Colossus 2007b for download!

[Eelco de Groot]

A new version of Colossus is out! Thanks Martin Bryant! See http://www.colossus.demon.co.uk/chess/c ... essuci.htm

It is a bugfix version, but the general search seems to be different more than you would expect from just a bugfix, see the results in time to plies for the initial position:

rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -

Engine: Colossus 2007b (256 MB)
gemaakt door Martin Bryant

4/06 0:00 +0.16 1.e4 d5 (254)

5.01 0:00 +0.16 1.e4 d5 (425) 424

6.01 0:00 +0.16 1.e4 d5 (595) 594

7.01 0:00 +0.16 1.e4 d5 (803) 802

8.01 0:00 +0.16 1.e4 d5 (1.080) 1079

9.01 0:00 +0.16 1.e4 d5 (1.400) 1399

10.01 0:00 +0.16 1.e4 d5 (1.762) 1761

11.01 0:00 +0.16 1.e4 d5 (2.380) 2379

12.01 0:00 +0.16 1.e4 d5 (4.492) 4491

13.01 0:00 +0.16 1.e4 d5 (22.710) 473

14.01 0:00 +0.16 1.e4 d5 (49.578) 450

15.01 0:11 +0.20 1.e4 d5 2.exd5 (5.648.487) 463

16.01 1:37 +0.14 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Pgxe2 c6 6.d3 e5 7.f4 exf4 8.Lxf4 Pf6
9.O-O-O Lb4 10.Te1 Le6 (50.661.939) 503

17.01 6:22 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 Pxd4 10.Lxd7+ Kxd7 11.Td1 Td8 (195.657.049) 507

18.01 28:52 +0.22 1.e4 e5 2.Pf3 Pc6 3.d4 exd4 4.Pxd4 Dh4
5.Pc3 Lb4 6.Le2 d6 7.Pdb5 La5 8.Dd5 Pf6
9.Dg5 Dxg5 10.Lxg5 (884.087.611) 508


rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -

Engine: Colossus 2007a (256 MB)
gemaakt door Martin Bryant

4/08 0:00 +0.08 1.e4 e5 2.d4 Dh4 (988)

5.01 0:00 -0.08 1.e4 d5 2.Df3 dxe4 3.Dxe4 e6 (3.396) 18

5.01 0:00 +0.20 1.d4 d5 2.Pc3 Pf6 3.Lf4 (4.281) 18

6.01 0:00 0.00 1.d4 d5 2.Pc3 Pf6 3.Lf4 e6 (8.309) 57

6.01 0:00 +0.06 1.e3 d5 2.Pf3 Pf6 3.Lb5+ c6 4.Le2 (10.743) 57

7.01 0:00 +0.06 1.e3 d5 2.Pf3 Pf6 3.Pc3 Lg4 4.d4 (27.228) 148

7.01 0:00 +0.22 1.d4 d5 2.Pc3 Pf6 3.Lf4 e6 4.Pf3 (36.408) 148

8.01 0:00 +0.04 1.d4 d5 2.Pc3 Pf6 3.Lg5 Lg4 4.f3 Lf5 (65.320) 210

8.01 0:00 +0.06 1.e3 Pf6 2.Pf3 c6 3.Le2 e6 4.d4 Ld6 (94.174) 210

8.01 0:00 +0.18 1.e4 d5 2.e5 Pc6 3.d4 Ld7 4.Pc3 e6 (163.703) 210

9.01 0:00 +0.16 1.e4 d5 2.e5 d4 3.Df3 f6 4.Le2 Pc6
5.Dh5+ g6 (251.301) 366

10.01 0:01 +0.20 1.e4 e5 2.Pf3 Df6 3.d4 exd4 4.Dxd4 Pc6
5.Dd5 Pge7 6.Dg5 (861.830) 430

11.01 0:04 +0.10 1.e4 e5 2.d4 exd4 3.Dxd4 Pc6 4.Dc4 Df6
5.Pf3 d6 6.Db5 Le7 (2.130.132) 473

11.02 0:05 +0.14 1.d4 d5 2.Lf4 Pf6 3.e3 e6 4.Pf3 Ld6
5.Lxd6 Dxd6 6.Lb5+ c6 7.Ld3 (2.853.624) 473

