Not logged inRybka Chess Community Forum
Forum RybkaChess Help Search Register Login
Up Topic The Rybka Lounge / Correspondence Chess / Deleted Vs. Vytron (DPA Vs. IA) 1/2 - 1/2
1 2 3 4 5 6 Previous Next  
Parent - - By Banned for Life (Gold) Date 2009-07-16 16:55
I am using the gui from Fritz 8.
Parent - By titanium cranium (***) [us] Date 2009-07-16 20:47
I am using the gui from Fritz 11.
Parent - - By Uly (Gold) [mx] Date 2009-07-16 17:31
You must have something configured wrong, AFAIK DPA works the same in all GUIs.
Parent - By titanium cranium (***) [us] Date 2009-07-16 20:48
Again, I've never been able to find Multipv_cp with Fritz 11, and my DPA output looks different too.
Parent - - By Uly (Gold) [mx] Date 2009-07-16 18:17
Well, I guess that's an example of a case in where, if there was a move better than d4 that required high depth to find, DPA wouldn't have found it.

Time for this move: 1:30:03 (29:22 of scale + 1:00:41 of bank)

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3 Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5 Re8 13. d4 Ne4

rn1qr1k1/pp3pp1/2pb3p/3pBp2/2PPnP2/1P2P3/P3B1PP/RN1Q1RK1 w - -


11.00   0:01   -0.03   13...Ne4 14.Nd2 Nc3 15.Qe1 Bb4 16.a3 Nxe2+ 17.Qxe2 Bxd2 18.Qxd2 f6 19.Bxb8 Rxb8 20.Qd3 Qd7 (187.545) 127
12.01   0:02   -0.04   13...Ne4 14.Nd2 Nc3 15.Qe1 Bb4 16.a3 Nxe2+ 17.Qxe2 Bxd2 18.Qxd2 f6 19.Bxb8 Rxb8 20.Rfe1 Qd7 21.Rac1 (259.992) 131
13.01   0:03   -0.03   13...Ne4 14.Nd2 Nc3 15.Qe1 Bb4 16.a3 Nxe2+ 17.Qxe2 Bxd2 18.Qxd2 f6 19.Bxb8 Rxb8 20.Rfe1 Qd7 21.Rac1 Rbd8 (466.732) 136
14.01   0:07   -0.03   13...Ne4 14.Nd2 Nc3 15.Qe1 Bb4 16.a3 Nxe2+ 17.Qxe2 Bxd2 18.Qxd2 f6 19.Bxb8 Rxb8 20.Rfe1 Qd7 21.Rac1 Rbd8 (945.547) 136
15.01   0:17   -0.03   13...Ne4 14.Bd3 Bf8 15.Qh5 g6 16.Qe2 Nd7 17.cxd5 cxd5 18.Bxe4 dxe4 (2.287.929) 136
16.01   0:54   -0.08   13...Ne4 14.Nd2 Nc3 15.Qe1 Bb4 16.a3 Nxe2+ 17.Qxe2 Bxd2 18.Qxd2 f6 19.Bxb8 Rxb8 20.Rf3 Qd7 21.Rc1 Re7 22.Qa5 dxc4 23.bxc4 Rbe8 24.Qc3 (7.779.101) 146
17.01   2:09   -0.12   13...Ne4 14.Nd2 Nc3 15.Qe1 Nxe2+ 16.Qxe2 Na6 17.Bxd6 Qxd6 18.c5 Qe6 19.Rae1 Nb4 20.Nb1 Re7 21.a3 Na6 22.Nd2 Rae8 (19.496.472) 154
18.01   4:14   -0.12   13...Ne4 14.Bd3 Bf8 15.Bxe4 dxe4 16.Bxb8 Rxb8 17.Qc2 Qf6 18.Nc3 Ba3 19.Rad1 Rbd8 20.c5 Bb4 21.Qb2 Qe6 (39.645.372) 159
19.01   6:53   -0.11   13...Ne4 14.Bd3 Bf8 15.Bxe4 dxe4 16.Bxb8 Rxb8 17.Qc2 Qd7 18.Nc3 Red8 19.Qb2 b5 20.cxb5 cxb5 21.Rac1 Rbc8 22.Rc2 Rc6 (65.698.126) 162
20.01  10:45   -0.10   13...Ne4 14.Bd3 Bf8 15.Bxe4 dxe4 16.Bxb8 Rxb8 17.Qc2 Qd7 18.Nc3 Red8 19.Qb2 b5 20.cxb5 cxb5 21.Rac1 Rbc8 22.Ne2 Bb4 23.Rxc8 Rxc8 (102.089.960) 161
21.01  29:03   -0.11   13...Ne4 14.Bd3 Bf8 15.Bxe4 dxe4 16.Bxb8 Rxb8 17.Qc2 b6 18.Nc3 Ba3 19.Nb1 Be7 20.Nc3 Qd6 21.Qe2 Rbd8 22.Rfd1 Qb4 23.Qd2 (287.182.408) 168
best move: Nf6-e4 time: 31:18.562 min  n/s: 166.150  nodes: 304.786.636

Remaining bank: 58:44
Parent - - By yanquis1972 (****) [us] Date 2009-07-16 19:02
are you subtracting from your bank when he goes under 30 mins?
Parent - By Uly (Gold) [mx] Date 2009-07-16 19:27
I'm subtracting from the scale, it's the same, i.e.:

Time for this move: 1:30:03 (46:20 of time for this move + 43:43 of bank)

But faster to calculate.
Parent - - By titanium cranium (***) [us] Date 2009-07-16 20:59
Isn't it true that, at this point in the game, both DPA and IA have similar assessments of the position?
Parent - By Uly (Gold) [mx] Date 2009-07-16 22:46
Yes, on the last output DPA was just a thicker version of IA, I think, of about depth 18. The fun will begin when either method plays an unexpected critical move, but so far I think DPA has been better (the DPA output seems more useful than a single PV.)
Parent - - By Banned for Life (Gold) Date 2009-07-17 00:03
Time consumed for this move: 25 minutes

14. Qc1

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1

A lot of time was spent considering g4, but this never made it into the move list. Of course this is not the only anti-thematic element here. There's a bishop sitting in my knight outpost! :-D

