Here we're in rare agreement.....mclane wrote:it would be a small step. because on slower hardware (1 cpu or 2 cpu's) zappa is not equal to rybka. on those machines there is still 100 ELO difference, or ?Sebastian Boehme wrote:This match is probably going to break the Rybka domination in engine chess, if Zappa should deliver more games like the last 2.
I really look forward to see this fish swimming dead on the top of the water
Devilish regards,
Sebastian
Mexico - Zappa 5.5 v Rybka 4.5- Zappa wins $10,000
Moderators: Harvey Williamson, Steve B, Watchman
Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
-
Terry McCracken
- Senior Member
- Posts: 1300
- Joined: Tue Jul 31, 2007 5:49 pm
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
-
Terry McCracken
- Senior Member
- Posts: 1300
- Joined: Tue Jul 31, 2007 5:49 pm
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
Round 6 4.0-2.0
[Event "Mexico"]
[Site "?"]
[Date "2007.09.24"]
[Round "?"]
[White "Rybka"]
[Black "Zappa Mexico X64"]
[Result "1/2-1/2"]
[ECO "C92"]
[PlyCount "103"]
[EventDate "2007.??.??"]
[TimeControl "3600+10"]
{384MB, ZapM X Black.ctg, ZAPPA} 1. e4 {0} e5 {B/0 0} 2. Nf3 {8} Nc6 {B/0 0} 3.
Bb5 {7} a6 {B/0 0} 4. Ba4 {9} Nf6 {B/0 0} 5. O-O {8} Be7 {B/0 0} 6. Re1 {11} b5
{B/0 0} 7. Bb3 {8} d6 {B/0 0} 8. c3 {8} O-O {B/0 0} 9. h3 {8} Bb7 {B/0 0} 10.
d4 {10} Re8 {B/0 0} 11. Nbd2 {12} Bf8 {B/0 0} 12. a4 {9} h6 {B/0 0} 13. Bc2 {8}
exd4 {B/0 0} 14. cxd4 {20} Nb4 {B/0 0} 15. Bb1 {7} c5 {B/0 0} 16. d5 {7} Nd7 {
B/0 0} 17. Ra3 {10} c4 {B/0 0} 18. axb5 {8} axb5 {B/0 0} 19. Nd4 {8} Qb6 {B/0 0
} 20. Nf5 {9} Ne5 {B/0 0} 21. Rg3 {9} g6 {B/0 0} 22. Nf3 {8} Ned3 {B/0 0} 23.
Be3 {11} Qd8 {B/0 0} 24. Bxh6 {17} Qf6 {B/0 0} 25. Qd2 {9} Nxe1 {B/0 0} 26.
Bxf8 {9} Nxf3+ {B/0 0} 27. Rxf3 {10} gxf5 {B/0 0} 28. Rxf5 {10} Qg6 {B/0 0} 29.
Rg5 {8} Qxg5 {B/0 0} 30. Qxg5+ {7} Kxf8 {B/0 0} 31. Qh6+ {11} Kg8 {B/0 0} 32.
Kh2 {8} Ra1 {B/0 0} 33. Qxd6 {144} Rxb1 {B/0 0} 34. Qxb4 {97} Rxe4 {B/0 0} 35.
Qxb5 {66} c3 {B/0 0} 36. Qxb7 {41} Kg7 {B/0 0} 37. d6 {98} c2 {0.00/18 179} 38.
Qxe4 {98} c1=Q {0.00/19 0} 39. d7 {258} Qg1+ {0.00/20 0} 40. Kg3 {25} Rd1 {
0.00/20 0} 41. Qf5 {231} f6 {0.00/21 0} 42. Qe4 {34} Rd2 {0.00/20 0} 43. d8=Q {
67} Qxf2+ {0.00/20 0} 44. Kh2 {43} Rxd8 {0.00/20 1} 45. Qe7+ {33} Kg6 {
0.00/21 142} 46. Qxd8 {46} Qf4+ {0.00/22 0} 47. Kg1 {126} Qc1+ {0.00/21 22} 48.
Kf2 {69} Qxb2+ {0.00/21 120} 49. Kf3 {(Ke3) 15} Qc3+ {0.00/21 105} 50. Ke2 {
(Kf4) 42} Qe5+ {0.00/21 140} 51. Kd1 {(Kd3) 61} Qa1+ {0.00/21 73} 52. Kd2 {
(Ke2) 64} 1/2-1/2
[Site "?"]
