endgame: passed pawns vs. pieces

[PauloSoare]

Thanks, Eelco, but you are my friend, so....
I think if white can draw, Bh4 is less complex,
Jon is right.

Paulo Soares

[Eelco de Groot]

Dialogue between Leonard Nimoy as Spock and DeForest Kelley as Leonard H. "Bones" McCoy in Star Trek IV - The Voyage Home:

(As Spock looks baffled, Bones ENTERS SHOT.)

BONES
You, ah...
(diplomatically)
You present the appearance of a man
with a problem.

SPOCK
Your perception is correct, Doctor...
In order to return us to the exact
moment at which we left the 23rd
Century, I have used our journey back
through time as a referent,
calculating the coefficient of
elapsed time in relation to the
acceleration curve.

BONES
Naturally.
(then)
So what is your problem?

SPOCK
Acceleration is no longer a constant.

BONES
Well, then you're just gonna have to take your
best shot.

SPOCK
... Best shot...?

BONES
Yes, Spock! Your best guess.

SPOCK
"Guessing" is not in my nature Doctor...

BONES
(has waited 10
years for this)
Well nobody's perfect...

I must admit that 50% of the time I think that Toga Mara is just a big joke and the other 50% I think it is just a big joke but maybe there is something there if we build it ourselves. At the moment despite the general interest I think I may be the only one who is actually testing something because nobody else has the time for it or if they do there is only a fragment of code and on our small forum not many have any compiling capacity to work with that and our resident compiling expert has no computer

As the actual programmer does not seem to be about at the moment, most other programmers are still discussing Rybka 1.0 Beta and myself I have not touched the Toga code in months and have forgotten how it even looks like, I think we are progressing on a glacial scale here. But even if what is there at the moment holds only the barest glimmer of hope it may be important.

I can only offer at the moment a result from Toga Mara Beta in Walter Eigenmann's SwissTest 4 with two different values for "Tactical" and the only thing that is really interesting is that if this variable would not be doing anything, then I would expect the resultmatrices to be almost identical, but the funny thing is they are not; both versions have different positions where either one version or the other is doing better. A pity is that the version that did worse, with value = 20, actually seemed to have more "chess knowledge" when staring at the output than the default value = 30, but maybe the change resulted in too long computations so the result suffered... anything is possible. Also I made a lot of other more or less random changes in the code that we have so maybe this does not apply to the actual Toga Mara anymore.

Toga Mara Beta SwissTest 4 Results

Code: Select all

Result with default Tactical = 30, Athlon 2009 MHz, 64 MB hashtables, 60 seconds per position, Shredder 9 interface:

Right until now: 40 of 64  ;  32:15m
I/O Fault!

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   -   0   -   -  24   0   0   -   0  31  39   -   0   8   0   -   -   -  38   -
  20 |   -   0   -   0   -   0   0   0   0   -   2   0   2  50  46  20   -   0   0   -
  40 |  19  50   -   3   -   -   -  54   -   0   0   0   -   3  17  18   0   -   -   3
  60 |  34  23   -   0

   1 sec ->  20/64
   2 sec ->  20/64
   3 sec ->  22/64
   4 sec ->  25/64
   5 sec ->  25/64
   6 sec ->  25/64
   7 sec ->  25/64
   8 sec ->  25/64
   9 sec ->  26/64
  10 sec ->  26/64
  11 sec ->  26/64
  12 sec ->  26/64
  13 sec ->  26/64
  14 sec ->  26/64
  15 sec ->  26/64
  20 sec ->  29/64
  25 sec ->  32/64
  30 sec ->  32/64
  40 sec ->  36/64
  50 sec ->  37/64
  60 sec ->  40/64
  n/s: 694.524  
  TotTime: 64:02m    SolTime: 32:15m

Result with Tactical = 20:

Right until now: 36 of 64  ;  34:23m
I/O Fault!

