Author: Robert Hyatt
Date: 18:15:03 08/24/99
Go up one level in this thread
On August 24, 1999 at 17:06:45, Mark Young wrote: >On August 24, 1999 at 14:48:52, Robert Hyatt wrote: > >>On August 24, 1999 at 13:24:38, rich wrote: >> >>>On August 24, 1999 at 10:05:35, Robert Hyatt wrote: >>> >>>>On August 24, 1999 at 04:02:10, rich wrote: >>>> >>>>>First of all, how can you know for sure this person is a cheater? >>>>>Innocent until found guilty, anyone ?!? >>>> >>>>If you had read _my_ post, you would have noticed that I did look at the >>>>games carefully, and ran them thru my 'computer screening' methodology. And >>>>that I reported this person to the FICS admins. >>>> >>> >>>>He was _definitely_ cheating. >>> >>>> Could you explain this "computer screening methodology" ? >>> >>>> >> >> >> >>briefly I look at time/move for several games. computers are pretty >>consistent. Humans are not. I am particularly suspicious of a human >>that loses in a zero increment game where the last 20 moves each took >>3-4 seconds, while a _real_ player would speed up. But it is hard to >>speed up if you are transferring moves manually from an ICC interface to >>a chess program interface and back again. >> >>If the games 'match' then I run several 'blunder check' runs with crafty's >>annotate feature. I look for (a) no blunders and (b) moves that seem to lose >>material at shallow searches but which don't at much deeper levels. After >>those, I look for obvious computer moves that no human would consider. >> >> >> >>>> >>>> >>>>>Second, I can't find any evidence in any of the posts in this thread. >>>>>Third, maybe all of you just can't realize that some unrated players >>>>>are very good chessplayers.I have myself beaten a GM once(I'm 1761 club-rating) >>>>>and from time to time I win CM 6000 in long games(I'm a CM connaisseur). >>>>>Furthermore a friend of mine ( unrated, never participates in tournament of any >>>>>kind)regularly whoops my bottom at the chessboard.Every time I play him I >>>>>FEEL I am getting slowly, positionally grinded down by a 2400 player. >>>> >>>>Beating a GM one game out of how many? That is a _lot_ easier than beating >>>>Crafty 3 games in a row, something that I have _never_ seen any GM player do. >>>> >>>> A: Out of 1 game May I ask for statistics?(crafty vs. GM's , number of >>> games and wins losses draws) >>> >> >> >>go to ICC and type this: (where "GM" should be replaced by any GM handle >>that plays on ICC) >> >>search +=crafty GM >>search -=crafty GM >> >>here's a few: >> >>Crafty vs GM Sierawan 110 wins 52 losses 40 draws >>Crafty vs Roman 72 wins 16 losses 30 draws (this for only one roman >> handle, he has many) >>Crafty vs Kamsky 173 wins 38 losses 39 draws >>Crafty vs cptnbluebear 286 wins 49 losses 139 draws >>Crafty vs Dlugy 37 wins 12 losses 11 draws > > >Great results, and shows just how much the ratings are skewed on ICC and other >servers. Either Crafty is way underrated or some of the GM's and IM's are over >rated!! > >GM Dlugy better then Crafty, I don't think so.... Several things account for this... 'sub-pools' is one thing... Dlugy plays a lot of players that Crafty doesn't play. Crafty plays a _lot_ of games against other computers... I'd bet that if I disallowed computers, Crafty could pass its 3249 record and probably stay there most of the time. As it is likely 400 rating points better than most good GM players at blitz and bullet (the above stats are blitz only, ICC doesn't store bullet games). But Crafty clearly isn't 400 points better than other computers... even with the quad xeon hardware... so it's rating is going to depend on who it has played the most, recently. It peaked at 3249 at a time when all the real computer operators were gone to Paderborn... so it was mostly playing GM players all day long... > > Blitz Bullet >2837 OzsO(C) 3178 Mofongo(C) 3095 KillerGrob(C) >2702 TheComputer(C) 3105 A-Morozevi(GM) 2994 crafty(C) >2694 crafty(C) 3057 Goldmund(GM) 2926 scrappy(C) >2681 mundilfari(C) 3056 KingLoek(GM) 2882 Masterchess(C) >2673 RAINHA(C) 3053 EggSalad(C) 2879 garompon(GM) >2668 FritzX(C) 3036 MachineHead(C) 2875 ZarkovX(C) >2668 Good-Boy(C) 3002 stein(GM) 2867 Indiana-Jones >2654 SandmanJr(C) 2999 junior(GM) 2853 Hawkeye(IM) >2647 singacrafty(C) 2999 pap(GM) 2810 EggSalad(C) >2611 BountyHunte(C) 2997 Sammour-Hasbun 2802 nasenmann(IM) >2610 Netsurfer(C) 2989 Dlugy(GM) 2801 CraftyCraft(C) >2603 High-Comman(C) 2983 crafty(C) 2797 MachineHead(C) >2583 RETI(C) 2965 Good-Boy(C) 2791 bot(C) >2582 ET2000(C) 2964 Hawkeye(IM) 2776 counterplay(C) >2580 Nottobad(C) 2961 Varguz(C) 2755 Data(C) >2564 KillerGrob(C) 2959 ban(C) 2755 junior(GM) >2550 Steffie(C) 2955 ZarkovX(C) 2747 sicilliankn(C) >2544 MoonShot(C) 2949 bot(C) 2745 WchessX(C) >2533 bot(C) 2945 Tendulkar(GM) 2741 rude-agent(C) >2532 garompon(GM) 2934 chaozz(GM) 2741 Slopawn(C) >2530 attakinski(C) 2931 Guarapuava 2736 Varguz(C) >2530 scrappy(C) 2925 Gela(IM) > > > >> >>etc. >> >> >>>> >>>>>Actually it feels like playing Karpov or Petrosjan.Oh and by the way did I >>>>>mention that he(my friend) has played my CM 6000 with a record of 13 wins, >>>>>1 draw and 1 loss? >>>>> >>>> >>>> >>>>it is easy to do this in uncontrolled conditions. Watching the analysis >>>>window. Taking back moves. Slow computer. etc... He should try dropping >>>>over to ICC and demonstrating his skill against some of the very strong >>>>programs there where he has to play like he would in a real game. I'd bet >>>>he has _zero_ chances of beating _any_ program 13 times. Unless he plays a >>>>million games or two. >>>> >>>> The conditions were not uncontrolled.The games were played at my home, >>> without aids of any kind at G/35 >> >> >>I simply don't believe it. Have your friend log on to ICC. He should be >>able to beat Crafty just as easily 3 in a row... >> >> >> >> >>>> >>>> >>>>> So that's that.Thinking somebody is a cheater is not enough. >>>>>Hard evidence is required, IMHO >>>> >>>> >>>>I agree. But it is not hard to figure out if you go over the games with a >>>>careful eye and use a program to check for blunders (or the lack thereof). >>> >>> So if someone doesn't drop a piece in a blitz game against Crafty that person >>>is likely to have cheated.?!?? Get a grip.Could it be that you are simply >>>underestimating human tactical ability?I know some players that are tactical >>>genuises with ~1900 elo and no better because they are bad strategicallly. >> >>No... I check a lot more than that. And I "have a grip" thank you. If someone >>is a 'tactical genius' then their blitz rating will be way higher than normal, >>because blitz is much less positional and more about tactics. Bring 'em on... >>a demonstration is much better than 'claims'... I've only watched several >>_thousand_ human vs computer games... I have a pretty good idea what a human >>can do. And if a GM can't do it, no 'unrated player' is going to do it either.
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.