[Date "2007.09.24"]
[Round "?"]
[White "Rybka"]
[Black "Zappa Mexico X64"]
[Result "1/2-1/2"]
[ECO "C92"]
[PlyCount "103"]
[EventDate "2007.??.??"]
[TimeControl "3600+10"]
{384MB, ZapM X Black.ctg, ZAPPA} 1. e4 {0} e5 {B/0 0} 2. Nf3 {8} Nc6 {B/0 0} 3.
Bb5 {7} a6 {B/0 0} 4. Ba4 {9} Nf6 {B/0 0} 5. O-O {8} Be7 {B/0 0} 6. Re1 {11} b5
{B/0 0} 7. Bb3 {8} d6 {B/0 0} 8. c3 {8} O-O {B/0 0} 9. h3 {8} Bb7 {B/0 0} 10.
d4 {10} Re8 {B/0 0} 11. Nbd2 {12} Bf8 {B/0 0} 12. a4 {9} h6 {B/0 0} 13. Bc2 {8}
exd4 {B/0 0} 14. cxd4 {20} Nb4 {B/0 0} 15. Bb1 {7} c5 {B/0 0} 16. d5 {7} Nd7 {
B/0 0} 17. Ra3 {10} c4 {B/0 0} 18. axb5 {8} axb5 {B/0 0} 19. Nd4 {8} Qb6 {B/0 0
} 20. Nf5 {9} Ne5 {B/0 0} 21. Rg3 {9} g6 {B/0 0} 22. Nf3 {8} Ned3 {B/0 0} 23.
Be3 {11} Qd8 {B/0 0} 24. Bxh6 {17} Qf6 {B/0 0} 25. Qd2 {9} Nxe1 {B/0 0} 26.
Bxf8 {9} Nxf3+ {B/0 0} 27. Rxf3 {10} gxf5 {B/0 0} 28. Rxf5 {10} Qg6 {B/0 0} 29.
Rg5 {8} Qxg5 {B/0 0} 30. Qxg5+ {7} Kxf8 {B/0 0} 31. Qh6+ {11} Kg8 {B/0 0} 32.
Kh2 {8} Ra1 {B/0 0} 33. Qxd6 {144} Rxb1 {B/0 0} 34. Qxb4 {97} Rxe4 {B/0 0} 35.
Qxb5 {66} c3 {B/0 0} 36. Qxb7 {41} Kg7 {B/0 0} 37. d6 {98} c2 {0.00/18 179} 38.
Qxe4 {98} c1=Q {0.00/19 0} 39. d7 {258} Qg1+ {0.00/20 0} 40. Kg3 {25} Rd1 {
0.00/20 0} 41. Qf5 {231} f6 {0.00/21 0} 42. Qe4 {34} Rd2 {0.00/20 0} 43. d8=Q {
67} Qxf2+ {0.00/20 0} 44. Kh2 {43} Rxd8 {0.00/20 1} 45. Qe7+ {33} Kg6 {
0.00/21 142} 46. Qxd8 {46} Qf4+ {0.00/22 0} 47. Kg1 {126} Qc1+ {0.00/21 22} 48.
Kf2 {69} Qxb2+ {0.00/21 120} 49. Kf3 {(Ke3) 15} Qc3+ {0.00/21 105} 50. Ke2 {
(Kf4) 42} Qe5+ {0.00/21 140} 51. Kd1 {(Kd3) 61} Qa1+ {0.00/21 73} 52. Kd2 {
(Ke2) 64} 1/2-1/2
-
Ted Summers
- Member
- Posts: 269
- Joined: Tue Jul 31, 2007 2:20 pm
- Location: Marietta, GA
- Contact:
-
Eelco de Groot
- Member
- Posts: 52
- Joined: Tue Jul 31, 2007 5:42 pm
- Location: Groningen
Re: Round 6 4.0-2.0
Hello Harvey,Harvey Williamson wrote:[Event "Mexico"]
[Site "?"]
[Date "2007.09.24"]
[Round "?"]
[White "Rybka"]
[Black "Zappa Mexico X64"]
[Result "1/2-1/2"]
[ECO "C92"]
[PlyCount "103"]
[EventDate "2007.??.??"]
[TimeControl "3600+10"]
{384MB, ZapM X Black.ctg, ZAPPA} 1. e4 {0} e5 {B/0 0} 2. Nf3 {8} Nc6 {B/0 0} 3.
