Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rebel Century 4.0 - Crafty

Author: Slater Wold

Date: 21:19:46 11/03/01

Go up one level in this thread


On November 03, 2001 at 18:21:54, Uri Blass wrote:

>On November 03, 2001 at 18:06:36, Dan Andersson wrote:
>
>>I would say that the mistake was 31... Qxf4 The game seems terribly one sided
>>after that. Not that blacks position is that tasty before, except for white that
>>is:)
>>
>>MvH Dan Andersson
>
>I can say that I do not know what was the losing mistake and the only clear
>thing is that there is a drop in the evaluation of more than 1 pawn after Qg5
>
>The only logical alternative is Qh6 and we need to analyze the position after
>Qh6 in order to prove a clear advantage for white before claiming that Qxf4 was
>a mistake.
>
>I am also not sure that Qg5 is the losing mistake because I did not analyze it
>enough and evaluation of +1.xx is not a proof but I believe that black is losing
>after Qg5.
>
>Uri

I analyzed this position and game for some time.  It was a mixture of things.

Right after 33...Rff8 Michel disconnected from my computer.  (From what I hear,
the internet connection at the tournament hall is very bad.)  When Crafty made
the move 33...Rff8, it scored it 0.69 for white.  From the log:

White(33): Kh1 [pondering]
              time limit 1:27 (10:11)
         nss  depth   time  score   variation (13)
               14     3.37     --   33. ... Rff8
White(33): Kh1
predicted move made.
               14     2:46   0.69   33. ... Rff8 34. Rf3 Qg5 35. Rg3 Qh4
                                    36. Nxd7 Nxd7 37. Bxg6 Qf6 38. Bf5+
                                    Kh8 39. Qxf6+ Nxf6 40. Rc7 Rae8 41.
                                    h4 Re1+ 42. Kh2 a6
         (2)   14->   5:11   0.69   33. ... Rff8 34. Rf3 Qg5 35. Rg3 Qh4
                                    36. Nxd7 Nxd7 37. Bxg6 Qf6 38. Bf5+
                                    Kh8 39. Qxf6+ Nxf6 40. Rc7 Rae8 41.
                                    h4 Re1+ 42. Kh2 a6
              time=5:13  cpu=198%  mat=-1  n=463626612  fh=92%  nps=1479k
              ext-> chk=25658935 cap=1009410 pp=24540 1rep=2507646 mate=364712
              predicted=24  nodes=463626612  evals=85063890
              endgame tablebase-> probes done=0  successful=0
              hashing-> trans/ref=25%  pawn=99%  used=6%
              SMP->  split=631  stop=113  data=9/32  cpu=10:23  elap=5:13
              time used:   1:26


Afterwards, Rf3 was played, and from the log:

White(34): Rf3 [pondering]
              time limit 1:21 (9:31)
         nss  depth   time  score   variation (13)
               13    24.18   0.79   34. ... Qg5 35. Rg3 Qh4 36. Nxd7 Nxd7
                                    37. Bxg6 Qf6 38. Bf5+ Kh8 39. Qxf6+
                                    Nxf6 40. Rc7 Rae8 41. h4 Re5 42. Rf3
                                    a6
               13->  29.40   0.79   34. ... Qg5 35. Rg3 Qh4 36. Nxd7 Nxd7
                                    37. Bxg6 Qf6 38. Bf5+ Kh8 39. Qxf6+
                                    Nxf6 40. Rc7 Rae8 41. h4 Re5 42. Rf3
                                    a6
               14     1:35   0.94   34. ... Qg5 35. Rg3 Qh4 36. Bxg6 Ng4
                                    37. Nxg4 Bxg4 38. Bd3 Rac8 39. Rxc8
                                    Rxc8 40. Bf5 Qxg3 41. Bxh7+ Kxh7 42.
                                    hxg3 Rh8
               14->   1:50   0.94   34. ... Qg5 35. Rg3 Qh4 36. Bxg6 Ng4
                                    37. Nxg4 Bxg4 38. Bd3 Rac8 39. Rxc8
                                    Rxc8 40. Bf5 Qxg3 41. Bxh7+ Kxh7 42.
                                    hxg3 Rh8

And this is where Michel got disconnected, and had to use Bob's computer
instead.  Starting a fresh search is a bummer.  And this would probably be the
*worst* position in the world for this to happen.

Qg5 is the loser.  Qh6 shines a lot more hope on a draw.