{2048MB, vytron.ctg, ATS-PENRYN} 1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 (14. Bf3 Bf8 (14... Na6 15. Bxe4 (15. Bxd6 Qxd6 (15...
Nxd6 16. cxd5 $11 {[%eval -10,15]}) 16. Qe2 $11 {[%eval -17,15]}) (15. Qe2 Be7
16. Bxe4 $17 {[%eval -74,15]}) 15... fxe4 (15... dxe4 16. Bxd6 $11 {
[%eval 6,15]}) 16. Bxd6 $11 {[%eval -5,15]}) (14... Be7 15. Bxe4 (15. Bxb8 Rxb8
(15... Qxb8 16. Bxe4 $11 {[%eval -9,15]}) 16. Bxe4 $11 {[%eval -15,15]}) (15.
Nd2 Na6 16. Bh5 $15 {[%eval -63,15]}) 15... fxe4 (15... dxe4 16. Qh5 $11 {
[%eval 0,15]}) 16. Qg4 $11 {[%eval -1,15]}) 15. Bxe4 (15. Nd2 Na6 (15... Nd7
16. cxd5 $11 {[%eval -23,15]}) 16. cxd5 $15 {[%eval -33,15]}) (15. a3 Nd7 (
15... Na6 16. cxd5 $11 {[%eval -20,15]}) 16. Nd2 $15 {[%eval -46,15]}) 15...
fxe4 (15... dxe4 16. Bxb8 $11 {[%eval -12,15]}) 16. Bxb8 $11 {[%eval -16,15]})
(14. Bd3 Bf8 (14... Qd7 15. a3 (15. Nd2 Bf8 (15... Bb4 16. Nxe4 $11 {
[%eval 3,15]}) 16. a3 $11 {[%eval 2,15]}) (15. Qe1 Na6 (15... Be7 16. Bxe4 $11
{[%eval 1,15]}) 16. Bxe4 $11 {[%eval 0,15]}) 15... Bf8 (15... Be7 16. Qe2 $11 {
[%eval 6,15]}) 16. Nd2 $11 {[%eval 2,15]}) (14... Na6 15. cxd5 (15. Bxd6 Nxd6 (
15... Qxd6 16. c5 $11 {[%eval -3,15]}) 16. c5 $11 {[%eval -5,15]}) (15. Bxe4
fxe4 (15... dxe4 16. Bxd6 $11 {[%eval 6,15]}) 16. Bxd6 $11 {[%eval -5,15]})
15... cxd5 (15... Bxe5 16. fxe5 $14 {[%eval 51,15]}) 16. Nd2 $11 {[%eval 4,15]}
) 15. Nd2 (15. Bxe4 fxe4 (15... dxe4 16. Bxb8 $11 {[%eval -12,15]}) 16. Bxb8
$11 {[%eval -16,15]}) (15. Qh5 g6 16. Qh3 $15 {[%eval -31,15]}) 15... Qa5 (
15... Nd7 16. cxd5 $11 {[%eval -15,15]}) 16. Nxe4 $11 {[%eval -16,15]}) (14.
Nd2 Nc3 15. Qe1 Nxe2+ (15... Bb4 16. a3 $11 {[%eval -10,15]}) 16. Qxe2 $11 {
[%eval -12,15]}) 14... Nd7 (14... Qa5 15. Bxd6 (15. a3 Nd7 (15... dxc4 16. bxc4
$11 {[%eval -5,15]}) 16. Ra2 $11 {[%eval -15,15]}) (15. Bd3 Be7 (15... Bf8 16.
Bxe4 $11 {[%eval -14,15]}) 16. Bxe4 $15 {[%eval -27,15]}) 15... Nxd6 16. Nc3
$11 {[%eval -1,15]}) (14... Qe7 15. Bd3 (15. Bf3 Nd7 (15... Na6 16. Bxe4 $11 {
[%eval 1,15]}) 16. Bxe4 $11 {[%eval 0,15]}) (15. cxd5 cxd5 16. Nc3 $11 {
[%eval -7,15]}) 15... Nd7 (15... Na6 16. a3 $11 {[%eval 5,15]}) 16. Bxe4 $11 {
[%eval 0,15]}) 15. Nc3 (15. Nd2 Ndf6 (15... Ba3 16. Qxa3 $11 {[%eval -2,15]})
16. c5 $11 {[%eval -13,15]}) (15. Bxd6 Nxd6 16. c5 $11 {[%eval -18,15]}) 15...
Ndf6 (15... Qe7 16. Nxe4 $11 {[%eval 8,15]}) 16. Nxe4 $11 {[%eval -5,15]} *
Parent - - By Mark (****) [us] Date 2009-07-17 00:50
I've been following this game with IDeA and move 14 was the first move where IDeA differed from the game moves.  Some of the listed IDeA moves had other moves with the same eval.

14. Bd3 was also at -.11. 
14. Qc1 was at -.15 with more positions than the other moves. 

IDeA setting (300 sec or d=15)
Q6600 @ 2.4GHz (3 cores)

Moves that IDeA would have chosen:

    DPA  IA      IDeA  eval  time
-----------------------------------------------
6.  Nh4  ...     Nh4  -.07  60 min
     ...   O-O   O-O  -.07  56 min
7.  Be2  ...     Be2  -.07  55 min
     ...   Be4    Be4  -.11  64 min
8.  d3   ...      d3  -.07  80 min
     ...   Bf5    Bf5  -.11  58 min
9.  O-O  ...    O-O  -.11  70 min
     ...   h6     h6  -.11  65 min
10.  c4  ...    c4  -.11  70 min 
     ...   c6     c6  -.04  58 min 
11.  Nxf5  ...  Nxf5  -.07  8 hours 20 min   
     ...   exf5  exf5  -.04  skipped
12.  Be5  ...   Be5  -.04  60 min
     ...   Re8   Re8  -.07  64 min
13.  d4  ...    d4  -.15  110 min
     ...   Ne4  Ne4  -.11  30 min
14.  Qc1  ...  Bf3   -.11  90 min
Parent - - By Uly (Gold) [mx] Date 2009-07-17 01:57
It's likely that those are Rybka moves that would be played no matter the analysis method. 10...c6 was interesting, I think if Deleted played faster up to that point, 10...Bh7 would have been played instead.
Parent - - By Mark (****) [us] Date 2009-07-17 02:20
IDeA had 10. ... Bh7 at -.04 also, and with many more positions evaluated than c6.  I can post the tree after the game ends.
Parent - By Banned for Life (Gold) Date 2009-07-17 04:52
You don't have to worry about influencing the game here. The only variation should be due to the settings I choose for DPA and how they affect my time usage, and how Vytron uses his banked time and responds to positions with a forced move.
Parent - - By Banned for Life (Gold) Date 2009-07-17 04:55
A lot of the white moves here are rather strange:

12. Be5, and 14. Qc1 aren't really in keeping with the spirit of the opening. Putting a bishop on e5 rather than a knight is pretty strange too. I guess that's why we have opening books! :-)
Parent - By Uly (Gold) [mx] Date 2009-07-17 06:50
There should be some engine out there that manages this opening better, even when automated, though.
Parent - - By Mark (****) [us] Date 2009-07-17 10:16
12. Be5 was the clear favorite with IDeA.
Parent - By Banned for Life (Gold) Date 2009-07-17 14:14
Not too surprising given the small number of positions. I think IDeA would take a long, long time to come up with the best kingside attacks on its own. If you fed it with 30 or 40 good lines it might be a different story though.
Parent - By Uly (Gold) [mx] Date 2009-07-17 21:02
Time for move: 1:37:22 (38:38 of scale + 58:44 of bank)