Bb5 {7} a6 {B/0 0} 4. Ba4 {9} Nf6 {B/0 0} 5. O-O {8} Be7 {B/0 0} 6. Re1 {11} b5
{B/0 0} 7. Bb3 {8} d6 {B/0 0} 8. c3 {8} O-O {B/0 0} 9. h3 {8} Bb7 {B/0 0} 10.
d4 {10} Re8 {B/0 0} 11. Nbd2 {12} Bf8 {B/0 0} 12. a4 {9} h6 {B/0 0} 13. Bc2 {8}
exd4 {B/0 0} 14. cxd4 {20} Nb4 {B/0 0} 15. Bb1 {7} c5 {B/0 0} 16. d5 {7} Nd7 {
B/0 0} 17. Ra3 {10} c4 {B/0 0} 18. axb5 {8} axb5 {B/0 0} 19. Nd4 {8} Qb6 {B/0 0
} 20. Nf5 {9} Ne5 {B/0 0} 21. Rg3 {9} g6 {B/0 0} 22. Nf3 {8} Ned3 {B/0 0} 23.
Be3 {11} Qd8 {B/0 0} 24. Bxh6 {17} Qf6 {B/0 0} 25. Qd2 {9} Nxe1 {B/0 0} 26.
Bxf8 {9} Nxf3+ {B/0 0} 27. Rxf3 {10} gxf5 {B/0 0} 28. Rxf5 {10} Qg6 {B/0 0} 29.
Rg5 {8} Qxg5 {B/0 0} 30. Qxg5+ {7} Kxf8 {B/0 0} 31. Qh6+ {11} Kg8 {B/0 0} 32.
Kh2 {8} Ra1 {B/0 0} 33. Qxd6 {144} Rxb1 {B/0 0} 34. Qxb4 {97} Rxe4 {B/0 0} 35.
Qxb5 {66} c3 {B/0 0} 36. Qxb7 {41} Kg7 {B/0 0} 37. d6 {98} c2 {0.00/18 179} 38.
Qxe4 {98} c1=Q {0.00/19 0} 39. d7 {258} Qg1+ {0.00/20 0} 40. Kg3 {25} Rd1 {
0.00/20 0} 41. Qf5 {231} f6 {0.00/21 0} 42. Qe4 {34} Rd2 {0.00/20 0} 43. d8=Q {
67} Qxf2+ {0.00/20 0} 44. Kh2 {43} Rxd8 {0.00/20 1} 45. Qe7+ {33} Kg6 {
0.00/21 142} 46. Qxd8 {46} Qf4+ {0.00/22 0} 47. Kg1 {126} Qc1+ {0.00/21 22} 48.
Kf2 {69} Qxb2+ {0.00/21 120} 49. Kf3 {(Ke3) 15} Qc3+ {0.00/21 105} 50. Ke2 {
(Kf4) 42} Qe5+ {0.00/21 140} 51. Kd1 {(Kd3) 61} Qa1+ {0.00/21 73} 52. Kd2 {
(Ke2) 64} 1/2-1/2
Impressive opening preparation, I wonder if Erdo had this prepared long ago!? Harvey I had a little question, do you know maybe whether this version of Zappa has some form of result learning? I'm asking because as far as I can see the very first move that Zappa had to calculate on its own, after 37. d6 from Rybka, it was already showing a 0.00 evaluation for 37... c2. If I let Toga calculate it sees many moves of the played variation ahead so I'll buy it if it is just Zappa's Singular Extensions kicking in that let it see so far ahead with a correct evaluation, but on its own Toga does not see any draw evaluation here. It does for the first iterations, but then it thinks that White is still almost a pawn ahead, although it sees the whole game continuation till move 48 or so already in its analysis. I wonder what Rybka was evaluating at this point?