BTW:  This was just a little test to see how long it would take Crafty to get
Qh6, with the hash cleared:

              clearing hash tables
              time surplus   0.00  time limit 60:00 (60:00)
         nss  depth   time  score   variation (1)
starting thread 1
                6     0.09     --   1. ... Qg5
                6     0.12   0.19   1. ... Qg5 2. Rcf1 Be8 3. Nxg6 Bxg6
                                    4. Bxg6 Qxg6 5. Rxf6
                6->   0.15   0.19   1. ... Qg5 2. Rcf1 Be8 3. Nxg6 Bxg6
                                    4. Bxg6 Qxg6 5. Rxf6
                7     0.28     ++   1. ... Qg5!!
                7->   0.43  -0.20   1. ... Qg5 2. Rcf1 Be8 3. Nxg6 Bxg6
                                    4. Bxg6 Qxg6 5. Rxf6
                8     0.71  -0.10   1. ... Qg5 2. Rcf1 Bg4 3. Nxg4 Nxg4
                                    4. Rxf8+ Rxf8 5. Rxf8+ Kxf8 6. Qh8+
                                    Ke7 7. Qxh7+ Kf6 8. Qh8+ Ke6 9. Qd4
                8->   0.79  -0.10   1. ... Qg5 2. Rcf1 Bg4 3. Nxg4 Nxg4
                                    4. Rxf8+ Rxf8 5. Rxf8+ Kxf8 6. Qh8+
                                    Ke7 7. Qxh7+ Kf6 8. Qh8+ Ke6 9. Qd4
                9     1.53   0.00   1. ... Qg5 2. Rcf1 Bg4 3. Nxg4 Nxg4
                                    4. Rxf8+ Rxf8 5. Rxf8+ Kxf8 6. Qh8+
                                    Ke7 7. Qxh7+ Kf6 8. Qh8+ Kf7 9. Qh7+
                                    Kf6
                9->   1.68   0.00   1. ... Qg5 2. Rcf1 Bg4 3. Nxg4 Nxg4
                                    4. Rxf8+ Rxf8 5. Rxf8+ Kxf8 6. Qh8+
                                    Ke7 7. Qxh7+ Kf6 8. Qh8+ Kf7 9. Qh7+
                                    Kf6
               10     8.23  -0.32   1. ... Qg5 2. Rc7 Ba4 3. Qd4 Rae8 4.
                                    Re3 a5 5. h4 Qh6 6. Rf3
               10->   8.50  -0.32   1. ... Qg5 2. Rc7 Ba4 3. Qd4 Rae8 4.
                                    Re3 a5 5. h4 Qh6 6. Rf3
               11    15.81     --   1. ... Qg5
               11    23.59   0.78   1. ... Qg5 2. Rg3 Bg4 3. h3 Qf4 4.
                                    Rxg4 Nxg4 5. Nxg4 Rad8 6. Rf1 Qg3 7.
                                    Rxf8+ Kxf8
               11->  24.54   0.78   1. ... Qg5 2. Rg3 Bg4 3. h3 Qf4 4.
                                    Rxg4 Nxg4 5. Nxg4 Rad8 6. Rf1 Qg3 7.
                                    Rxf8+ Kxf8
               12    34.20   0.96   1. ... Qg5 2. Rg3 Qh4 3. Nxd7 Nxd7
                                    4. Bxg6 Qf6 5. Qb1 Kh8 6. Rh3 Ne5 7.
                                    Rxh7+ Kg8 8. Bd3 a6
               12->  36.06   0.96   1. ... Qg5 2. Rg3 Qh4 3. Nxd7 Nxd7
                                    4. Bxg6 Qf6 5. Qb1 Kh8 6. Rh3 Ne5 7.
                                    Rxh7+ Kg8 8. Bd3 a6
               13     1:02   0.94   1. ... Qg5 2. Rg3 Qh4 3. Bxg6 Ng4 4.
                                    Nxg4 Bxg4 5. Bd3 Rfc8 6. Rxc8+ Rxc8
                                    7. Bf5 Qxg3 8. Bxh7+ Kxh7 9. hxg3 Rh8
               13->   1:07   0.94   1. ... Qg5 2. Rg3 Qh4 3. Bxg6 Ng4 4.
                                    Nxg4 Bxg4 5. Bd3 Rfc8 6. Rxc8+ Rxc8
                                    7. Bf5 Qxg3 8. Bxh7+ Kxh7 9. hxg3 Rh8
               14     2:00   1.22   1. ... Qg5 2. Rg3 Qh4 3. Bxg6 Ng4 4.
                                    Nxg4 Bxg4 5. Bd3 Rac8 6. Rxc8 Rxc8
                                    7. Bf5 Qxg3 8. Bxh7+ Kxh7 9. hxg3 Rh8
                                    <HT>
               14    16:31   1.08   1. ... Qh6 2. Rxf6 Rxf6 3. Nxd7 Rf2
                                    4. Qxf2 Qxc1+ 5. Qf1 Qxa3 6. Bb5 Qxb4
                                    7. Bc6 Rc8 8. Bxd5+ Kg7 9. Qa1+ Qc3
                                    10. Qxa7
               14->  16:31   1.08   1. ... Qh6 2. Rxf6 Rxf6 3. Nxd7 Rf2
                                    4. Qxf2 Qxc1+ 5. Qf1 Qxa3 6. Bb5 Qxb4
                                    7. Bc6 Rc8 8. Bxd5+ Kg7 9. Qa1+ Qc3
                                    10. Qxa7


Century is a tough opponent.  No one else at the tournament has beat it, other
than CT.  It played very well.  In such a tough game, the worse thing that could
possibly happen, is that the computer crashes (or in this case, you lose your
connection).

The Rebel programs have done extremely well.  I congratulate both Chris and Ed.



Slate



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.