Originally posted by EmLaskerEdit: On second thought I'm going to take this as a joke. Tell me you're not that stupid...
something just looks suspicious to me, it seems like the difference between every single game that they play is decided by a pawn. Somethings not right...
Very accurate play from both.
Here's what Fritz says:
Fritz 11 @ 30 seconds per move
Pentium 4 2.93GHz 1GB RAM
Hash Table 192MB
Database used www.chesslive.de
[Event "Challenge"]
[Site "http://www.playtheimmortalgame.com"]
[Date "2006.08.27"]
[EndDate "2007.07.16"]
[Round "?"]
[White "Weyerstrass"]
[Black "Akizy"]
[WhiteRating "2451"]
[BlackRating "2455"]
[WhiteELO "2451"]
[BlackELO "2455"]
[Result "1-0"]
[GameId "2423892"]
1. c4 Nf6 2. g3 c6 3. Nf3 d5 4. Bg2 Bf5 5. cxd5 cxd5 6. Qb3 Nc6 7. Qxb7 Bd7 8.
Qb3 e5 9. d3 Rb8 10. Qd1 Bb4+ 11. Nfd2 {Takes game out of book; 2nd choice} Qc8
{Not in top 3} 12. O-O {1st choice} Nd4 {1st choice} 13. Nc3 {1st choice} Bg4 {
2nd choice} 14. Nf3 {1st choice} Bxf3 {1st choice} 15. exf3 {1st choice} Bxc3 {
1st choice} 16. bxc3 {1st choice} Qxc3 {1st choice} 17. Qa4+ {1st choice} Qc6 {
1st choice} 18. Qxc6+ {1st choice} Nxc6 {1st choice} 19. f4 {1st choice} e4 {
2nd choice} 20. Ba3 {2nd choice} exd3 {1st choice} 21. Rfd1 {1st choice} Rb5 {
3rd choice} 22. Rxd3 {2nd choice} Kd7 {1st choice} 23. Rc1 {2nd choice} Rhb8 {
1st choice} 24. Rdc3 {1st choice} R5b6 {1st choice} 25. Bh3+ {Not in top 3} Kc7
{1st choice} 26. R3c2 {1st choice} a6 {Not in top 3} 27. f3 {3rd choice} a5 {
1st choice} 28. Kf2 {1st choice} h5 {2nd choice} 29. Bf1 {2nd choice} Kd7 {
2nd choice} 30. h4 {1st choice} Nb4 {Not in top 3} 31. Rc7+ {1st choice} Kd6 {
3rd choice} 32. Rxf7 {1st choice} R8b7 {1st choice} 33. Rxb7 {2nd choice} Rxb7
{1st choice} 34. Ke3 {1st choice} Nd7 {Not in top 3} 35. Bd3 {1st choice} Nc5 {
1st choice} 36. Bg6 {1st choice} Re7+ {1st choice} 37. Kd2 {1st choice} Na4 {
1st choice} 38. g4 {Not in top 3} d4 {3rd choice} 39. Bxh5 {Not in top 3} 1-0
Game 2423892
Result:
White: Weyerstrass
Top 1 Match: 19/29 (65,5% )
Top 2 Match: 25/29 (86,2% )
Top 3 Match: 26/29 (89,7% )
Black: Akizy
Top 1 Match: 17/28 (60,7% )
Top 2 Match: 21/28 (75,0% )
Top 3 Match: 24/28 (85,7% )
[Event "Challenge"]
[Site "http://www.playtheimmortalgame.com"]
[Date "2006.08.27"]
[EndDate "2007.07.13"]
[Round "?"]
[White "Akizy"]
[Black "Weyerstrass"]
[WhiteRating "2455"]
[BlackRating "2451"]
[WhiteELO "2455"]
[BlackELO "2451"]
[Result "1/2-1/2"]
[GameId "2423887"]
1. e4 e6 2. d4 d5 3. Nc3 Bb4 4. e5 Ne7 5. a3 Bxc3+ 6. bxc3 c5 7. Qg4 O-O 8. Bd3
f5 9. exf6 Rxf6 10. Bg5 Rf7 11. Qh5 g6 12. Qd1 Qa5 13. Bd2 Nbc6 14. Nf3 Qc7 15.
dxc5 e5 16. Ng5 Rf8 17. c4 e4 18. cxd5 exd3 19. d6 Qd8 {
Takes game out of book; 2nd choice} 20. dxe7 {3rd choice} Qxe7+ {1st choice}
21. Be3 {1st choice} Rf5 {2nd choice} 22. Nf3 {2nd choice} dxc2 {1st choice}
23. Qxc2 {1st choice} Be6 {1st choice} 24. Nd2 {Not in top 3} Rd8 {Not in top 3
} 25. O-O {1st choice} Nd4 {2nd choice} 26. Qe4 {3rd choice} Ne2+ {1st choice}
27. Kh1 {Only move} Nc3 {1st choice} 28. Qc2 {1st choice} Nd5 {1st choice} 29.
Rac1 {1st choice} Nxe3 {1st choice} 30. fxe3 {1st choice} Rxf1+ {2nd choice}
31. Nxf1 {2nd choice} Bd5 {1st choice} 32. Qc3 {2nd choice} Qg7 {2nd choice}
33. Kg1 {Not in top 3} Qxc3 {1st choice} 34. Rxc3 {1st choice} Bc6 {1st choice}
35. g4 {1st choice} Rd1 {1st choice} 36. Kf2 {1st choice} Bb5 {1st choice} 37.