12.01 0:08 +0.10 1.d4 Pf6 2.Pf3 d5 3.Lf4 Lf5 4.e3 e6
5.Le2 c6 6.O-O Ld6 7.Lxd6 (4.408.793) 496

13.01 0:22 +0.08 1.d4 d5 2.Lf4 Pf6 3.e3 e6 4.Pf3 Ld6
5.Lxd6 Dxd6 6.c3 Ld7 7.Ld3 O-O (11.441.164) 504

14.01 0:51 +0.08 1.d4 Pf6 2.c3 d5 3.Lf4 Lf5 4.e3 c6
5.Db3 b5 6.a4 Da5 7.Pf3 Lxb1 8.Txb1 (26.075.472) 506

14.02 1:45 +0.14 1.e4 d5 2.e5 c5 3.c3 Pc6 4.Pf3 Dc7
5.d4 Lg4 6.Lb5 cxd4 (54.021.480) 506

15.01 2:32 +0.16 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 c6
5.Pf3 Pf6 6.a3 Dd6 7.d3 Lg4 8.Lg5 Lxf3
9.Dxf3 De5+ 10.De3 (78.507.443) 510

16.01 3:44 +0.14 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 O-O-O 10.d5 (116.446.751) 513

17.01 11:49 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 Pxd4 10.Lxd7+ Kxd7 11.Td1 Td8 (367.226.210) 513

18.01 17:49 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+
4.Pge2 Pf6 5.d4 Db6 6.Tb1 e6 7.Lg5 Da5
8.Lxf6 gxf6 9.Dd3 Pc6 10.Df3 Le7
11.a3 (548.708.509) 515

Colossus 2007b seems to have much smaller searchtrees per ply in this position, except at ply 18 where Colossus 2007b had to change its PV from the Scandinavian to the Scottish opening, that took a lot of time.
I tried it again with 128 Mb hash and now the differences seemed to be less but version 2007b was still faster.

I played no less than two Blitz games of the latest version against each other, and it may be that Colossus 2007b was just unlucky, but it lost both and fairly early after the opening... (opening book is Shredder9.bkt, Shredder UCI interface, 128 Mb hash and 6 minutes + 6 seconds Fischer bonus timecontrol on Athlon 2009 MHz) Could it be that Colossus 2007b is searching a bit too selectively at present? Most likely cause is just an statistically unlucky opening for Colossus 2007b. The variations and evals are very similar but 2007b just ended up in the wrong variations. Here are the first two games:

[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.06.01"]
[Round "1"]
[White "Colossus 2007b"]
[Black "Colossus 2007a"]
[Result "0-1"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. c4 {Boek 0s} e6 {Boek 0s}
3. Nc3 {Boek 0s} Bb4 {Boek 0s} 4. e3 {Boek 0s} c5 {Boek 0s}
5. Nge2 {Boek 0s} d5 {Boek 0s} 6. a3 {Boek 0s} Bxc3+ {Boek
0s} 7. Nxc3 {Boek 0s} cxd4 {Boek 0s} 8. exd4 {Boek 0s} dxc4
{Boek 0s} 9. Bxc4 {Boek 0s} Nc6 {Boek 0s} 10. Be3 {Boek 0s}
O-O {Boek 0s} 11. O-O {Boek 0s} b6 {Boek 0s} 12. Qf3 {Boek
0s} Bb7 {-0.02/13 20s} 13. Bd3 {Boek 0s} Rc8 {-0.04/12 23s}
14. Rad1 {Boek 0s} Na5 {-0.04/12 27s} 15. Qg3 {-0.04/13
16s} Nc4 {+0.04/13 18s} 16. Bxc4 {0.00/13 34s} Rxc4
{0.00/13 16s} 17. Rc1 {0.00/13 15s} Kh8 {0.00/13 22s}
18. Nb5 {0.00/13 22s} Bd5 {+0.04/13 20s} 19. Nxa7 {-0.04/12
15s} Ne4 {+0.22/14 18s} 20. Qf4 {-0.08/14 21s} Nd6
{+0.14/13 19s} 21. Bd2 {-0.08/13 23s} Re8 {+0.02/13 17s}
22. b3 {-0.28/13 25s} e5 {+0.02/14 15s} 23. bxc4 {-0.14/15
19s} exf4 {+0.38/15 15s} 24. cxd5 {-0.30/15 19s} Qg5
{+0.68/14 16s} 25. Nc6 {-0.74/13 18s} Qxd5 {+0.84/14 16s}
26. Bxf4 {-0.76/14 12s} Nc4 {+0.76/14 15s} 27. Ne5
{-0.76/15 18s} b5 {+0.78/14 15s} 28. Nxc4 {-0.86/14 21s}
bxc4 {+0.84/15 14s} 29. Be3 {-0.84/15 18s} f5 {+0.84/15
15s} 30. Rfe1 {-0.84/14 16s} h5 {+0.84/14 14s} 31. h3
{-1.00/13 18s} Kh7 {+0.98/14 16s} 32. Red1 {-1.02/14 11s}
Re4 {+1.32/15 13s} 33. g3 {-1.68/15 15s} Rxe3 {+1.92/16
12s} 34. fxe3 {-1.96/17 10s} Qf3 {+2.10/16 11s} 35. Re1
{-2.14/16 17s} Qxg3+ {+2.12/16 12s} 36. Kf1 {-2.16/16 13s}
Qxh3+ {+2.20/16 11s} 37. Ke2 {-2.20/16 17s} Qg2+ {+2.26/16
11s} 38. Kd1 {-2.20/4 0s} f4 {+2.26/17 11s} 39. Rc2
{-2.26/15 13s} Qe4 {+2.38/15 11s} 40. exf4 {-2.36/14 12s}
Qxd4+ {+2.58/15 11s} 41. Kc1 {-2.52/15 15s} Qxf4+ {+2.46/15
11s} 42. Kb2 {-2.52/16 12s} Qf3 {+2.46/14 11s} 43. a4
{-2.46/14 15s} Qb3+ {+2.64/13 11s} 44. Kc1 {-2.84/15 11s}
Qa3+ {+2.84/14 11s} 45. Kd2 {-2.58/15 10s} Qd3+ {+2.66/14
10s} 46. Kc1 {-2.58/4 0s} h4 {+2.52/14 8s} 47. Kb1
{-2.56/14 12s} Qb3+ {+2.96/14 10s} 48. Rb2 {-2.86/14 11s}
Qxa4 {+3.20/13 8s} 49. Rc1 {-2.78/14 13s} g5 {+2.80/13 11s}
50. Rc3 {-3.06/13 10s} Qd1+ {+2.98/14 8s} 51. Ka2 {-2.98/15
10s} Kg6 {+2.98/14 4s} 52. Rxc4 {-3.22/13 10s} Qd5
{+4.46/13 2s} 53. Kb3 {-4.46/12 2s} h3 {+5.90/14 10s}
54. Rbc2 {-5.58/13 11s} g4 {+6.04/11 1s} 55. Rh2 {-6.04/11
4s} Qb5+ {+6.22/11 9s} 0-1