When I go to move 48 and let Toga Checkov look at the position it sees the draw, then I go backwards again to move 37 step by step, and I get the following analysis with this hindsight information:
[fen]8/1Q3pk1/3P4/8/4r3/2p4P/1P3PPK/1r6 [/fen]
Diagram from White's viewpoint, Black to move:
[fen]8/1Q3pk1/3P4/8/4r3/2p4P/1P3PPK/1r6 w - - 0 1[/fen]
8/1Q3pk1/3P4/8/4r3/2p4P/1P3PPK/1r6 b - -
Engine: Toga II 1.3.4 Checkov 2c bitbases (256 MB)
by Thomas Gaksch and Fabien Letouzey
6/23 0:00 -1.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qxb2 (54.997)
7/25 0:00 -1.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qxb2 (81.034)
8/28 0:00 -1.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qxb2 (138.239)
9/30 0:00 -0.94 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qxb2+
49.Kg3 (235.051)
10/32 0:00 -0.94 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qxb2+
49.Kg3 (473.569)
11/34 0:01 -0.94 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qxb2+
49.Kg3 (916.488) 710
12/38 0:02 -0.94 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf1 Qb1+ (1.659.379) 735
13/38 0:02 0.00++ 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (2.117.027) 729
14/38 0:02 0.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (2.129.019) 728
15/40 0:07 0.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (4.934.627) 698
16/65 0:49 0.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (36.017.543) 723
17/65 6:47 0.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (299.655.339) 734
18/65 15:07 0.00 37...c2 38.Qxe4 c1Q 39.d7 Qg1+
40.Kg3 Rd1 41.Qf5 f6 42.Qe4 Rd2
43.d8Q Qxf2+ 44.Kh2 Rxd8 45.Qe7+ Kg6
46.Qxd8 Qf4+ 47.Kg1 Qc1+ 48.Kf2 Qf4+
49.Ke2 Qe4+ 50.Kf2 Qf4+ (661.465.442) 728
but calculating without any hashtable contents, Toga thinks much -I don't know how much-, longer that White has still a significant advantage. It must be missing the correct line somewhere deep. So my question was whether Zappa actually could see the draw already here right away, or was it a stored result?
Boy I think team Rybka will not be very pleased with the game today
Regards, Eelco
Careful! Even moonlit dewdrops,
If you’re lured to watch,
Are a wall before the truth.
- Sogyo (18th century)
If you’re lured to watch,
Are a wall before the truth.
- Sogyo (18th century)
-
Mark Uniacke
- Hiarcs Author
- Posts: 1459
- Joined: Sun Jul 29, 2007 1:32 pm
- Location: United Kingdom
- Contact:
-
Eelco de Groot
- Member
- Posts: 52
- Joined: Tue Jul 31, 2007 5:42 pm
- Location: Groningen
Wow! Thanks Mark, for finding out and asking!Mark Uniacke wrote:Harvey confirms yes Zappa was scoring 0.00 from the end of book.
We don't believe any learning was involved just a very good search! (and great prep from Erdo)
Zappa has a very good search indeed then !
That makes this kind of opening preparation also a little bit easier. But quiting his daytime job for a month just for this match, that was quite a leap of faith from Erdo!
Best Regards,
Eelco
Careful! Even moonlit dewdrops,
If you’re lured to watch,
Are a wall before the truth.
- Sogyo (18th century)
If you’re lured to watch,
Are a wall before the truth.
- Sogyo (18th century)
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
Round 7 4.5 - 2.5
[Event "Blitz:5'+10"]
[Site "?"]
[Date "2007.09.25"]
[Round "?"]
[White "Zappa Mexico X64 "]
[Black "Rybka"]
[Result "1/2-1/2"]
[ECO "B35"]
[PlyCount "133"]
[EventDate "2007.??.??"]
[TimeControl "300+10"]
{4MB, ZapM X Wihte.ctg, ZAPPA} 1. e4 {B/0 0} c5 {6} 2. Nf3 {B/0 0} Nc6 {7} 3.