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3 Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5 Re8 13. d4 Ne4 14. Qc1 Nd7

r2qr1k1/pp1n1pp1/2pb3p/3pBp2/2PPnP2/1P2P3/P3B1PP/RNQ2RK1 w - -


11.00   0:01   -0.11   14...Nd7 15.Nd2 Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.Qc3 Qd7 (181.132) 123
12.01   0:02   -0.12   14...Nd7 15.Nd2 Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.Nf3 Ne4 19.Bd3 Bf6 (249.939) 122
13.01   0:02   -0.07   14...Nd7 15.Nd2 Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.Nf3 Ne4 19.Bd3 Bf6 (316.437) 122
14.01   0:05   -0.11   14...Nd7 15.Nd2 Ndf6 16.c5 Nxd2 17.Bxf6 Qxf6 18.Qxd2 Bf8 19.Rae1 b6 20.b4 Rab8 21.a3 (664.268) 131
15.01   0:08   -0.10   14...Nd7 15.Nd2 Ndf6 16.c5 Nxd2 17.Bxf6 Qxf6 18.Qxd2 Bf8 19.Bd3 b6 20.b4 Rab8 21.a3 g6 (1.137.655) 132
16.01   0:17   -0.10   14...Nd7 15.Nd2 Ndf6 16.c5 Nxd2 17.Bxf6 Qxf6 18.Qxd2 Bf8 19.Bd3 b6 20.b4 Rab8 21.Rfe1 g6 22.Rab1 (2.363.708) 139
17.01   0:43   -0.12   14...Nd7 15.Nd2 Ndf6 16.c5 Nxd2 17.Qxd2 Be7 18.Bxf6 Bxf6 19.Bd3 g6 20.Rf3 Qd7 21.Rb1 Re7 22.b4 (6.328.736) 148
18.01   1:54   -0.13   14...Nd7 15.Nd2 Ndf6 16.c5 Nxd2 17.Qxd2 Be7 18.Bxf6 Bxf6 19.Bd3 g6 20.Rf3 Qd7 21.Rb1 Re7 22.b4 a6 23.Rbf1 (16.718.067) 150
19.01   7:54   -0.10   14...Nd7 15.Nd2 Ndf6 16.Nf3 Rc8 17.Bxd6 Qxd6 18.c5 Qc7 19.b4 a5 20.b5 b6 21.Ne5 bxc5 22.bxc6 c4 23.Qc2 (71.583.729) 154
20.01  10:51   -0.10   14...Nd7 15.Nd2 Ndf6 16.Nf3 Rc8 17.Bxd6 Qxd6 18.c5 Qc7 19.b4 a5 20.b5 b6 21.Ne5 bxc5 (97.952.267) 153
21.01  24:34   -0.08   14...Nd7 15.Nc3 Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Qd2 Be7 20.Rac1 Qd7 21.f5 Rac8 22.Qe1 Rf8 23.Qd2 b6 (217.920.465) 151
best move: Nb8-d7 time: 28:01.937 min  n/s: 149.017  nodes: 244.744.192

Remaining bank: 1:09:20
Parent - - By Banned for Life (Gold) Date 2009-07-18 20:19
Time consumed for this move: 22 minutes

15. Nc3?

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3

15. Nd2 is clearly better and is absolutely necessary to eventually establish a knight outpost on e5.

{2048MB, vytron.ctg, ATS-PENRYN} 1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 (15. Nd2 Ndf6 (15... Ba3 16. Qxa3 Nxd2 17.
Rfd1 $11 {[%eval -1,15]}) (15... Bb4 16. Nxe4 dxe4 (16... fxe4 17. a3 $11 {
[%eval 7,15]}) 17. a3 $11 {[%eval 4,15]}) 16. Nxe4 (16. Bd3 Be7 (16... Bb4 17.
Nxe4 $11 {[%eval 0,15]}) 17. Nxe4 $11 {[%eval -2,15]}) (16. Nf3 Bb4 (16... Be7
17. Bd3 $11 {[%eval -9,15]}) 17. Bd3 $11 {[%eval -10,15]}) 16... Nxe4 (16...
dxe4 17. Qe1 $11 {[%eval 12,15]}) 17. Bf3 $11 {[%eval -2,15]}) (15. Bxd6 Nxd6
16. c5 (16. Re1 dxc4 (16... Nf6 17. Nc3 $11 {[%eval -23,15]}) 17. bxc4 $11 {
[%eval -24,15]}) (16. Nc3 dxc4 (16... Nf6 17. c5 $11 {[%eval -23,15]}) 17. bxc4
$11 {[%eval -23,15]}) 16... Ne4 (16... Nb5 17. Bxb5 $11 {[%eval -1,15]}) 17.
Nd2 $11 {[%eval -18,15]}) (15. Qb2 Ndf6 (15... dxc4 16. bxc4 (16. Bxc4 Bf8 (
16... Bb4 17. a3 $11 {[%eval -20,15]}) 17. Nc3 $15 {[%eval -67,15]}) 16... Qb6
(16... Be7 17. Nc3 $11 {[%eval -1,15]}) 17. Qc1 $11 {[%eval -11,15]}) (15...
Qa5 16. Bxd6 (16. a3 dxc4 (16... Ndf6 17. b4 $11 {[%eval -11,15]}) 17. bxc4 $15
{[%eval -36,15]}) (16. c5 Be7 (16... Ndxc5 17. dxc5 $11 {[%eval -21,15]}) 17.
b4 $15 {[%eval -41,15]}) 16... Nxd6 17. b4 $11 {[%eval -15,15]}) 16. Bf3 (16.
Rc1 Bf8 (16... Be7 17. Bf3 $15 {[%eval -37,15]}) 17. Bf3 $15 {[%eval -40,15]})
(16. Qc1 Be7 (16... Bf8 17. Bd3 $15 {[%eval -33,15]}) 17. c5 $15 {
[%eval -54,15]}) 16... Bf8 (16... Be7 17. Nd2 $11 {[%eval -22,15]}) 17. Rc1 $15
{[%eval -40,15]}) 15... Ndf6 (15... Ba3 16. Qxa3 (16. Qc2 Ndf6 (16... a5 17.
Nxe4 $11 {[%eval -2,15]}) 17. Bf3 $11 {[%eval -21,15]}) 16... Nxc3 17. Bd3 $11
{[%eval 6,15]}) (15... Nxc3 16. Qxc3 Nf6 17. Bd3 $11 {[%eval 0,15]}) 16. Nxe4 (
16. Bd3 Bf8 (16... Be7 17. Nxe4 $11 {[%eval -2,15]}) 17. Bxf6 $11 {
[%eval -6,15]}) (16. Bf3 Bf8 (16... Be7 17. Nxe4 $11 {[%eval -2,15]}) 17. Qc2
$11 {[%eval -17,15]}) 16... Nxe4 (16... dxe4 17. Qe1 $11 {[%eval 12,15]}) 17.
Bf3 $11 {[%eval -2,15]} *
Parent - - By Kappatoo (****) [de] Date 2009-07-18 21:16
Sorry, I don't get this comment. How can you ever establish a knight outpost on e5? Black always has f6.
Parent - - By Banned for Life (Gold) Date 2009-07-18 23:21
There's nothing wrong with having the knight on g6...
Parent - - By Kappatoo (****) [de] Date 2009-07-19 08:53
I'm not sure if Nd2 was better or worse than Nc3. I was just wondering, since as long as there's a black pawn on f7, e5 doesn't match my unerstanding of an outpost.