Ng3 {1st choice} Rd2+ {1st choice} 38. Kf3 {1st choice} Rxh2 {2nd choice} 39.
Ne4 {1st choice} Bc6 {1st choice} 40. Kg3 {1st choice} Re2 {2nd choice} 41. Kf4
{2nd choice} 1/2-1/2
Game 2423887
Result:
White: Akizy
Top 1 Match: 13/22 (59,1% )
Top 2 Match: 17/22 (77,3% )
Top 3 Match: 19/22 (86,4% )
Black: Weyerstrass
Top 1 Match: 14/22 (63,6% )
Top 2 Match: 21/22 (95,5% )
Top 3 Match: 21/22 (95,5% )
Originally posted by SquelchbelchI wish I were Fritz. He's so smart and I'm sure he gets all the women. 🙁
Very accurate play from both.
Here's what Fritz says:
Fritz 11 @ 30 seconds per move
Pentium 4 2.93GHz 1GB RAM
Hash Table 192MB
Database used www.chesslive.de
[Event "Challenge"]
[Site "http://www.playtheimmortalgame.com"]
[Date "2006.08.27"]
[EndDate "2007.07.16"]
[Round "?"]
[White "Weyerstrass"]
[Black "Akizy"]
[WhiteRating "2451"]
...[text shortened]... 1/22 (95,5% )
Top 3 Match: 21/22 (95,5% )
Originally posted by SquelchbelchWhy only 30 seconds?
Very accurate play from both.
Here's what Fritz says:
Fritz 11 @ 30 seconds per move
Pentium 4 2.93GHz 1GB RAM
Hash Table 192MB
Database used www.chesslive.de
[Event "Challenge"]
[Site "http://www.playtheimmortalgame.com"]
[Date "2006.08.27"]
[EndDate "2007.07.16"]
[Round "?"]
[White "Weyerstrass"]
[Black "Akizy"]
[WhiteRating "2451"]
...[text shortened]... 1/22 (95,5% )
Top 3 Match: 21/22 (95,5% )
The positions after the last database moves are dynamically complex with unbalanced material. The games took a year to play and the resultant positions could have been subject to long analysis several weeks in advance of appearing in the game. My feeling that 30 seconds of Fritz analysis is sufficient for no more than a first impression?
Originally posted by Quiet InterludeYou're quite wrong actually.
Why only 30 seconds?
...My feeling that 30 seconds of Fritz analysis is sufficient for no more than a first impression?
In my experience it makes little difference if you leave Fritz to look for the top 3 choices for 30 seconds or 30 minutes. Little statistical difference anyway. When I analyse the games, I do so from start to finish in 1 hit, therefore Fritz actually has much longer time to look at each line because of cumulative analysis & the hash table.
In a game like the one above, Fritz largely agreed with the choices of almost all moves, so the cumulative stuff is of great relevance.
If you get 2 players who consistently choose moves way outside the top 3 30 second Fritz choice, then yes - the cumulative analysis of lines won't apply because Fritz is going into the position cold, or at least with scant analysis of disregarded lines.
The 2nd reason that 30 second analysis is fine is that to analyse an average game of, say, 90 ply with 30 moves from both out of book takes about 2 hours.
What do you want? 5 minutes per move & you not only get very similar end result figures, but you've also wasted about a whole day analysing/writing-up 1 single game.
For the reasons above, this is why Games Moderators on this site use the same methods I do for matchup analysis.
Originally posted by SquelchbelchFair enough if you say so. What then is the significance of moves that are not top three choice?
You're quite wrong actually.
In my experience it makes little difference if you leave Fritz to look for the top 3 choices for 30 seconds or 30 minutes. Little statistical difference anyway...
Originally posted by Quiet InterludeStatistically, the very best CC/OTB GM players can match Fritz 11's top 3 choices in many games once out of book around 80% of the time.
Fair enough if you say so. What then is the significance of moves that are not top three choice?
Games Mods analysed pre-computer CC World Championship matches, I myself have fully analysed Fischer-Spassky 1972 & others have analysed Kasparov-Karpov games all with similar results.
From a technical point of view, the moves outside the top 3 often have had scant analysis from Fritz during the 30 second time slot.
Of course many times they have very similar evaluation scores to the top 3 choices and very occasionally a move that Fritz disregarded "out-Fritzes" the engine & beats the top choice hands-down.
I've analysed maybe 150-200 games & I reckon this happens once in every 2 or 3 games. Therefore, these moves have no statistical relevance whatsoever & further support the case that 30 seconds is plenty of time.
Plug any top GM PGN into Fritz yourself, give it 30 seconds to analyse each top 3 choice & provided Fritz is set-up correctly on a decent PC & you'll soon get the picture.
Fritz has a very good idea indeed at 30 seconds and in typical positions is able to complete between 4 & 6 cycles of analysis per move.
I wasn't really asking about a statistical relevance to the non top three choice - I was more interested in whether investigation into them could determine anything significant.
Sadly I cannot replicate the Fritz trial - I fell out with the chessbase stable sometime ago when I spent money on an earlier chessbase version with Fritz 7 which didn't maintain compatibility with later windows updates. I think it was service pack 1 that finally killed it off and I decided not to reinvest at £80 or more. I can enter three or four weekend tournaments for that and have five lessons each time from someone better than me! I use a crafty engine to look over my OTB games for missed opportunities or unseen howlers and to sink my opponents inaccurate post mortem analysis...