d4 {B/0 0} cxd4 {8} 4. Nxd4 {B/0 0} g6 {10} 5. Nc3 {B/0 0} Bg7 {9} 6. Be3 {
B/0 0} Nf6 {11} 7. Bc4 {B/0 0} Qa5 {10} 8. O-O {B/0 0} O-O {10} 9. Nb3 {B/0 0}
Qd8 {10} 10. Be2 {0.34/18 379} d6 {81} 11. f4 {B/0 0} b6 {13} 12. a3 {
0.32/16 272} Bb7 {(Ld7) 171} 13. Qd3 {0.29/18 339} Nd7 {(Tc8) 137} 14. Rad1 {
0.22/18 223} Nc5 {(Tc8) 9} 15. Nxc5 {0.07/19 304} bxc5 {111} 16. Qd2 {
0.22/18 54} Nd4 {130} 17. Bd3 {0.24/18 49} Rc8 {101} 18. Ne2 {0.09/18 439} Qc7
{(Sxe2+) 215} 19. b3 {0.17/17 132} a5 {(Kh8) 190} 20. f5 {0.29/18 206} a4 {270}
21. Nxd4 {0.52/20 0} cxd4 {106} 22. Bh6 {0.26/19 0} axb3 {86} 23. Bxg7 {
0.35/18 82} Kxg7 {90} 24. cxb3 {0.35/19 0} Qc3 {92} 25. Qa2 {0.39/18 33} Qa5 {
(e5) 133} 26. b4 {0.22/18 223} Qe5 {99} 27. Qd2 {0.39/18 210} Ra8 {(Lxe4) 180}
28. Ra1 {0.44/18 92} Rfc8 {(Lc6) 9} 29. Rf3 {0.34/17 94} Ba6 {(Kg8) 66} 30. Rh3
{0.28/18 157} h5 {98} 31. a4 {0.36/18 86} Bxd3 {88} 32. Rxd3 {0.36/19 0} Rc4 {
(Ta7) 67} 33. fxg6 {0.20/18 112} fxg6 {24} 34. Rf3 {0.08/18 67} Rc7 {(Tf8) 9}
35. a5 {0.51/16 24} Qxe4 {20} 36. Raf1 {0.58/16 2} e5 {47} 37. b5 {0.50/17 9}
Qd5 {57} 38. b6 {0.54/17 0} Rd7 {41} 39. Qg5 {0.54/17 17} Kh7 {75} 40. g4 {
0.58/18 0} Rg7 {39} 41. Rg3 {0.58/18 0} d3 {55} 42. gxh5 {0.50/18 1} Qxa5 {32}
43. hxg6+ {0.44/17 19} Kg8 {6} 44. h4 {0.44/17 43} e4 {58} 45. Qg4 {1.38/16 9}
Qxb6+ {189} 46. Kh2 {0.90/19 0} Re8 {(Db2+) 296} 47. h5 {0.90/8 1} Qb2+ {34}
48. Kh3 {0.78/17 0} Qe2 {35} 49. Qf5 {0.78/18 38} d2 {9} 50. h6 {0.78/18 0} Qc4
{8} 51. Ra3 {0.78/18 66} Qc8 {(e3) 8} 52. Qxc8 {0.24/7 1} Rxc8 {10} 53. hxg7 {
0.24/18 16} e3 {8} 54. Rxe3 {0.24/18 0} Rc1 {9} 55. Rf7 {0.24/18 11} Rh1+ {8}
56. Kg2 {0.24/19 39} Rg1+ {8} 57. Kh2 {0.05/20 40} Rh1+ {6} 58. Kg3 {0.05/21 39
} Rg1+ {8} 59. Kf2 {0.00/21 18} Rf1+ {6} 60. Kxf1 {0.00/21 29} d1=Q+ {12} 61.
Kf2 {0.00/22 30} Qc2+ {17} 62. Kf3 {0.00/22 25} Qxg6 {(Dd1+) 7} 63. Rf4 {
0.00/18 15} Qh5+ {47} 64. Kg3 {0.00/22 0} Qg5+ {422} 65. Kf3 {0.00/25 0} Qh5+ {
(Dd5+) 35} 66. Kf2 {0.00/21 25} Kxg7 {12} 67. Rg3+ {0.00/21 11} 1/2-1/2
[Site "?"]
[Date "2007.09.25"]
[Round "?"]
[White "Zappa Mexico X64 "]
[Black "Rybka"]
[Result "1/2-1/2"]
[ECO "B35"]
[PlyCount "133"]
[EventDate "2007.??.??"]
[TimeControl "300+10"]
{4MB, ZapM X Wihte.ctg, ZAPPA} 1. e4 {B/0 0} c5 {6} 2. Nf3 {B/0 0} Nc6 {7} 3.