By the way, I haven't quite followed the rules of this games, but I was having the impression that each of you is using some automatical procedure without human input. So I hope it's okay to comment?!
Parent - By Banned for Life (Gold) Date 2009-07-19 14:16
There's no problem commenting. I'm sure neither DPA nor IA will be insulted! :-)

I will concede the point that e5 wouldn't be a true outpost for the knight, but if black's position is weakened by f6, the knight could stay on e5 for quite a while.
Parent - - By Uly (Gold) [mx] Date 2009-07-20 22:14

> 15. Nc3?


I just had to know, what move would IA have played, and it turns out Nc3 was discarded really quick in favor of Nd2:

10.00   0:01   -0.16   15.Nc3 Ndf6 16.Nxe4 Nxe4 17.Bd3 Bb4 18.Bxe4 dxe4 19.d5 cxd5 (142.098) 117
11.00   0:02   -0.14   15.Nc3 Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.Qd2 Qd7 19.h3 (258.099) 127
11.03   0:02   -0.06   15.Nd2 Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.Nf3 Ne4 19.Qc2 (320.066) 127
12.01   0:03   -0.13   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 (495.356) 135
13.01   0:04   -0.13   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 (604.231) 137
14.01   0:07   -0.12   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 bxc5 21.bxc5 Rab8 22.Bd3 (981.364) 139
15.01   0:11   -0.10   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 f6 21.Ng6 Rab8 22.Nxf8 Rxf8 23.Qc2 (1.594.447) 143
16.01   0:17   -0.10   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 f6 21.Ng6 Rab8 22.Nxf8 Rxf8 23.Qc2 (2.403.205) 142
17.01   0:37   -0.09   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 f6 21.Ng6 Rab8 22.Nxf8 Rxf8 23.Qc2 bxc5 24.bxc5 Rf7 (5.372.721) 148
18.01   1:05   -0.09   15.Nd2 Ndf6 16.c5 Bf8 17.Bxf6 Qxf6 18.Nf3 b6 19.b4 Qe6 20.Ne5 f6 21.Ng6 Rab8 22.Nxf8 Rxf8 23.Qc2 bxc5 24.bxc5 (9.526.628) 150
19.01   4:52   -0.12   15.Nd2 Ndf6 16.c5 Nxd2 17.Bxf6 Qxf6 18.Qxd2 Bf8 19.b4 Qe6 20.Rf3 b6 21.Bd3 g6 22.Rb1 Rab8 23.Qc2 bxc5 (43.649.445) 152
20.01  19:43   -0.12   15.Nd2 Ndf6 16.Bd3 Bf8 17.Bxf6 Qxf6 18.Nf3 c5 19.Ne5 Rac8 20.Qb2 cxd4 21.exd4 Bb4 22.Bxe4 fxe4 23.a3 Ba5 24.Rad1 Red8 25.c5 (180.868.885) 156
best move: Nb1-d2 time: 34:18.547 min  n/s: 158.060  nodes: 317.726.720
Parent - By Banned for Life (Gold) Date 2009-07-21 14:04
Thanks for posting this. IA gives a much better line here.
Parent - - By Vempele (Silver) [fi] Date 2009-07-21 14:10
What happens if you use Multi-PV? I noticed the moves have identical scores in the DPA output.
Parent - - By Banned for Life (Gold) Date 2009-07-21 15:49
Multi-PV (used extensively by DPA of course) should see the difference fairly quickly. I'm surprised that DPA went for Nc3 at depth 16.
Parent - By Banned for Life (Gold) Date 2009-07-21 15:51
Actually it was depth 15 except at the root for this move...
Parent - By Uly (Gold) [mx] Date 2009-07-19 21:15
Time for this move: 1:43:19 (39:59 of scale + 1:09:20 of bank)

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3 Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5 Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6

r2qr1k1/pp3pp1/2pb1n1p/3pBp2/2PPnP2/1PN1P3/P3B1PP/R1Q2RK1 w - -


11.00   0:01   -0.13   15...Ndf6 16.c5 Be7 17.Bf3 Qa5 18.Bxe4 fxe4 19.Bxf6 Bxf6 20.Qb2 Qc7 (154.092) 129
12.01   0:02   -0.16   15...Ndf6 16.c5 Be7 17.Bf3 Qa5 18.Bxe4 fxe4 19.Bxf6 Bxf6 20.Qb2 Qc7 21.Rac1 (310.555) 137
13.01   0:02   -0.12   15...Ndf6 16.c5 Be7 17.Bf3 Qa5 18.Bxe4 fxe4 19.Bxf6 Bxf6 20.Qb2 Qc7 21.Rac1 Rad8 (404.505) 139
14.01   0:05   -0.11   15...Ndf6 16.c5 Be7 17.Bf3 Qa5 18.Bxf6 Qxc3 19.Be5 Qd3 20.Rd1 Qb5 21.Bxe4 fxe4 22.Qc2 (727.020) 144
15.01   0:17   -0.18   15...Ndf6 16.c5 Be7 17.Bxf6 Nxf6 18.b4 Bf8 19.Re1 b6 20.Qd2 Qd7 (2.617.192) 154
16.01   0:33   -0.06   15...Ndf6 16.Nxe4 Nxe4 17.Bf3 Be7 18.Bxe4 dxe4 19.d5 cxd5 20.Rd1 Qb6 21.Rxd5 Bc5 22.Bd4 Bxd4 23.Rxd4 Rad8 24.Rd5 Rxd5 25.cxd5 Qb4 26.Qe1 Qd6 (4.730.345) 146
17.01   1:55   -0.11   15...Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Qd2 Rc8 20.Rac1 Be7 21.f5 Ba3 22.Ra1 Bb4 (16.318.624) 145
18.01   2:34   -0.11   15...Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Rb1 Be7 20.c5 b6 21.b4 Qc8 22.Qd2 Qc7 23.Qe2 bxc5 24.bxc5 Rab8 25.Qa6 (21.754.172) 143
19.01   3:43   -0.10   15...Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Rb1 Be7 20.f5 Rc8 21.Qe1 Bf6 22.Rd1 Qd6 23.Qe2 Qb4 24.Rc1 dxc4 25.bxc4 Rcd8 (31.627.066) 145
20.01   6:38   -0.08   15...Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Qe1 Rc8 20.Rc1 Bh4 21.Qe2 Be7 22.f5 Qa5 23.Qe1 Kh8 24.Qf2 Red8 25.Qd2 b6 (56.338.170) 144
21.01  17:42   -0.08   15...Ndf6 16.Bd3 Be7 17.Bxf6 Bxf6 18.Bxe4 fxe4 19.Qe1 Rc8 20.Rc1 Bh4 21.Qd2 Be7 22.cxd5 (160.421.109) 154
best move: Nd7-f6 time: 19:43.937 min  n/s: 152.031  nodes: 175.756.083