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   -   0   4   -   -   0   0   -   0  42   -   -   6  52   0  40   -  51   0   -
  20 |   -   0   -   0   -   0   0   0   0   -   0   -  39   -   -   -   -  14   0  46
  40 |  20   3   -   -   -   -   -   0   0   0   0   0   -   3  44   2   0   -   -   1
  60 |   -   2   -   -

   1 sec ->  20/64
   2 sec ->  21/64
   3 sec ->  23/64
   4 sec ->  25/64
   5 sec ->  26/64
   6 sec ->  26/64
   7 sec ->  27/64
   8 sec ->  27/64
   9 sec ->  27/64
  10 sec ->  27/64
  11 sec ->  27/64
  12 sec ->  27/64
  13 sec ->  27/64
  14 sec ->  27/64
  15 sec ->  28/64
  20 sec ->  28/64
  25 sec ->  29/64
  30 sec ->  29/64
  40 sec ->  30/64
  50 sec ->  34/64
  60 sec ->  36/64
  n/s: 695.933  
  TotTime: 64:02m    SolTime: 34:23m

The most interesting thing about this is that the resultmatrices are dissimilar.
40 solutions is also what Toga 1.3.1 is getting but on a faster Core 2 Duo, see http://forum.computerschach.de/cgi-bin/ ... .pl?tid=20 in the renewed CSS forum. Congratulations guys with the transition from the Parsimony Forum! It will take some getting used to the new look just as in the CCC make-over but I think there are advantages too!

Regards, Eelco

[Eelco de Groot]

Just a couple of games just played with experimental Toga Mara code, not much better or much worse I think than standard Toga, but it is hard to judge this because only very few games were played. I also posted this in the Toga forum but most can not read there, so crossposting for once:

Edit: the discussion, a reply from me to a post from Zach that people who improve Toga substantially would probably post their creations is maybe a bit relevant so I crosspost that part also to add to other 55+ pages of posts about certain subject.

Just for the sake of argument; I think there might be quite a few programmers out there who are trying their ideas in a well tested and strong program like Toga. Some may have improved Toga but why would that mean they publish their code? They just run the risk of getting accused of making clone programs whatever programs they make after. So that is no argument.

It does not make any sense of course to have huge material values for Kings that would only make some difference if one side had two or more of them... Obviously not necessary in chess. But it is not proven that this material.cpp is not going to work no matter what. I made some changes, played a couple of bad games yesterday with it but just finished a small match against an older Belka that seemed okay:

[Event "12 Minutes/Game + 3 Seconds/Move"]
[Site "Engine Match"]
[Date "2008.08.26"]
[Round "1"]
[White "Belka 1.8.11"]
[Black "Toga Mara Beta 1"]
[Result "1/2-1/2"]
[ECO "B46 - Sicilian/Taimanov Variation"]