d4 {B/0 0} cxd4 {8} 4. Nxd4 {B/0 0} g6 {10} 5. Nc3 {B/0 0} Bg7 {9} 6. Be3 {
B/0 0} Nf6 {11} 7. Bc4 {B/0 0} Qa5 {10} 8. O-O {B/0 0} O-O {10} 9. Nb3 {B/0 0}
Qd8 {10} 10. Be2 {0.34/18 379} d6 {81} 11. f4 {B/0 0} b6 {13} 12. a3 {
0.32/16 272} Bb7 {(Ld7) 171} 13. Qd3 {0.29/18 339} Nd7 {(Tc8) 137} 14. Rad1 {
0.22/18 223} Nc5 {(Tc8) 9} 15. Nxc5 {0.07/19 304} bxc5 {111} 16. Qd2 {
0.22/18 54} Nd4 {130} 17. Bd3 {0.24/18 49} Rc8 {101} 18. Ne2 {0.09/18 439} Qc7
{(Sxe2+) 215} 19. b3 {0.17/17 132} a5 {(Kh8) 190} 20. f5 {0.29/18 206} a4 {270}
21. Nxd4 {0.52/20 0} cxd4 {106} 22. Bh6 {0.26/19 0} axb3 {86} 23. Bxg7 {
0.35/18 82} Kxg7 {90} 24. cxb3 {0.35/19 0} Qc3 {92} 25. Qa2 {0.39/18 33} Qa5 {
(e5) 133} 26. b4 {0.22/18 223} Qe5 {99} 27. Qd2 {0.39/18 210} Ra8 {(Lxe4) 180}
28. Ra1 {0.44/18 92} Rfc8 {(Lc6) 9} 29. Rf3 {0.34/17 94} Ba6 {(Kg8) 66} 30. Rh3
{0.28/18 157} h5 {98} 31. a4 {0.36/18 86} Bxd3 {88} 32. Rxd3 {0.36/19 0} Rc4 {
(Ta7) 67} 33. fxg6 {0.20/18 112} fxg6 {24} 34. Rf3 {0.08/18 67} Rc7 {(Tf8) 9}
35. a5 {0.51/16 24} Qxe4 {20} 36. Raf1 {0.58/16 2} e5 {47} 37. b5 {0.50/17 9}
Qd5 {57} 38. b6 {0.54/17 0} Rd7 {41} 39. Qg5 {0.54/17 17} Kh7 {75} 40. g4 {
0.58/18 0} Rg7 {39} 41. Rg3 {0.58/18 0} d3 {55} 42. gxh5 {0.50/18 1} Qxa5 {32}
43. hxg6+ {0.44/17 19} Kg8 {6} 44. h4 {0.44/17 43} e4 {58} 45. Qg4 {1.38/16 9}
Qxb6+ {189} 46. Kh2 {0.90/19 0} Re8 {(Db2+) 296} 47. h5 {0.90/8 1} Qb2+ {34}
48. Kh3 {0.78/17 0} Qe2 {35} 49. Qf5 {0.78/18 38} d2 {9} 50. h6 {0.78/18 0} Qc4
{8} 51. Ra3 {0.78/18 66} Qc8 {(e3) 8} 52. Qxc8 {0.24/7 1} Rxc8 {10} 53. hxg7 {
0.24/18 16} e3 {8} 54. Rxe3 {0.24/18 0} Rc1 {9} 55. Rf7 {0.24/18 11} Rh1+ {8}
56. Kg2 {0.24/19 39} Rg1+ {8} 57. Kh2 {0.05/20 40} Rh1+ {6} 58. Kg3 {0.05/21 39
} Rg1+ {8} 59. Kf2 {0.00/21 18} Rf1+ {6} 60. Kxf1 {0.00/21 29} d1=Q+ {12} 61.
Kf2 {0.00/22 30} Qc2+ {17} 62. Kf3 {0.00/22 25} Qxg6 {(Dd1+) 7} 63. Rf4 {
0.00/18 15} Qh5+ {47} 64. Kg3 {0.00/22 0} Qg5+ {422} 65. Kf3 {0.00/25 0} Qh5+ {
(Dd5+) 35} 66. Kf2 {0.00/21 25} Kxg7 {12} 67. Rg3+ {0.00/21 11} 1/2-1/2
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
My opinion is that zappa's advantage relative to rybka is the evaluation and not the search.Eelco de Groot wrote:Wow! Thanks Mark, for finding out and asking!Mark Uniacke wrote:Harvey confirms yes Zappa was scoring 0.00 from the end of book.
We don't believe any learning was involved just a very good search! (and great prep from Erdo
Zappa has a very good search indeed then !
That makes this kind of opening preparation also a little bit easier. But quiting his daytime job for a month just for this match, that was quite a leap of faith from Erdo!
Best Regards,
Eelco
I hope that rybka's superior search will help her to score at least 2.5 out of 3 but I am passimistic about it.