[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.06.01"]
[Round "2"]
[White "Colossus 2007a"]
[Black "Colossus 2007b"]
[Result "1-0"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. c4 {Boek 0s} e6 {Boek 0s}
3. Nc3 {Boek 0s} Bb4 {Boek 0s} 4. e3 {Boek 0s} c5 {Boek 0s}
5. Nge2 {Boek 0s} d5 {Boek 0s} 6. a3 {Boek 0s} Bxc3+ {Boek
0s} 7. Nxc3 {Boek 0s} cxd4 {Boek 0s} 8. exd4 {Boek 0s} dxc4
{Boek 0s} 9. Bxc4 {Boek 0s} Nc6 {Boek 0s} 10. Be3 {Boek 0s}
O-O {Boek 0s} 11. O-O {Boek 0s} b6 {Boek 0s} 12. Qf3 {Boek
0s} Bb7 {0.00/13 20s} 13. Bd3 {Boek 0s} Rc8 {-0.02/12 19s}
14. Rad1 {Boek 0s} Qe7 {-0.04/12 27s} 15. Qg3 {+0.02/13
31s} Rfd8 {-0.02/12 21s} 16. Rfe1 {+0.04/12 18s} Rd7
{0.00/12 19s} 17. Be2 {+0.10/11 21s} Qd8 {-0.08/11 15s}
18. Bh6 {+0.46/13 26s} Qf8 {-0.30/14 25s} 19. Bg5 {+0.42/13
16s} Nd5 {-0.62/13 22s} 20. Nxd5 {+0.96/13 13s} Rxd5
{-0.96/13 13s} 21. Bf3 {+1.24/14 14s} Rd7 {-1.20/14 23s}
22. d5 {+1.24/14 13s} Nd8 {-1.24/13 14s} 23. dxe6 {+1.12/13
22s} Rxd1 {-1.06/13 14s} 24. exf7+ {+1.10/13 21s} Nxf7
{-1.12/15 16s} 25. Bxd1 {+1.10/14 15s} Nxg5 {-1.12/13 13s}
26. Qxg5 {+1.10/14 19s} Bc6 {-1.12/13 20s} 27. Bb3+
{+1.12/13 24s} Kh8 {-1.16/15 12s} 28. f3 {+1.12/14 18s} Re8
{-1.16/14 17s} 29. Rxe8 {+1.18/13 13s} Qxe8 {-1.18/16 13s}
30. Kf2 {+1.18/15 11s} Qf8 {-1.18/14 16s} 31. Qe5 {+1.18/14
17s} Qd8 {-1.20/14 14s} 32. Kg1 {+1.20/14 17s} Qe8
{-1.12/13 11s} 33. f4 {+1.14/15 18s} Qc8 {-1.12/13 9s}
34. Qe7 {+1.14/14 16s} Bb7 {-1.16/13 13s} 35. Kf2 {+1.14/14
15s} a5 {-1.14/12 10s} 36. g4 {+1.20/13 17s} Qb8 {-1.22/12
13s} 37. Ke1 {+1.22/13 14s} Bc6 {-1.44/13 14s} 38. f5
{+2.16/15 14s} Qe8 {-2.00/14 12s} 39. Qxe8+ {+2.14/18 12s}
Bxe8 {-2.00/4 0s} 40. g5 {+1.84/19 10s} Bh5 {-1.92/18 13s}
41. f6 {+2.16/19 11s} gxf6 {-1.24/19 11s} 42. gxf6
{+2.12/19 8s} h6 {-2.14/19 8s} 43. Be6 {+2.44/20 8s} Kh7
{-2.38/20 9s} 44. Kd2 {+2.46/21 11s} Kg6 {-2.42/20 13s}
45. f7 {+2.44/22 10s} Kg7 {-2.42/21 9s} 46. Kd3 {+2.44/21
7s} Bf3 {-2.42/20 11s} 47. Kc4 {+2.44/19 8s} Be2+ {-2.40/20
10s} 48. Kd5 {+2.42/20 10s} a4 {-2.42/20 8s} 49. Kc6
{+2.62/21 9s} b5 {-2.72/22 9s} 50. Kc5 {+2.80/21 11s} Kf8
{-2.86/21 8s} 51. Bd7 {+3.00/22 11s} Kxf7 {-3.04/21 9s}
52. Bxb5 {+3.22/21 9s} Bd1 {-3.66/21 7s} 53. Kb4 {+3.70/21
8s} Bg4 {-5.24/22 11s} 54. Bxa4 {+3.70/23 9s} Ke6 {-5.32/21
3s} 55. Kc5 {+4.72/18 3s} 1-0

Version 2007a was clearly better than Colossus 2006f, I am curious how it will turn out with the new 2007b! Kirill Kryukov just posted some early results for 2007a but now that version is already obsolete!

Eelco

[mclane]

it depends WHERE the bug fix is.

this

Fixed bug in null move (using wrong sides variables for single piece r=2 decision

could have e.g. a big influence.

[Eelco de Groot]

Hello Thorsten,

Yes, I suppose that could be. I don't have a real idea what a single piece null move reduction is, sounds very interesting and it probably does not mean something specific to endgames. Martin Bryant does not expect much of an elo jump though? Maybe also reducing search instability could be important. The engine does as I said earlier a little bit faster going through the plies, so playing against the previous version I think that should give some extra advantage, just like in those old Belle vs. Belle experiments from Ken Thompson where one ply deeper I believe made a huge difference in strength.