1. e4 {book 0s} c5 {book 0s} 2. Nf3 {book 0s} e6 {book 0s}
3. d4 {book 0s} cxd4 {-0.33/11 18s} 4. Nxd4 {book 0s} a6
{book 0s} 5. Nc3 {book 0s} Nc6 {book 0s} 6. Nxc6 {book 0s}
bxc6 {book 0s} 7. Bd3 {book 0s} Bd6 {book 0s} 8. O-O
{+0.57/13 17s} Nf6 {book 0s} 9. Qe2 {+0.48/14 24s} Qc7
{book 0s} 10. f4 {+0.42/15 34s} e5 {book 0s} 11. f5
{+0.36/14 21s} h6 {-0.70/13 37s} 12. Be3 {+0.49/14 21s} O-O
{-0.65/13 50s} 13. Qf3 {+0.46/13 36s} Bb4 {-0.59/13 39s}
14. Qg3 {+0.45/13 22s} Kh8 {-0.59/12 13s} 15. Qh4 {+0.60/13
43s} Qd8 {-0.65/12 29s} 16. Rf3 {+0.72/12 14s} Nh7
{-0.74/13 21s} 17. Qg3 {+0.72/13 28s} Bxc3 {-0.50/12 14s}
18. bxc3 {+0.87/13 7s} d6 {-0.41/12 13s} 19. Rb1 {+0.70/13
21s} Qa5 {-0.41/12 17s} 20. c4 {+1.10/12 13s} c5 {-0.34/12
11s} 21. f6 {+1.48/13 10s} g5 {+0.18/13 14s} 22. h4
{+1.03/13 15s} Rg8 {+0.25/13 18s} 23. Rf2 {+0.68/13 30s}
Qxa2 {+0.43/12 18s} 24. Rff1 {+0.76/14 23s} Qa5 {+0.76/13
18s} 25. hxg5 {+1.00/13 12s} hxg5 {0.00/13 53s} 26. Qh2
{+0.97/13 11s} Rg6 {+0.15/13 21s} 27. g3 {+1.21/14 8s} Qd8
{-0.04/14 15s} 28. Kg2 {+1.36/15 8s} a5 {+0.03/13 10s}
29. Rh1 {+1.89/15 8s} Qg8 {+0.12/13 3s} 30. Qh5 {+2.19/16
8s} Rxf6 {-2.03/13 9s} 31. Rh2 {+2.17/16 11s} Rg6 {-1.77/13
15s} 32. Rbh1 {+2.08/16 20s} a4 {-1.74/12 10s} 33. Qxh7+
{+2.07/15 15s} Qxh7 {-1.76/4 0s} 34. Rxh7+ {+2.06/14 9s}
Kg8 {-1.76/4 0s} 35. Bc1 {+2.06/16 25s} Bb7 {-1.65/12 9s}
36. g4 {+2.27/13 11s} Ra7 {-1.59/12 11s} 37. Rh8+ {+2.32/13
7s} Kg7 {-1.66/4 0s} 38. Ba3 {+2.19/14 12s} Bc6 {-1.48/13
17s} 39. Kg3 {+2.14/15 15s} Bd7 {-1.46/13 17s} 40. R1h7+
{+2.37/15 11s} Kf6 {-1.47/4 0s} 41. Rb8 {+2.14/16 43s} Ke7
{-1.47/14 12s} 42. Rh1 {+2.10/15 15s} Rg7 {-1.40/13 19s}
43. Rb6 {+2.35/14 5s} Ra8 {-1.42/13 12s} 44. Be2 {+2.29/15
9s} Rc8 {-1.33/13 11s} 45. Ra6 {+2.19/15 14s} Ke6 {-1.67/13
11s} 46. Rb1 {+2.21/15 12s} Bc6 {-1.88/13 10s} 47. Rbb6
{+2.58/16 5s} Kd7 {-1.70/13 8s} 48. c3 {+2.59/16 4s} Rg6
{-1.79/13 8s} 49. Bf3 {+2.57/15 5s} Rc7 {-1.93/14 10s}
50. Bc1 {+2.69/16 4s} Kc8 {-2.09/14 10s} 51. Be3 {+2.61/16
14s} Kd7 {-2.34/14 11s} 52. Bd2 {+2.62/16 5s} Rc8 {-2.32/14
11s} 53. Kf2 {+2.75/16 11s} Rc7 {-2.34/14 13s} 54. Ke2
{+2.64/16 5s} f6 {-2.31/13 8s} 55. Kd3 {+2.69/16 4s} Rh6
{-2.24/13 5s} 56. Bd1 {+2.73/15 4s} Bb7 {-2.19/12 8s}
57. Rxa4 {+2.77/13 2s} Ke7 {-2.10/13 8s} 58. Bf3 {+2.72/14
5s} Bc8 {-1.86/13 8s} 59. Rb8 {+2.61/15 9s} Rh3 {-1.26/14