Uri
In the game zappa used more timeHarvey Williamson wrote:[fen]4r1k1/6r1/3p2PP/5Q2/2q1p3/6RK/3p4/5R2 w - - 0 51[/fen]
In this position it seems hxg7 is winning mate in 31 - Anthony told me that with clean hash Zappa finds this in 1 minute!
Oh well 4.5/7 3 games to go is ok,
Harvey
51. Ra3 {0.78/18 66}
I wonder if the 1 minute is deterministic or based on luck and can be also 90 seconds in case of bad luck(I assume zappa is not deterministic with parallel search).
Uri
-
Terry McCracken
- Senior Member
- Posts: 1300
- Joined: Tue Jul 31, 2007 5:49 pm
Why the evaluation rather than search? It's known that Rybka scores better on one or two cpus than Zappa.Uri Blass wrote:My opinion is that zappa's advantage relative to rybka is the evaluation and not the search.Eelco de Groot wrote:Wow! Thanks Mark, for finding out and asking!Mark Uniacke wrote:Harvey confirms yes Zappa was scoring 0.00 from the end of book.
We don't believe any learning was involved just a very good search! (and great prep from Erdo
Zappa has a very good search indeed then !
That makes this kind of opening preparation also a little bit easier. But quiting his daytime job for a month just for this match, that was quite a leap of faith from Erdo!
Best Regards,
Eelco
I hope that rybka's superior search will help her to score at least 2.5 out of 3 but I am passimistic about it.
Uri
Mystified Regards,
Terry
I think that rybka scores better on one cpu thanks to superior search.Terry McCracken wrote:Why the evaluation rather than search? It's known that Rybka scores better on one or two cpus than Zappa.Uri Blass wrote:My opinion is that zappa's advantage relative to rybka is the evaluation and not the search.Eelco de Groot wrote:Wow! Thanks Mark, for finding out and asking!Mark Uniacke wrote:Harvey confirms yes Zappa was scoring 0.00 from the end of book.
We don't believe any learning was involved just a very good search! (and great prep from Erdo
Zappa has a very good search indeed then !
That makes this kind of opening preparation also a little bit easier. But quiting his daytime job for a month just for this match, that was quite a leap of faith from Erdo!
Best Regards,
Eelco
I hope that rybka's superior search will help her to score at least 2.5 out of 3 but I am passimistic about it.
Uri
Mystified Regards,
Terry
The fact that zappa scales better with more cpu reduce the search advantage of rybka but rybka has still better search.
I think that the last game demonstrates the search advantage of rybka and the evaluation advantage of zappa.
It seems to me that Rybka evaluated the position as equal when Zappa was right in evaluating the position as advantage for Zappa.
Rybka also saw that she is losing by fxg7 but fortunately zappa did not see the win probably because of inferior search.
Uri
-
Harvey Williamson
- Site Admin
- Posts: 6079
- Joined: Sun Jul 29, 2007 6:57 am
- Location: Media City, UK
- Contact:
Uri Blass wrote:I think that rybka scores better on one cpu thanks to superior search.Terry McCracken wrote:Why the evaluation rather than search? It's known that Rybka scores better on one or two cpus than Zappa.Uri Blass wrote:My opinion is that zappa's advantage relative to rybka is the evaluation and not the search.Eelco de Groot wrote:Wow! Thanks Mark, for finding out and asking!Mark Uniacke wrote:Harvey confirms yes Zappa was scoring 0.00 from the end of book.
We don't believe any learning was involved just a very good search! (and great prep from Erdo
Zappa has a very good search indeed then !
That makes this kind of opening preparation also a little bit easier. But quiting his daytime job for a month just for this match, that was quite a leap of faith from Erdo!
Best Regards,
Eelco
I hope that rybka's superior search will help her to score at least 2.5 out of 3 but I am passimistic about it.
Uri
Mystified Regards,
Terry
The fact that zappa scales better with more cpu reduce the search advantage of rybka but rybka has still better search.
I think that the last game demonstrates the search advantage of rybka and the evaluation advantage of zappa.
It seems to me that Rybka evaluated the position as equal when Zappa was right in evaluating the position as advantage for Zappa.
Rybka also saw that she is losing by fxg7 but fortunately zappa did not see the win probably because of inferior search.
Uri
Hi Uri,
First fxg is not possible so I assume you mean hxg. zappa on Anthonys laptop finds this in 1 minute. Most engines find the move instantly and then go away from it. Are you sure Rybka found this in the game and stayed with it? Or is it like Zappa and found it later with clean hash?
Harvey