Remaining bank: 1:23:35
Parent - - By Banned for Life (Gold) Date 2009-07-20 21:42
Time consumed for this move: 34 minutes

16. Bf3

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6 16. Bf3

Changed to depth 16 to return to ~30 minutes per move at 4-3-3-2 branching factor.

{2048MB, vytron.ctg, ATS-PENRYN} 1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6 16. Bf3 (16. Bd3 Be7 (16... Bf8 17.
Nxe4 (17. Bxe4 Nxe4 (17... dxe4 18. Bxf6 $11 {[%eval -8,16]}) 18. Nxe4 $11 {
[%eval -8,16]}) (17. Bxf6 Nxf6 (17... Qxf6 18. Bxe4 $11 {[%eval -8,16]}) 18.
Bxf5 $11 {[%eval -17,16]}) 17... Nxe4 (17... dxe4 18. Be2 $11 {[%eval -7,16]})
18. Bxe4 $11 {[%eval -8,16]}) (16... Nxc3 17. Qxc3 g6 (17... Ne4 18. Bxe4 $11 {
[%eval 15,16]}) 18. Qd2 $11 {[%eval -2,16]}) 17. Nxe4 (17. Bxe4 Nxe4 (17...
dxe4 18. Bxf6 $11 {[%eval -5,16]}) 18. Nxe4 $11 {[%eval -8,16]}) (17. Bxf6 Bxf6
(17... Nxf6 18. Bxf5 $11 {[%eval -2,16]}) 18. Bxe4 $11 {[%eval -8,16]}) 17...
Nxe4 (17... dxe4 18. Be2 $11 {[%eval 5,16]}) 18. Bxe4 $11 {[%eval -8,16]}) (16.
Nxe4 Nxe4 (16... dxe4 17. c5 (17. Qc3 Be7 (17... a5 18. Bxf6 $11 {[%eval 16,16]
}) 18. Rad1 $11 {[%eval 12,16]}) (17. Rb1 a5 (17... a6 18. b4 $11 {
[%eval 16,16]}) 18. Qe1 $11 {[%eval 10,16]}) 17... Bb8 (17... Bc7 18. b4 $11 {
[%eval 21,16]}) 18. Bxf6 $11 {[%eval 19,16]}) 17. Bf3 (17. Bd3 Bf8 (17... Be7
18. Bxe4 $11 {[%eval -8,16]}) 18. Bxe4 $11 {[%eval -8,16]}) (17. Bxd6 Qxd6 (
17... Nxd6 18. cxd5 $11 {[%eval -15,16]}) 18. c5 $11 {[%eval -24,16]}) 17...
Bf8 (17... Be7 18. Bxe4 $11 {[%eval -8,16]}) 18. Bxe4 $11 {[%eval -8,16]}) (16.
c5 Bf8 (16... Be7 17. Bxf6 (17. Bf3 Nh7 (17... Nxc3 18. Bxf6 $11 {
[%eval -19,16]}) 18. Nxe4 $15 {[%eval -35,16]}) (17. Bd3 Nh7 (17... Nd7 18. Na4
$15 {[%eval -28,16]}) 18. Bxe4 $15 {[%eval -42,16]}) 17... Nxf6 (17... Nxc3 18.
Qxc3 $11 {[%eval -18,16]}) 18. b4 $11 {[%eval -21,16]}) (16... Nxc3 17. Qxc3
Be7 (17... Bf8 18. Bxf6 $11 {[%eval -14,16]}) 18. Bxf6 $11 {[%eval -18,16]})
17. Bf3 (17. Bxf6 Nxf6 (17... Qxf6 18. b4 $11 {[%eval -16,16]}) 18. b4 $15 {
[%eval -33,16]}) (17. Bd3 Nh7 (17... Nxc3 18. Qxc3 $15 {[%eval -27,16]}) 18.
Nxe4 $15 {[%eval -38,16]}) 17... Nxc3 (17... Qd7 18. b4 $11 {[%eval -21,16]})
18. Qxc3 $15 {[%eval -27,16]}) 16... Be7 (16... Bf8 17. Nxe4 (17. Bxe4 dxe4 (
17... Nxe4 18. Nxe4 $11 {[%eval -8,16]}) 18. Bxf6 $11 {[%eval -8,16]}) (17. a3
Qd7 (17... Qe7 18. Re1 $11 {[%eval -17,16]}) 18. Qb2 $11 {[%eval -21,16]})
17... Nxe4 (17... dxe4 18. Be2 $11 {[%eval 0,16]}) 18. Bxe4 $11 {[%eval -8,16]}
) (16... Bb4 17. Nxe4 (17. Bxe4 Bxc3 (17... dxe4 18. Bxf6 $11 {[%eval -3,16]})
18. Bxf5 $11 {[%eval -16,16]}) (17. Bxf6 Nxf6 (17... Qa5 18. Bxe4 $11 {
[%eval 0,16]}) 18. a3 $11 {[%eval -20,16]}) 17... Nxe4 (17... dxe4 18. Be2 $11
{[%eval 0,16]}) 18. Bxe4 $11 {[%eval 0,16]}) 17. Bxe4 (17. Nxe4 Nxe4 (17...
dxe4 18. Be2 $11 {[%eval 6,16]}) 18. Bxe4 $11 {[%eval -8,16]}) (17. Bxf6 Nxf6 (
17... Bxf6 18. Bxe4 $11 {[%eval -8,16]}) 18. Re1 $11 {[%eval -21,16]}) 17...
Nxe4 (17... dxe4 18. Bxf6 $11 {[%eval -5,16]}) 18. Nxe4 $11 {[%eval -8,16]} *
Parent - By Uly (Gold) [mx] Date 2009-07-21 22:06
Oops, the move I posted 4 minutes ago was from Rybka Dynamic, sorry about that, real move coming soon...
Parent - Date 2009-07-21 22:02
[topic branch moved]
Parent - - By Uly (Gold) [mx] Date 2009-07-21 22:47
Here the real move, and it mattered, Rybka Default picks Bf8 instead. Same branching 5 happened, only earlier, risk of taking 125 minutes I don't deem worthwhile, so I'll keep accumulating bank.

