuck with extremely nonobjective statements >about the knowledge added to deeper blue. when I say non-object I mean >completely unkown!! What did they add in the year? It seems like some moose >lodge fraternal seceret! Everyone was appearantly sworn to secrecy to add to >the Deep blue 2 mystique. The only concrete thing I heard was some vague >statements about some opposite bishops endgame position. the Deep blue team was >well paid but they had alot to do with the whole knew hardware etc. The >programers today focus all their attention on the knowledge of their programs >they are not spending time designing new hardware. They have also been doing it >for alot longer than one year. right again.... >4) In sum I believe Kramnik will win convincingly. But I just don't think the >six games against Kasparov is a meaningful way to draw conclusions about how >strong Deep blue is and how it would compare to todays computers. Again when I >say what about Gulko?? I can take it a different way and say he had a winning >record agaisnt Kasparov in 1997. Does this show he was just as good as Deeper >blue? Of course not. It may seem absurd to many here to say he may have been >better than Deeper blue but why? What did Deeper Blue do that Gulko didn't??? >And if you say Deeper Blue played a match and not just random games and that >would somehow help Gary, I will remind you that Gary did not have any way to >prepare for the match against Deeper Blue. Given the small number of games from >DB that we have to go on I think making blanket statements about how it is >obviously better than the best comp opponents is just as misguided as saying the >comp opponents would clearly beat it. We just don't know. Joel Benjamin would >probably have the best guess. and again... it's just that this thread has been like "you don't know what you're talking about and micros are better than DB and over 2700", so i decided to answer with a similar statement taking the reverse position :-) i think it's clear that when you compare the only standard time control matches, micros are definitely not over 2700. the match aspect is important IMO, because humans can learn how to play against the machines. gulko claims he would have won the next 4 games... i just wanted to point out that DB did beat a 2800 player, and fritz drew a 2600 player, so claiming that fritz is better than DB is ridiculous IMO. even if you factor in that it was kasparov, and that he was not prepared as well as huebner might have been (how good would that have been? i don't know. gulko for instance was NOT prepared AT ALL). aloha martin
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.