From the next two matches Colossus 2007b lost one with ½-1½ and won one 1½-½ (that was not with the intended book but accidently with my own experiment that only knows how to play one line of the French so far, but that does not matter much)

So now the standings is Colossus 2007b - Colossus 2007a 2-4, that looks a bit better already .

[tjfroh]

A new version of Colossus is out! Thanks Martin Bryant! See http://www.colossus.demon.co.uk/chess/c ... essuci.htm

It is a bugfix version, but the general search seems to be different more than you would expect from just a bugfix, see the results in time to plies for the initial position:

[d]rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -

Engine: Colossus 2007b (256 MB)
gemaakt door Martin Bryant

4/06 0:00 +0.16 1.e4 d5 (254)

5.01 0:00 +0.16 1.e4 d5 (425) 424

6.01 0:00 +0.16 1.e4 d5 (595) 594

7.01 0:00 +0.16 1.e4 d5 (803) 802

8.01 0:00 +0.16 1.e4 d5 (1.080) 1079

9.01 0:00 +0.16 1.e4 d5 (1.400) 1399

10.01 0:00 +0.16 1.e4 d5 (1.762) 1761

11.01 0:00 +0.16 1.e4 d5 (2.380) 2379

12.01 0:00 +0.16 1.e4 d5 (4.492) 4491

13.01 0:00 +0.16 1.e4 d5 (22.710) 473

14.01 0:00 +0.16 1.e4 d5 (49.578) 450

15.01 0:11 +0.20 1.e4 d5 2.exd5 (5.648.487) 463

16.01 1:37 +0.14 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Pgxe2 c6 6.d3 e5 7.f4 exf4 8.Lxf4 Pf6
9.O-O-O Lb4 10.Te1 Le6 (50.661.939) 503

17.01 6:22 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 Pxd4 10.Lxd7+ Kxd7 11.Td1 Td8 (195.657.049) 507

18.01 28:52 +0.22 1.e4 e5 2.Pf3 Pc6 3.d4 exd4 4.Pxd4 Dh4
5.Pc3 Lb4 6.Le2 d6 7.Pdb5 La5 8.Dd5 Pf6
9.Dg5 Dxg5 10.Lxg5 (884.087.611) 508


[d]rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -

Engine: Colossus 2007a (256 MB)
gemaakt door Martin Bryant

4/08 0:00 +0.08 1.e4 e5 2.d4 Dh4 (988)

5.01 0:00 -0.08 1.e4 d5 2.Df3 dxe4 3.Dxe4 e6 (3.396) 18

5.01 0:00 +0.20 1.d4 d5 2.Pc3 Pf6 3.Lf4 (4.281) 18

6.01 0:00 0.00 1.d4 d5 2.Pc3 Pf6 3.Lf4 e6 (8.309) 57

6.01 0:00 +0.06 1.e3 d5 2.Pf3 Pf6 3.Lb5+ c6 4.Le2 (10.743) 57

7.01 0:00 +0.06 1.e3 d5 2.Pf3 Pf6 3.Pc3 Lg4 4.d4 (27.228) 148

7.01 0:00 +0.22 1.d4 d5 2.Pc3 Pf6 3.Lf4 e6 4.Pf3 (36.408) 148

8.01 0:00 +0.04 1.d4 d5 2.Pc3 Pf6 3.Lg5 Lg4 4.f3 Lf5 (65.320) 210

8.01 0:00 +0.06 1.e3 Pf6 2.Pf3 c6 3.Le2 e6 4.d4 Ld6 (94.174) 210

8.01 0:00 +0.18 1.e4 d5 2.e5 Pc6 3.d4 Ld7 4.Pc3 e6 (163.703) 210

9.01 0:00 +0.16 1.e4 d5 2.e5 d4 3.Df3 f6 4.Le2 Pc6
5.Dh5+ g6 (251.301) 366

10.01 0:01 +0.20 1.e4 e5 2.Pf3 Df6 3.d4 exd4 4.Dxd4 Pc6
5.Dd5 Pge7 6.Dg5 (861.830) 430

11.01 0:04 +0.10 1.e4 e5 2.d4 exd4 3.Dxd4 Pc6 4.Dc4 Df6
5.Pf3 d6 6.Db5 Le7 (2.130.132) 473

11.02 0:05 +0.14 1.d4 d5 2.Lf4 Pf6 3.e3 e6 4.Pf3 Ld6
5.Lxd6 Dxd6 6.Lb5+ c6 7.Ld3 (2.853.624) 473

12.01 0:08 +0.10 1.d4 Pf6 2.Pf3 d5 3.Lf4 Lf5 4.e3 e6
5.Le2 c6 6.O-O Ld6 7.Lxd6 (4.408.793) 496

13.01 0:22 +0.08 1.d4 d5 2.Lf4 Pf6 3.e3 e6 4.Pf3 Ld6
5.Lxd6 Dxd6 6.c3 Ld7 7.Ld3 O-O (11.441.164) 504

14.01 0:51 +0.08 1.d4 Pf6 2.c3 d5 3.Lf4 Lf5 4.e3 c6
5.Db3 b5 6.a4 Da5 7.Pf3 Lxb1 8.Txb1 (26.075.472) 506

14.02 1:45 +0.14 1.e4 d5 2.e5 c5 3.c3 Pc6 4.Pf3 Dc7
5.d4 Lg4 6.Lb5 cxd4 (54.021.480) 506

15.01 2:32 +0.16 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 c6
5.Pf3 Pf6 6.a3 Dd6 7.d3 Lg4 8.Lg5 Lxf3
9.Dxf3 De5+ 10.De3 (78.507.443) 510

16.01 3:44 +0.14 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 O-O-O 10.d5 (116.446.751) 513

17.01 11:49 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+ 4.De2 Dxe2+
5.Lxe2 e5 6.Pf3 Pc6 7.Lb5 Ld7 8.O-O f6
9.d4 Pxd4 10.Lxd7+ Kxd7 11.Td1 Td8 (367.226.210) 513

18.01 17:49 +0.20 1.e4 d5 2.exd5 Dxd5 3.Pc3 De6+
4.Pge2 Pf6 5.d4 Db6 6.Tb1 e6 7.Lg5 Da5
8.Lxf6 gxf6 9.Dd3 Pc6 10.Df3 Le7
11.a3 (548.708.509) 515

Colossus 2007b seems to have much smaller searchtrees per ply in this position, except at ply 18 where Colossus 2007b had to change its PV from the Scandinavian to the Scottish opening, that took a lot of time.
I tried it again with 128 Mb hash and now the differences seemed to be less but version 2007b was still faster.

I played no less than two Blitz games of the latest version against each other, and it may be that Colossus 2007b was just unlucky, but it lost both and fairly early after the opening... (opening book is Shredder9.bkt, Shredder UCI interface, 128 Mb hash and 6 minutes + 6 seconds Fischer bonus timecontrol on Athlon 2009 MHz) Could it be that Colossus 2007b is searching a bit too selectively at present? Most likely cause is just an statistically unlucky opening for Colossus 2007b. The variations and evals are very similar but 2007b just ended up in the wrong variations. Here are the first two games:

[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.06.01"]
[Round "1"]
[White "Colossus 2007b"]
[Black "Colossus 2007a"]
[Result "0-1"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. c4 {Boek 0s} e6 {Boek 0s}
3. Nc3 {Boek 0s} Bb4 {Boek 0s} 4. e3 {Boek 0s} c5 {Boek 0s}
5. Nge2 {Boek 0s} d5 {Boek 0s} 6. a3 {Boek 0s} Bxc3+ {Boek
0s} 7. Nxc3 {Boek 0s} cxd4 {Boek 0s} 8. exd4 {Boek 0s} dxc4
{Boek 0s} 9. Bxc4 {Boek 0s} Nc6 {Boek 0s} 10. Be3 {Boek 0s}
O-O {Boek 0s} 11. O-O {Boek 0s} b6 {Boek 0s} 12. Qf3 {Boek
0s} Bb7 {-0.02/13 20s} 13. Bd3 {Boek 0s} Rc8 {-0.04/12 23s}
14. Rad1 {Boek 0s} Na5 {-0.04/12 27s} 15. Qg3 {-0.04/13
16s} Nc4 {+0.04/13 18s} 16. Bxc4 {0.00/13 34s} Rxc4
{0.00/13 16s} 17. Rc1 {0.00/13 15s} Kh8 {0.00/13 22s}
18. Nb5 {0.00/13 22s} Bd5 {+0.04/13 20s} 19. Nxa7 {-0.04/12
15s} Ne4 {+0.22/14 18s} 20. Qf4 {-0.08/14 21s} Nd6
{+0.14/13 19s} 21. Bd2 {-0.08/13 23s} Re8 {+0.02/13 17s}
22. b3 {-0.28/13 25s} e5 {+0.02/14 15s} 23. bxc4 {-0.14/15
19s} exf4 {+0.38/15 15s} 24. cxd5 {-0.30/15 19s} Qg5
{+0.68/14 16s} 25. Nc6 {-0.74/13 18s} Qxd5 {+0.84/14 16s}
26. Bxf4 {-0.76/14 12s} Nc4 {+0.76/14 15s} 27. Ne5
{-0.76/15 18s} b5 {+0.78/14 15s} 28. Nxc4 {-0.86/14 21s}
bxc4 {+0.84/15 14s} 29. Be3 {-0.84/15 18s} f5 {+0.84/15
15s} 30. Rfe1 {-0.84/14 16s} h5 {+0.84/14 14s} 31. h3
{-1.00/13 18s} Kh7 {+0.98/14 16s} 32. Red1 {-1.02/14 11s}
Re4 {+1.32/15 13s} 33. g3 {-1.68/15 15s} Rxe3 {+1.92/16
12s} 34. fxe3 {-1.96/17 10s} Qf3 {+2.10/16 11s} 35. Re1
{-2.14/16 17s} Qxg3+ {+2.12/16 12s} 36. Kf1 {-2.16/16 13s}
Qxh3+ {+2.20/16 11s} 37. Ke2 {-2.20/16 17s} Qg2+ {+2.26/16
11s} 38. Kd1 {-2.20/4 0s} f4 {+2.26/17 11s} 39. Rc2
{-2.26/15 13s} Qe4 {+2.38/15 11s} 40. exf4 {-2.36/14 12s}
Qxd4+ {+2.58/15 11s} 41. Kc1 {-2.52/15 15s} Qxf4+ {+2.46/15
11s} 42. Kb2 {-2.52/16 12s} Qf3 {+2.46/14 11s} 43. a4
{-2.46/14 15s} Qb3+ {+2.64/13 11s} 44. Kc1 {-2.84/15 11s}
Qa3+ {+2.84/14 11s} 45. Kd2 {-2.58/15 10s} Qd3+ {+2.66/14
10s} 46. Kc1 {-2.58/4 0s} h4 {+2.52/14 8s} 47. Kb1
{-2.56/14 12s} Qb3+ {+2.96/14 10s} 48. Rb2 {-2.86/14 11s}
Qxa4 {+3.20/13 8s} 49. Rc1 {-2.78/14 13s} g5 {+2.80/13 11s}
50. Rc3 {-3.06/13 10s} Qd1+ {+2.98/14 8s} 51. Ka2 {-2.98/15
10s} Kg6 {+2.98/14 4s} 52. Rxc4 {-3.22/13 10s} Qd5
{+4.46/13 2s} 53. Kb3 {-4.46/12 2s} h3 {+5.90/14 10s}
54. Rbc2 {-5.58/13 11s} g4 {+6.04/11 1s} 55. Rh2 {-6.04/11
4s} Qb5+ {+6.22/11 9s} 0-1

[Event "6 minuten per partij + 6 seconden per zet"]
[Site "Engine Match"]
[Date "2007.06.01"]
[Round "2"]
[White "Colossus 2007a"]
[Black "Colossus 2007b"]
[Result "1-0"]

1. d4 {Boek 0s} Nf6 {Boek 0s} 2. c4 {Boek 0s} e6 {Boek 0s}
3. Nc3 {Boek 0s} Bb4 {Boek 0s} 4. e3 {Boek 0s} c5 {Boek 0s}
5. Nge2 {Boek 0s} d5 {Boek 0s} 6. a3 {Boek 0s} Bxc3+ {Boek
0s} 7. Nxc3 {Boek 0s} cxd4 {Boek 0s} 8. exd4 {Boek 0s} dxc4
{Boek 0s} 9. Bxc4 {Boek 0s} Nc6 {Boek 0s} 10. Be3 {Boek 0s}
O-O {Boek 0s} 11. O-O {Boek 0s} b6 {Boek 0s} 12. Qf3 {Boek
0s} Bb7 {0.00/13 20s} 13. Bd3 {Boek 0s} Rc8 {-0.02/12 19s}
14. Rad1 {Boek 0s} Qe7 {-0.04/12 27s} 15. Qg3 {+0.02/13
31s} Rfd8 {-0.02/12 21s} 16. Rfe1 {+0.04/12 18s} Rd7
{0.00/12 19s} 17. Be2 {+0.10/11 21s} Qd8 {-0.08/11 15s}
18. Bh6 {+0.46/13 26s} Qf8 {-0.30/14 25s} 19. Bg5 {+0.42/13
16s} Nd5 {-0.62/13 22s} 20. Nxd5 {+0.96/13 13s} Rxd5
{-0.96/13 13s} 21. Bf3 {+1.24/14 14s} Rd7 {-1.20/14 23s}
22. d5 {+1.24/14 13s} Nd8 {-1.24/13 14s} 23. dxe6 {+1.12/13
22s} Rxd1 {-1.06/13 14s} 24. exf7+ {+1.10/13 21s} Nxf7
{-1.12/15 16s} 25. Bxd1 {+1.10/14 15s} Nxg5 {-1.12/13 13s}
26. Qxg5 {+1.10/14 19s} Bc6 {-1.12/13 20s} 27. Bb3+
{+1.12/13 24s} Kh8 {-1.16/15 12s} 28. f3 {+1.12/14 18s} Re8
{-1.16/14 17s} 29. Rxe8 {+1.18/13 13s} Qxe8 {-1.18/16 13s}
30. Kf2 {+1.18/15 11s} Qf8 {-1.18/14 16s} 31. Qe5 {+1.18/14
17s} Qd8 {-1.20/14 14s} 32. Kg1 {+1.20/14 17s} Qe8
{-1.12/13 11s} 33. f4 {+1.14/15 18s} Qc8 {-1.12/13 9s}
34. Qe7 {+1.14/14 16s} Bb7 {-1.16/13 13s} 35. Kf2 {+1.14/14
15s} a5 {-1.14/12 10s} 36. g4 {+1.20/13 17s} Qb8 {-1.22/12
13s} 37. Ke1 {+1.22/13 14s} Bc6 {-1.44/13 14s} 38. f5
{+2.16/15 14s} Qe8 {-2.00/14 12s} 39. Qxe8+ {+2.14/18 12s}
Bxe8 {-2.00/4 0s} 40. g5 {+1.84/19 10s} Bh5 {-1.92/18 13s}
41. f6 {+2.16/19 11s} gxf6 {-1.24/19 11s} 42. gxf6
{+2.12/19 8s} h6 {-2.14/19 8s} 43. Be6 {+2.44/20 8s} Kh7
{-2.38/20 9s} 44. Kd2 {+2.46/21 11s} Kg6 {-2.42/20 13s}
45. f7 {+2.44/22 10s} Kg7 {-2.42/21 9s} 46. Kd3 {+2.44/21
7s} Bf3 {-2.42/20 11s} 47. Kc4 {+2.44/19 8s} Be2+ {-2.40/20
10s} 48. Kd5 {+2.42/20 10s} a4 {-2.42/20 8s} 49. Kc6
{+2.62/21 9s} b5 {-2.72/22 9s} 50. Kc5 {+2.80/21 11s} Kf8
{-2.86/21 8s} 51. Bd7 {+3.00/22 11s} Kxf7 {-3.04/21 9s}
52. Bxb5 {+3.22/21 9s} Bd1 {-3.66/21 7s} 53. Kb4 {+3.70/21
8s} Bg4 {-5.24/22 11s} 54. Bxa4 {+3.70/23 9s} Ke6 {-5.32/21
3s} 55. Kc5 {+4.72/18 3s} 1-0

Version 2007a was clearly better than Colossus 2006f, I am curious how it will turn out with the new 2007b! Kirill Kryukov just posted some early results for 2007a but now that version is already obsolete!

Eelco

[mclane]

is there any statement from you other than quoting ?

[Dr.Wael Deeb]

is there any statement from you other than quoting ?

Hi Thorsten,
This is a pure sample of spamming

[mariaclara]

.............. .....thanks Martin...............