Time for this move: 2:16:07 (52:32 from scale + 1:23:35 of bank)

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3 Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5 Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6 16. Bf3 Bf8

r2qrbk1/pp3pp1/2p2n1p/3pBp2/2PPnP2/1PN1PB2/P5PP/R1Q2RK1 w - -


10.00   0:01   -0.09   16...Bb4 17.Nxe4 Nxe4 18.Bxe4 dxe4 19.c5 Qa5 20.Rd1 Bc3 21.Rb1 b6 (170.937) 131
11.00   0:01   -0.09   16...Bb4 17.Nxe4 Nxe4 18.Bxe4 dxe4 19.c5 Qa5 20.Rd1 Bc3 21.Rb1 b6 (206.311) 128
12.01   0:02   -0.11   16...Bb4 17.Nxe4 Nxe4 18.Bxe4 dxe4 19.c5 Qa5 20.Rd1 Bc3 21.Rb1 b6 22.Bd6 Red8 23.Qc2 (336.123) 134
13.01   0:04   -0.04   16...Bb4 17.Nxe4 dxe4 18.Bxf6 Qxf6 19.Be2 b6 20.a3 Be7 21.Qd2 Rad8 22.b4 (552.289) 135
13.36   0:14   -0.08   16...Bf8 17.Qc2 Nxc3 18.Bxf6 Qxf6 19.Qxc3 Rac8 20.Rac1 Qe7 21.cxd5 cxd5 22.Qd2 Qe6 23.Rxc8 Rxc8 24.Re1 Rc6 (2.067.168) 144
14.01   0:19   -0.07   16...Bf8 17.Qc2 Nxc3 18.Bxf6 Qxf6 19.Qxc3 Rac8 20.Rfe1 b6 21.b4 Qd6 22.c5 Qe6 23.Rad1 (2.828.391) 147
15.01   0:25   -0.08   16...Bf8 17.Qc2 Nxc3 18.Bxf6 Qxf6 19.Qxc3 Qe7 20.Rfe1 Qd7 21.cxd5 cxd5 22.Qa5 Rad8 (3.692.956) 148
16.01   0:39   -0.08   16...Bf8 17.Qc2 Nxc3 18.Bxf6 Qxf6 19.Qxc3 Qe7 20.Rae1 a5 21.Qd3 Qd7 22.c5 g6 23.Qc2 Bg7 24.a3 (5.989.236) 153
17.01   2:21   -0.20   16...Bf8 17.Qc2 Qa5 18.Nxe4 Nxe4 19.Bxe4 dxe4 20.c5 b6 21.Rfc1 bxc5 22.dxc5 Red8 23.Bd4 Rab8 24.Rd1 Qb4 25.Rac1 Rb5 (23.175.576) 167
18.01   3:01   -0.19   16...Bf8 17.Qc2 Qa5 18.Nxe4 Nxe4 19.Bxe4 dxe4 20.c5 b6 21.Rfc1 bxc5 22.dxc5 Red8 23.Bd4 Rab8 24.Qc4 Qa3 25.Qe2 g6 26.Rc4 (29.623.676) 167
19.01   5:17   -0.19   16...Bf8 17.Qc2 Qa5 18.Nxe4 Nxe4 19.Bxe4 dxe4 20.c5 Qa3 21.Qe2 b6 22.cxb6 axb6 23.Qh5 g6 24.Qh4 Bg7 25.Bxg7 Kxg7 26.Qe1 Re6 27.Qc3 Kg8 (53.320.086) 172
20.01  25:45   -0.20   16...Bf8 17.a3 Rc8 18.c5 Qd7 19.b4 Nxc3 20.Bxf6 Ne4 21.Bh4 Qc7 22.Qc2 b6 23.Rac1 Re6 24.Rfe1 Qb7 (254.477.298) 168
best move: Bd6-f8 time: 27:02.156 min  n/s: 166.511  nodes: 263.751.270

Remaining bank: 1:49:05

My comment "Now the differences in scores is [12cp], I suspect IA is extending beyond DPA's horizon" still applies.
Parent - - By Banned for Life (Gold) Date 2009-07-22 04:04
My comment "Now the differences in scores is [12cp], I suspect IA is extending beyond DPA's horizon" still applies.

The comment makes sense now because the eval jumps at depth 17, which is deeper than DPA is searching. It didn't make sense before when it jumped at depth 14.
Parent - By Vempele (Silver) [fi] Date 2009-07-22 08:31

> The comment makes sense now because the eval jumps at depth 17, which is deeper than DPA is searching.


DPA gets depth 16 (and then some) for the position following Bf8, IA only gets 16 at that point.
Parent - By Uly (Gold) [mx] Date 2009-07-22 09:08
I'm talking about score end results, regardless of depth (if we're taking relatively the same time per analysis method, I just hope at some point in the game it matters.)
Parent - - By Dadi Jonsson (Silver) [is] Date 2009-07-22 16:39
Here is what Rybka says at depth 25 :smile:

[-0.23]  d=25  1.a3 Qd7 2.Bxf6 Nxf6 3.Re1 b6 4.Qc2 Rac8 5.h3 g6 6.Qa2 a5 7.Qd2 Re6 8.Qd3 Rce8 9.Re2 Kg7 10.Ree1 Be7 (7:08.50) 6412444kN
[-0.23]  d=24  1.a3 Rc8 2.Bxf6 Nxf6 3.Re1 Re6 4.Qd2 Qd7 5.c5 Rce8 6.Re2 b6 7.b4 g6 8.Qd3 Bg7 9.Rae1 bxc5 10.bxc5 Rb8 11.Rb1 (4:03.58) 3736360kN
[-0.26]  d=24  1.Qc2 Qa5 2.Nxe4 Nxe4 3.Bxe4 dxe4 4.c5 Qa3 5.Rfc1 Rac8 6.Qc4 Red8 7.Rc2 b6 8.Kf2 bxc5 9.dxc5 Rd5 10.Bd4 Rcd8 11.Rac1 Kh7 12.Kg1 g6 13.Qc3 (1:32.48) 1505394kN
[-0.22]  d=23  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Bh4 Ne4 4.Rfd1 g6 5.Be1 Qa3 6.Rd3 Qe7 7.a4 Qd7 8.Rc1 Rad8 9.Rdd1 Qe6 10.Qe2 Ba3 11.Rc2 Be7 (0:32.38) 498941kN
[-0.21]  d=22  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Bh4 Ne4 4.Rfd1 Rab8 5.Be1 Qa3 6.cxd5 cxd5 7.Qe2 g6 8.Rac1 Rbc8 9.Rc2 Rxc2 10.Qxc2 Qa6 11.a4 Re6 12.Qb2 (0:19.53) 295845kN
[-0.21]  d=21  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Bh4 Ne4 4.Rfd1 Rac8 5.Be1 Qc7 6.Rac1 Qd7 7.Qd3 Qe6 8.Bxe4 fxe4 9.Qe2 Ba3 10.Rc2 Be7 11.cxd5 Qxd5 12.Qc4 Qh5 (0:15.15) 230909kN
[-0.21]  d=20  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Bh4 Ne4 4.Rfd1 Rac8 5.Be1 Qc7 6.Rac1 Qd7 7.Qd3 Qe6 8.Bxe4 fxe4 9.Qe2 Ba3 10.Rc2 Be7 11.cxd5 Qxd5 12.Qc4 Qh5 (0:15.15) 230909kN
[-0.21]  d=19  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Bh4 Ne4 4.Rfd1 Rac8 5.Be1 Qc7 6.Rac1 Qd7 7.Qd3 Qe6 8.Bxe4 fxe4 9.Qe2 Ba3 10.Rc2 Be7 11.cxd5 Qxd5 12.Qc4 Qh5 (0:15.15) 230909kN
[-0.24]  d=19  1.a3 Rc8 (0:08.15) 122285kN
[-0.21]  d=18  1.a3 Rc8 2.Qb2 Qd7 3.Bxf6 Nxf6 4.Rfe1 Re6 5.Re2 dxc4 6.bxc4 b5 7.c5 Rce8 8.Qd2 a5 9.Rb1 Qc7 (0:04.44) 70790kN
[-0.23]  d=18  1.Nxe4 Nxe4 2.Bxe4 dxe4 3.d5 cxd5 4.Rd1 Bc5 5.b4 Bxb4 6.Rxd5 Qe7 7.Bd4 Rac8 8.Rxf5 f6 9.Rd5 b6 10.Qc2 Rc6 11.Qb3 Rec8 12.Rc1 (0:02.08) 32439kN
[-0.11]  d=17  1.Nxe4 Nxe4 2.Bxe4 dxe4 3.d5 Qb6 4.Rd1 Rad8 5.Bd4 Qa6 6.dxc6 bxc6 7.Qc2 Qa3 8.Kf2 f6 9.Qb2 Qa6 10.c5 (0:01.26) 20157kN
[-0.09]  d=16  1.Nxe4 Nxe4 2.Bxe4 dxe4 3.d5 Qb6 4.Rd1 Rad8 5.Bd4 Qa6 6.dxc6 bxc6 7.c5 Re6 8.Qc2 Rg6 9.Kh1 Be7 (0:01.18) 18196kN
[-0.21]  d=16  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Be5 Ne4 4.Bxe4 dxe4 5.c5 Rad8 6.Rfb1 f6 7.b4 Qa6 8.Bc7 Rd7 9.Ba5 Rc8 10.Qb3+ Rd5 11.Rc1 Kh7 (0:00.49) 10341kN
[-0.08]  d=15  1.Qc2 Nxc3 2.Bxf6 Qxf6 3.Qxc3 Qd6 4.a3 a5 5.c5 Qe6 6.Rfe1 g6 7.Rad1 Qd7 8.Qd3 (0:00.27) 4999kN
[-0.08]  d=14  1.Qc2 Nxc3 2.Bxf6 Qxf6 3.Qxc3 Qe7 4.Rae1 Qd7 5.Rc1 Rac8 6.cxd5 cxd5 7.Qa5 b5 8.Qd2 Qe6 (0:00.18) 3271kN
[-0.11]  d=13  1.Qc2 Nxc3 2.Bxf6 Qxf6 3.Qxc3 Qe7 4.cxd5 cxd5 5.Rfe1 a5 6.a3 Qd7 7.Qd2 Rec8 (0:00.12) 2100kN
[-0.07]  d=12  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Be5 Ne4 4.Bxe4 dxe4 5.c5 Rad8 6.a4 f6 (0:00.08) 1454kN
[-0.07]  d=11  1.Qc2 Nxc3 2.Bxf6 Qa5 3.Be5 Ne4 4.Bxe4 dxe4 5.c5 Rad8 6.a4 f6 (0:00.07) 1218kN
[-0.11]  d=11  1.c5 Qd7 2.b4 a5 3.bxa5 Rxa5 4.Qb2 Raa8 5.a4 (0:00.05) 1071kN
Parent - - By Vempele (Silver) [fi] Date 2009-07-22 16:43
I think it's safe to say you're outsearching both of them. :-)

Anyone else think that smiley looks creepy?
Parent - By keoki010 (Bronze) [us] Date 2009-07-22 17:23
Yep! :smile:
Parent - - By Banned for Life (Gold) Date 2009-07-22 20:47
Actually DPA comes up with the same PV with a transposed move ordering at depth 16.
Parent - - By Vempele (Silver) [fi] Date 2009-07-22 20:53
Can you add the PVs to the DPA output?
Parent - By Banned for Life (Gold) Date 2009-07-23 03:50
I don't know how to do this. Of course the PVs show up in the engine window, but they can't be saved (to the best of my knowledge). This would be a useful function, but Chessbase stopped working on DPA at least 7 years ago.
Parent - - By Banned for Life (Gold) Date 2009-07-22 20:45
Time consumed for this move: 51 minutes

17. a3

1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6 16. Bf3 Bf8 17. a3

DPA's assessment gets darker...