8s} 60. Ke2 {+2.32/16 6s} Rh2+ {-1.35/14 8s} 61. Ke3
{+2.22/16 3s} Rh3 {-1.35/14 5s} 62. Kf2 {+2.07/16 12s} Rh2+
{-1.33/13 5s} 63. Bg2 {+2.05/16 5s} Bxg4 {-1.43/13 6s}
64. Kg1 {+2.05/15 3s} Rh4 {-1.22/13 9s} 65. Be1 {+2.10/15
7s} Rh6 {-1.24/13 7s} 66. Ra2 {+2.06/14 15s} Kd7 {-1.29/12
9s} 67. Bg3 {+2.24/14 8s} Ke6 {-1.30/12 6s} 68. Bf2
{+2.12/14 10s} Ke7 {-1.27/12 6s} 69. Ra6 {+2.14/13 5s} Be2
{-1.05/13 8s} 70. Ra4 {+1.80/13 2s} Bd1 {-1.10/13 4s}
71. Raa8 {+1.76/14 3s} Be2 {-1.11/13 6s} 72. Re8+ {+1.71/13
4s} Kf7 {-1.10/13 6s} 73. Rf8+ {+1.67/14 12s} Ke6 {-1.04/13
6s} 74. Ra6 {+1.59/12 8s} Ke7 {-0.86/11 4s} 75. Rg8
{+1.69/14 3s} Rh7 {-1.23/13 6s} 76. Ra4 {+1.84/12 1s} Rb7
{-0.78/12 4s} 77. Be1 {+1.50/13 5s} Ke6 {-0.81/12 6s}
78. Kf2 {+1.47/13 6s} Rb2 {-0.85/11 4s} 79. Kg3 {+1.29/12
2s} Bd3 {-0.84/11 4s} 80. Re8+ {+1.31/12 4s} Kd7 {-0.71/13
6s} 81. Rf8 {+1.02/13 4s} Ke6 {-0.73/11 5s} 82. Ra6
{+1.07/13 11s} Bxc4 {-0.51/10 7s} 83. Raa8 {+1.02/13 5s}
Bb5 {-0.55/11 5s} 84. Bf3 {+1.01/13 4s} Be2 {-0.62/12 9s}
85. Rae8+ {+1.01/13 6s} Kd7 {-0.56/13 11s} 86. Bg2
{+0.74/13 5s} Rh4 {0.00/10 3s} 87. Ra8 {+0.67/12 2s} Bd3
{-0.13/10 3s} 88. Ra7+ {+0.59/11 1s} Kc6 {0.00/11 2s}
89. Rc8+ {+0.31/12 5s} Kb6 {0.00/11 1s} 90. Rg7 {+0.53/13
5s} Re2 {-0.06/11 5s} 91. Bf2 {+0.53/13 4s} Rc2 {-0.20/11
4s} 92. Rd7 {+0.20/13 11s} Rxc3 {-0.01/11 6s} 93. Rxd6+
{+0.19/13 4s} Kb7 {0.00/12 3s} 94. Rxc5 {+0.21/11 1s} Rb3
{0.00/12 4s} 95. Rd7+ {+0.01/12 3s} Kb8 {0.00/12 4s}
96. Rdc7 {+0.03/13 4s} Bxe4+ {+0.13/11 4s} 97. Rc3
{-0.03/13 1s} Rxc3+ {0.00/11 4s} 98. Rxc3 {-0.03/14 1s}
Bxg2 {0.00/12 4s} 99. Kxg2 {+0.03/12 0s} Ra4 {0.00/12 8s}
100. Rc6 {+0.03/13 2s} f5 {0.00/11 4s} 101. Rc5 {-0.06/13
3s} Rg4+ {0.00/12 3s} 102. Kf3 {-0.23/15 8s} Re4 {0.00/13
5s} 103. Rc2 {-0.22/14 5s} Kb7 {+0.26/11 3s} 104. Kg2
{-0.30/14 7s} Rg4+ {+0.38/12 4s} 105. Kh3 {-0.21/15 5s} Rf4
{+0.30/12 3s} 106. Kg2 {-0.32/14 3s} Re4 {+0.15/11 6s}
107. Kh3 {-0.16/14 3s} Ka6 {0.00/11 3s} 108. Rb2 {-0.18/14
6s} Rf4 {+0.36/11 2s} 109. Kg2 {-0.22/14 3s} Ka5 {+0.15/11
9s} 110. Bb6+ {-0.09/13 3s} Ka4 {+0.18/12 4s} 111. Bc7
{-0.08/14 3s} Rg4+ {0.00/11 2s} 112. Kf3 {-0.12/14 2s} Re4
{0.00/11 3s} 113. Bd6 {-0.05/14 2s} g4+ {0.00/12 3s}
114. Kf2 {0.00/17 2s} f4 {0.00/11 2s} 115. Bc5 {0.00/17 2s}
Ka5 {0.00/12 2s} 116. Bb6+ {-0.03/15 4s} Ka6 {0.00/13 3s}
117. Bc7 {0.00/16 3s} g3+ {0.00/13 3s} 118. Kf3 {0.00/15
2s} Re1 {0.00/13 3s} 119. Kg4 {+0.01/15 4s} Re3 {0.00/13
3s} 120. Rb6+ {+0.02/15 2s} Ka7 {0.00/14 3s} 121. Rb1
{0.00/16 3s} Re2 {0.00/13 3s} 122. Bxe5 {+0.02/15 2s} Rxe5
{0.00/14 2s} 123. Kxf4 {+0.02/16 0s} Re2 {0.00/23 2s}
124. Kxg3 {+0.03/17 2s} Ra2 {0.00/25 2s} 125. Kf3 {+0.09/16
2s} Ka6 {0.00/26 2s} 126. Ke3 {+0.07/18 6s} Ka5 {0.00/27
2s} 127. Kd4 {+0.08/17 2s} Ka6 {0.00/27 4s} 128. Kc5
{+0.08/17 2s} Ka7 {0.00/25 2s} 129. Kc6 {+0.09/16 1s} Ka6
{0.00/26 2s} 130. Rb8 {+0.08/18 3s} Ka7 {0.00/24 2s}
131. Re8 {+0.08/19 2s} Rf2 {0.00/21 2s} 132. Rd8 {+0.08/17
1s} Rf1 {0.00/19 2s} 133. Rd2 {+0.08/18 4s} Kb8 {0.00/23
3s} 134. Rg2 {+0.08/16 2s} Rc1+ {0.00/22 3s} 135. Kd5
{+0.07/19 6s} Kb7 {0.00/20 2s} 136. Rg6 {+0.08/18 4s} Rf1
{0.00/23 3s} 137. Kc5 {+0.08/17 2s} Rf2 {0.00/20 2s}
138. Rb6+ {+0.08/17 2s} Ka7 {0.00/23 2s} 139. Rd6 {+0.08/18
1s} Rf1 {0.00/18 2s} 140. Rg6 {+0.08/19 4s} Rf2 {0.00/22
3s} 141. Rg3 {+0.08/17 1s} Rf1 {0.00/18 2s} 142. Rb3
{+0.08/20 2s} Rf2 {0.00/23 2s} 143. Kd5 {+0.08/19 1s} Rc2
{0.00/26 3s} 144. Kd6 {+0.08/18 5s} Ka6 {0.00/27 3s}
145. Rb1 {+0.08/20 2s} Rf2 {0.00/26 4s} 146. Ke6 {+0.08/17
1s} Ka7 {0.00/27 4s} 147. Ke5 {+0.08/19 4s} Ka6 {0.00/27
3s} 148. Kd5 {+0.08/18 3s} Ka7 {0.00/27 4s} 149. Rb4
{+0.08/17 3s} Rc2 {0.00/26 3s} 150. Kd6 {+0.08/18 1s} Ka6
{0.00/21 3s} 151. Rb3 {+0.08/18 2s} Rf2 {0.00/27 3s}
152. Rb8 {+0.08/17 2s} Rf1 {0.00/19 2s} 153. Ke5 {+0.08/18
4s} Rf2 {0.00/24 3s} 154. Ke6 {+0.08/17 3s} Rf1 {0.00/19
2s} 155. Kd5 {+0.08/19 3s} Rf2 {0.00/26 5s} 156. Rb3
{+0.08/16 2s} Ka7 {0.00/26 3s} 157. Kc6 {+0.08/19 2s} Rf1
{0.00/25 3s} 158. Ra3+ {+0.08/18 3s} Kb8 {0.00/4 0s}
159. Kd5 {+0.08/16 3s} Rf2 {0.00/19 2s} 160. Re3 {+0.08/16
3s} Ka7 {0.00/17 3s} 161. Re6 {+0.08/17 2s} Kb7 {0.00/17
3s} 162. Kc4 {+0.07/16 2s} Ka7 {0.00/17 2s} 163. Re8
{+0.08/16 3s} Rf1 {0.00/18 4s} 164. Kd5 {+0.08/16 1s} Rg1
{0.00/19 3s} 165. Re6 {+0.08/17 3s} Kb7 {0.00/17 2s}
166. Rh6 {+0.08/17 4s} Rf1 {0.00/17 3s} 167. Kd4 {+0.07/15
2s} Rf2 {0.00/22 2s} 168. Ke5 {+0.07/14 3s} Ka7 {0.00/27
2s} 169. Rd6 {+0.07/13 3s} Rf1 {0.00/20 3s} 170. Rg6
{+0.03/13 3s} Rh1 {0.00/27 2s} 171. Kf4 {0.00/13 2s} Rb1
{0.00/30 4s} 172. Rh6 {0.00/15 2s} Kb7 {0.00/31 2s}
173. Rd6 {0.00/16 3s} Ka7 {0.00/20 3s} 174. Kf5 {0.00/18
2s} 1/2-1/2

[Event "12 Minutes/Game + 3 Seconds/Move"]
[Site "Engine Match"]
[Date "2008.08.26"]
[Round "2"]
[White "Toga Mara Beta 1"]
[Black "Belka 1.8.11"]
[Result "1-0"]
[ECO "B46 - Sicilian/Taimanov Variation"]

1. e4 {book 0s} c5 {book 0s} 2. Nf3 {book 0s} e6 {book 0s}
3. d4 {book 0s} cxd4 {-0.32/14 23s} 4. Nxd4 {book 0s} a6
{book 0s} 5. Nc3 {book 0s} Nc6 {book 0s} 6. Nxc6 {book 0s}
bxc6 {book 0s} 7. Bd3 {book 0s} Bd6 {book 0s} 8. O-O
{+0.73/12 29s} Nf6 {book 0s} 9. Kh1 {+0.73/13 25s} O-O
{-0.34/13 15s} 10. f4 {+0.59/13 31s} e5 {-0.11/15 56s}
11. g4 {+0.73/13 38s} exf4 {+0.27/15 30s} 12. g5 {+0.59/13
15s} Ne8 {+0.21/14 5s} 13. Qh5 {+0.75/13 17s} Rb8 {0.00/16
2:16m} 14. Bxf4 {+0.73/12 24s} g6 {-0.14/16 2:21m} 15. Qh6
{+0.79/13 16s} Qe7 {-0.13/15 1:08m} 16. Bxd6 {+0.81/13 20s}
Nxd6 {-0.16/15 20s} 17. Rf6 {+0.69/13 28s} Rxb2 {+0.30/13
12s} 18. Raf1 {+0.69/13 18s} Qe5 {-0.62/13 21s} 19. R1f3
{+1.18/12 12s} Re8 {-1.01/13 19s} 20. Rh3 {+1.14/12 18s}
Re6 {-0.86/12 8s} 21. Qxh7+ {+1.25/11 14s} Kf8 {-0.92/13
10s} 22. Qh8+ {+1.71/12 22s} Ke7 {-0.86/14 16s} 23. Rxe6+
{+1.71/13 17s} Qxe6 {-0.97/15 17s} 24. Re3 {+1.59/13 13s}
Rb4 {-1.07/13 17s} 25. a3 {+1.58/12 11s} Rb8 {-1.20/13 10s}
26. Rf3 {+1.72/12 14s} Ra8 {-1.35/13 7s} 27. e5 {+1.61/13
53s} Ne8 {-1.81/14 19s} 28. Rf6 {+1.86/12 12s} Qg4
{-2.33/14 14s} 29. Ne4 {+2.77/11 11s} Qh5 {-2.09/13 5s}
30. Qg8 {+3.04/11 8s} Kd8 {-6.27/13 33s} 31. Nd6 {+6.62/11
13s} Rb8 {-6.48/13 19s} 32. Qxe8+ {+7.22/10 13s} Kc7
{-6.29/14 16s} 33. Rxf7 {+7.10/11 12s} c5 {-6.45/13 10s}
34. Nxc8 {+7.55/10 8s} Qg4 {-6.79/13 8s} 35. Rf1 {+7.81/12
12s} Rxc8 {-6.97/13 2s} 36. Qxg6 {+9.13/11 5s} Qd4
{-7.81/13 8s} 37. Rb1 {+11.24/12 12s} c4 {-12.14/15 48s}
38. c3 {+15.90/12 12s} Qf2 {-13.03/12 3s} 39. Be4 {+26.34/9
9s} Kd8 {-20.67/12 3s} 40. Qg8+ {+M12/8 5s} Ke7 {-27.72/12
2s} 41. Qxc8 {+M8/9 4s} Qe3 {-M6/8 0s} 42. Qxc4 {+M7/10 9s}
Qxg5 {0.00/5 0s} 43. Qc5+ {+M6/10 3s} Kf7 {0.00/4 0s}
44. Rf1+ {+M5/16 2s} Qf5 {-M3/3 0s} 45. Rxf5+ {+M4/63 0s}
Kg7 {-M2/2 0s} 46. Qe7+ {+M3/63 0s} Kh8 {-M1/1 0s} 47. Rf8#
{+M1/63 0s} 1-0


Belka 1.8.11 - Toga Mara Beta/Toga II Checkov 1.4 Beta_4: ½ - 1½

The material values I used in this version are more standardized values for all the pieces, but to have more flexibility in values I changed the value for a pawn in the endgame to 0x100 like Tord does in Glaurung and MaterialWeight is decreased from 256 to 100 so that more or less cancels out. My latest experimental material.cpp is attached below. Silly things like counting number of Kings I did not change here, that should not hurt the performance. Some other values changed at random, it would be difficult to get good values for all the variables.

Eelco