{2048MB, vytron.ctg, ATS-PENRYN} 1. b3 d5 2. Bb2 Nf6 3. e3 Bf5 4. f4 e6 5. Nf3
Bd6 6. Nh4 O-O 7. Be2 Be4 8. d3 Bf5 9. O-O h6 10. c4 c6 11. Nxf5 exf5 12. Be5
Re8 13. d4 Ne4 14. Qc1 Nd7 15. Nc3 Ndf6 16. Bf3 Bf8 17. a3 (17. Bxe4 Nxe4 (
17... dxe4 18. Bxf6 (18. Rd1 Nd7 (18... Ng4 19. h3 $11 {[%eval -6,16]}) 19. c5
$11 {[%eval -18,16]}) (18. Na4 Nh7 (18... Nd7 19. c5 $11 {[%eval -23,16]}) 19.
c5 $15 {[%eval -27,16]}) 18... Qxf6 19. Rd1 $11 {[%eval -7,16]}) (17... fxe4
18. Bxf6 (18. Qd2 dxc4 (18... Rc8 19. Rac1 $11 {[%eval 0,16]}) 19. Bxf6 $11 {
[%eval 0,16]}) (18. f5 Rc8 (18... Nd7 19. Bg3 $11 {[%eval -2,16]}) 19. Bxf6 $11
{[%eval -2,16]}) 18... Qxf6 (18... gxf6 19. f5 $14 {[%eval 29,16]}) 19. cxd5
$11 {[%eval 3,16]}) 18. Nxe4 (18. cxd5 Nxc3 19. Qxc3 $19 {[%eval -142,16]})
18... dxe4 19. d5 $11 {[%eval -24,16]}) (17. Nxe4 Nxe4 (17... dxe4 18. Be2 (18.
Bxf6 Qxf6 19. Be2 $11 {[%eval -14,16]}) (18. Bd1 b6 (18... Nh7 19. c5 $11 {
[%eval -6,16]}) 19. Bxf6 $11 {[%eval -15,16]}) 18... b6 (18... Nd7 19. c5 $11 {
[%eval 2,16]}) 19. Qe1 $11 {[%eval 1,16]}) (17... fxe4 18. Be2 (18. Bd1 Nd7 (
18... Qd7 19. Bxf6 $11 {[%eval 10,16]}) 19. Bh5 $11 {[%eval 7,16]}) (18. Bxf6
Qxf6 19. Bg4 $15 {[%eval -54,16]}) 18... Nd7 (18... Nh7 19. Bg4 $11 {
[%eval 10,16]}) 19. Bh5 $11 {[%eval 7,16]}) 18. Bxe4 dxe4 (18... fxe4 19. f5
$11 {[%eval 8,16]}) 19. d5 $11 {[%eval -24,16]}) (17. Kh1 Qa5 (17... Nxc3 18.
Bxf6 (18. Qxc3 Ne4 19. Bxe4 $15 {[%eval -66,16]}) 18... Qa5 (18... Nxa2 19.
Rxa2 $11 {[%eval -19,16]}) 19. Bh4 $15 {[%eval -31,16]}) (17... Rc8 18. c5 (18.
a3 Qd7 (18... Qe7 19. c5 $15 {[%eval -30,16]}) 19. c5 $15 {[%eval -35,16]}) (
18. Bxf6 Nxf6 (18... Qxf6 19. Nxe4 $11 {[%eval -9,16]}) 19. Re1 $15 {
[%eval -30,16]}) 18... Qd7 (18... b6 19. b4 $11 {[%eval -17,16]}) 19. Na4 $11 {
[%eval -25,16]}) 18. Nxe4 (18. Bxe4 Nxe4 19. Nxe4 $15 {[%eval -35,16]}) (18.
Bxf6 Qxc3 (18... Ba3 19. Qc2 $15 {[%eval -62,16]}) 19. Be5 $15 {[%eval -63,16]}
) 18... Nxe4 (18... dxe4 19. Be2 $11 {[%eval -13,16]}) 19. Bxe4 $15 {
[%eval -35,16]}) 17... Rc8 (17... Qd7 18. Bxe4 (18. Qb2 Rac8 (18... Nxc3 19.
Qxc3 $11 {[%eval -22,16]}) 19. Nxe4 $15 {[%eval -27,16]}) (18. Nxe4 Nxe4 (18...
fxe4 19. Be2 $11 {[%eval 14,16]}) 19. Bxe4 $11 {[%eval -21,16]}) 18... Nxe4 19.
Nxe4 $11 {[%eval -21,16]}) (17... Qe7 18. Bxf6 (18. cxd5 cxd5 (18... Nxc3 19.
d6 $11 {[%eval 0,16]}) 19. Bxf6 $11 {[%eval -25,16]}) (18. Re1 Qd7 (18... Qe6
19. Bxf6 $11 {[%eval -21,16]}) 19. Bxf6 $11 {[%eval -21,16]}) 18... Nxf6 (18...
Qxf6 19. Bxe4 $11 {[%eval -5,16]}) 19. Re1 $11 {[%eval -19,16]}) 18. c5 (18.
Bxf6 Nxf6 (18... Qxf6 19. Nxe4 $11 {[%eval -7,16]}) 19. Re1 $11 {[%eval -22,16]
}) (18. Qb2 Qd7 (18... Qa5 19. b4 $11 {[%eval -21,16]}) 19. Nxe4 $15 {
[%eval -27,16]}) 18... Qd7 (18... Kh7 19. b4 $11 {[%eval -13,16]}) 19. Bxf6 $11
{[%eval -22,16]} *
Parent - By Vempele (Silver) [fi] Date 2009-07-22 20:51
DPA searches 1 ply deeper than IA did at the divergence point (19. b4 vs. 19.Bxf6), assuming no extensions.
Parent - - By Uly (Gold) [mx] Date 2009-07-22 21:10
Now, adding more time to my bank doesn't help, what about, in the spirit of keeping DPA to 30min/move, if you use 45 minutes one move, you have 15 minutes for your next move? Or you have 26:15 for your next 4 moves? (from now on) or something?

It just feels unfair that you got infinite time in your bank and can overstep at will.
Parent - - By Banned for Life (Gold) Date 2009-07-23 04:24
DPA can't be precisely controlled. Total analysis time doesn't work at all (it finished in less than 5 minutes and gave me a number of analysis depths of 5), and small changes can easily cause factor of 2 or more changes in analysis time. There is no way to predict how long a move will take a priori. Here are the times taken so far:

1     B   
2     B
3     B
4     B
5     B
6     53
7     34
8     51
9     15
10   22
11   37
12   38
13   19
14   25
15   22
16   34
17   51

Averaging about 34 1/2 minutes per move. This is a little high, but its not too bad considering the total lack of control in this mode.
Parent - By Uly (Gold) [mx] Date 2009-07-23 06:27

> Averaging about 34 1/2 minutes per move.


Ok, that's a good point, apparently I've been playing too fast while you've been taking reasonable time for your moves, sorry.

I'd just hope, when you go over the average to switch back to depth 15, and later on to depth 16, or something.
Parent - - By Dadi Jonsson (Silver) [is] Date 2009-07-23 07:41
Doesn't it produce any usable analysis if you stop it early? Can you, for instance, copy the analysis as it stands at any time? Can you see what is the best move found so far?
Parent - By Banned for Life (Gold) Date 2009-07-23 16:01
Yes, this is my normal mode of operation when I use DPA for correspondance chess. Whatever is available after many hours is available for inspection. Since the root moves are evaluated from best candidate to worst, the results are perfectly reasonable.

The only problem with this approach is that I have to be there to physically stop it at the designated time. Right now, I just restart the GUI and engine, queue up DPA, and wait until it finishes, then I read the time consumed from task manager (max of Rybka core times plus GUI time).

I suggested this approach, as an alternative, when it first because obvious that the total time didn't do what it was supposed to do (its hard to see how it could). It was Vytron who suggested the banking approach.
Up Topic The Rybka Lounge / Correspondence Chess / Deleted Vs. Vytron (DPA Vs. IA) 1/2 - 1/2
1 2 3 4 5 6 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill