Discussion of anything and everything relating to chess playing software and machines.
Moderators: hgm, Harvey Williamson, bob
Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
-
rvida
- Posts: 481
- Joined: Thu Apr 16, 2009 10:00 am
- Location: Slovakia, EU
Post
by rvida » Thu Mar 04, 2010 11:55 am
from a 40/2 match between Critter - Komodo:
[D]r4r2/1RPK2p1/5pkp/8/8/2R3PP/8/8 w - -
Critter played c8=N here and eventually won, but the move choice seems odd. It sees a better move only after half minute.
Code: Select all
2/ 3 00:00 592 0 +0.00 c8=Q Rf7+ Kd8
2/ 3 00:00 920 0 +3.04 Rc1 Rfc8
2/ 7 00:00 1223 0 +3.68 c8=N Kg5
3/ 9 00:00 1792 0 +3.91 c8=N Kf5 Nd6+ Ke5
4/10 00:00 4312 0 +4.00 c8=N f5 Nd6 f4
5/ 9 00:00 7069 0 +3.87 c8=N f5 Rc5 Ra3 Ne7+ Kg5
6/10 00:00 13121 820062 +3.92 c8=N f5 Rc6+ Kg5 Ne7 Rf7 Rcc7
7/13 00:00 26030 813437 +3.85 c8=N f5 Rc6+ Kh5 Nd6 f4 gxf4 Rxf4
8/13 00:00 44563 948148 +3.78 c8=N f5 Rc6+ Kh5 Rc5 g5 Ne7 f4 g4+ Kh4 Ng6+ Kxh3
9/15 00:00 96360 1025106 +3.76 c8=N f5 Rc6+ Kh5 Rb5 g6 Nd6 Rfb8 Rxb8 Rxb8 Rc2
10/21 00:00 900322 1129638 +3.91 c8=N Kg5 Rc5+ f5 Nd6 g6 Ke6 Ra4 Rb2 Rfa8 Nf7+ Kh5 Nd6
11/20 00:01 1966029 1143705 +3.92 c8=N Ra6 Nd6 Rfa8 Rcc7 Kg5 Rb5+ Kg6 h4 Ra3 h5+ Kh7 g4 Rd3
12/22 00:02 3336269 1147667 +3.89 c8=N Ra6 Ne7+ Kh7 Rcc7 Ra3 Nf5 Kg6 Ke6 Ra6+ Nd6 Re8+ Re7 Rxe7+ Kxe7 Ra5 Rc7
13/23 00:06 7836823 1152982 +3.84 c8=N h5 Rbc7 Ra5 Ne7+ Kh6 Kd6 Ra6+ Nc6 Re8 Rf3 h4 g4 Kg6 Rb3
14/23 00:08 9642466 1153405 +3.81 c8=N h5 Rbc7 Ra5 Ne7+ Kh6 Kd6 Ra6+ Nc6 Re8 Rf3 Re1 Kd5 Rd1+ Kc4 h4 g4
15/25 00:13 15915875 1158782 +3.86 c8=N h5 Rbc7 Ra5 Ne7+ Kh6 Rc3c5 Rxc5 Rxc5 g6 Rc6 Kg7 Nd5 h4 g4 f5 Rc4
16/28 00:32 37869983 1162975 +3.86 c8=N h5 Rbc7 Ra6 h4 Kh6 Ke7 Rh8 Nd6 Rb8 Rd3 Kg6 Kd7 Ra5 Rd1 Ra2
16/28+ 00:34 40981743 1174026 +4.25 Rb4
16/28 00:36 43111281 1180613 +4.51 Rb4 h5 c8=Q Rfxc8 Rxc8 Ra3 h4 Rxg3 Ke6 Re3+ Kd6 Rd3+ Ke7 Rg3 Rb7 Rg4 Kf8 Kf5 Rxg7
17/25 00:38 45900084 1187859 +4.47 Rb4 h5 c8=Q Rfxc8 Rxc8 Ra3 h4 Rxg3 Ke6 Re3+ Kd6 Kf5 Rc5+ Re5 Rxe5+ fxe5 Rb5 Kg4 Rxe5 Kxh4 Re4+ Kg5 Re7
Richard
-
hgm
- Posts: 22274
- Joined: Fri Mar 10, 2006 9:06 am
- Location: Amsterdam
- Full name: H G Muller
-
Contact:
Post
by hgm » Thu Mar 04, 2010 12:46 pm
If you promote to Q, he would play Rf7+, and have a perpetual on f7/f8 unless you abandon the Q. If you do the latter he would have either his two Rooks attacking the singly defended Q, or retaliate against Rb7 when you capture with the Rook. So that would just lose a Pawn.
By promoting to N he can interpose the Knight on e7 after Rf7+. In the short run that gains him a Knight, which might be valued more than a 7th-rank Pawn..
-
Sven
- Posts: 3576
- Joined: Thu May 15, 2008 7:57 pm
- Location: Berlin, Germany
Post
by Sven » Thu Mar 04, 2010 12:58 pm
rvida wrote:Critter played c8=N here and eventually won, but the move choice seems odd. It sees a better move only after half minute.
c8=N is not odd at all. It is the only promotion out of four possible ones that does not lead to a draw immediately. The reason is that all other promotions are answered by Rf7+, you can easily see what happens then. (In the meantime while I wrote this answer HGM has already explained it!

)
To find out that other moves, like Rb4 (or Rb1, or Re3) are even better than the winning move c8=N seems to require some deeper search, so it does not surprise me at all that Rb4 is not displayed as best move immediately. I analyzed the position a little bit with the help of free Rybka 2.2n2 32bit (I don't own Rybka 3) and found that after a move like Rb1, it takes about 18 plies until White can finally promote and win a Rook. That c8=N gives an advantage of N vs P can be seen much faster.
Sven
-
metax
- Posts: 344
- Joined: Wed Sep 23, 2009 3:56 pm
- Location: Germany
Post
by metax » Thu Mar 04, 2010 2:01 pm
This is not odd, but rather normal as HGM and Sven already explained. ChessMind has the same 'problem':
Code: Select all
FEN: r4r2/1RPK2p1/5pkp/8/8/2R3PP/8/8 w - - 0 1
ChessMind:
1/13 00:00 0 0 +2,75 c7c8S Ta8a6 Sc8e7+
1/13+ 00:00 0 0 +3,76 c7c8D
1/13 00:00 0 0 +4,26 c7c8D
2/13 00:00 9.238 73.000 -0,73 c7c8D Ta8xc8 Tc3xc8 Tf8f7+
2/13+ 00:00 9.238 73.000 +0,28 Tc3b3
2/14 00:00 9.238 73.000 +1,38 Tc3b3
2/18 00:00 9.238 73.000 +1,50 Tc3e3 Tf8f7+ Kd7c6 Tf7f8
3/18 00:00 40.899 174.000 +1,49 Tc3e3 h6h5 Tb7b4 Tf8f7+ Kd7c6 Ta8a6+
3/18+ 00:00 40.899 174.000 +2,50 c7c8S
3/18 00:00 40.899 174.000 +2,51 c7c8S
4/20 00:01 121.792 236.000 +2,64 c7c8S Ta8a5 Sc8e7+ Kg6f7 Kd7c6 Kf7e6 Tc3e3+
5/20 00:01 349.875 273.000 +2,58 c7c8S Ta8a5 Sc8e7+ Kg6f7 Kd7c6 Ta5e5 Se7d5+ Kf7g6 Sd5f4+
6/20 00:01 382.114 274.000 +2,33 c7c8S Ta8a5 Sc8e7+ Kg6f7 Tc3e3 Ta5e5 Se7d5 Kf7g6 Sd5f4+
7/23 00:02 575.140 277.000 +2,31 c7c8S Ta8a5 Sc8e7+ Kg6f7 Tc3e3 Ta5e5 Kd7c6 Te5e6+ Te3xe6 Kf7xe6
8/25 00:04 1.178.830 279.000 +2,38 c7c8S Ta8a5 Sc8e7+ Kg6h7 Tc3d3 Ta5e5 Td3d5 Tf8a8 h3h4 Ta8a4
9/28 00:08 2.301.869 282.000 +2,64 c7c8S Ta8a5 Sc8e7+ Kg6h7 Tc3d3 Ta5e5 h3h4 Tf8a8 Td3d2 Te5e4 Se7f5
10/30 00:19 5.551.590 283.000 +2,35 c7c8S Ta8a5 Sc8e7+ Kg6h7 Se7c6 Tf8f7+ Kd7d6 Tf7xb7 Sc6xa5 Tb7b5 Sa5c6 Kh7g6 Sc6e7+ Kg6g5 h3h4+
11/32 00:57 16.636.680 292.000 +2,44 c7c8S Ta8a5 Sc8e7+ Kg6h7 Tc3b3 Ta5e5 h3h4 Tf8a8 Se7c6 Te5e4 Tb3d3 Kh7g6 Kd7d6 f6f5
-
rvida
- Posts: 481
- Joined: Thu Apr 16, 2009 10:00 am
- Location: Slovakia, EU
Post
by rvida » Thu Mar 04, 2010 2:06 pm
hgm wrote:If you promote to Q, he would play Rf7+, and have a perpetual on f7/f8 unless you abandon the Q. If you do the latter he would have either his two Rooks attacking the singly defended Q, or retaliate against Rb7 when you capture with the Rook. So that would just lose a Pawn.
Thanks, now I understand.
hgm wrote:
In the short run that gains him a Knight, which might be valued more than a 7th-rank Pawn..
Indeed, this is the cause. The passed pawn is valued only 3.26. After promotion it is: -1.00 for loosing a pawn, +3.25 for gaining a knight, +0.87 bonus for being a minor piece up for a total of 3.12. After the knight gains some mobility and threatens black king, its value outweights the value of the passer.
After tweaking/increasing passed pawn evaluation now it switches to a different move much faster:
Code: Select all
2/ 4 00:00 322 20125 +0.00 c8=Q Rf7+ Kd8
2/ 4 00:00 748 46750 +3.41 Rd3 Rfc8
2/ 8 00:00 1479 92437 +3.82 c8=N Kg5
3/ 6 00:00 1819 113687 +3.91 c8=N Kg5 Nd6
4/ 7 00:00 2690 168125 +3.91 c8=N Kg5 Nd6 g6
5/12 00:00 4890 305625 +3.91 c8=N Kg5 Nd6 g6 Ke6
6/10 00:00 8004 258193 +3.90 c8=N Kg5 Nd6 g6 Ke6 f5
7/16 00:00 20607 438446 +3.92 c8=N Kg5 Nd6 g6 Rc5+ f5 Ke7 Kh5 Ke6
8/14 00:00 44318 703460 +3.83 c8=N f5 Ne7+ Kg5 Rc6 Rf7 Rg6+ Kh5 Rd6 g5 Kc6
9/16 00:00 76344 694036 +3.91 c8=N f5 Ne7+ Kg5 Rc6 Rf7 Rg6+ Kh5 Rd6 Rff8 Kc6 g5
10/17 00:00 150222 799053 +3.92 c8=N f5 Ne7+ Kg5 Rc6 Rf7 Rg6+ Kh5 Rd6 g6 Ke6 Rff8 Ke5
11/19 00:00 335162 857191 +3.90 c8=N f5 Ne7+ Kg5 Rc6 Rf7 Rg6+ Kh5 Rd6 g6 Ke6 Rff8 Ke5 Ra5+ Nd5
12/22 00:00 664876 886501 +3.90 c8=N f5 Ne7+ Kg5 Rc6 Rf7 Rd6 Ra3 Ke8 Rf6 h4+ Kg4 Rxf6 gxf6 Rb4+ f4
13/26 00:03 3552905 924272 +3.84 c8=N Ra1 Ne7+ Kh7 Kc6 Rfa8 Nd5 Rc8+ Kd7 Rxc3 Nxc3 g5 Nd5 Kg6 g4 h5
14/27 00:05 4904319 917724 +3.88 c8=N Ra1 Ne7+ Kh7 Kc6 Rfa8 Nd5 Rc8+ Kd7 Rxc3 Nxc3 g5 Nd5 Kg6 Rb6 Ra7+ Ke6 Ra3
15/27 00:07 6583808 920040 +3.89 c8=N Ra1 Ne7+ Kh7 Kd6 Rfa8 g4 Ra8a7 Rcc7 Rd1+ Nd5 Ra6+ Rc6 Ra8 Rc5 Kg6 Rcc7 Rd8+ Rd7
15/27+ 00:07 7116360 925645 +4.27 Rb6
15/27+ 00:08 8055360 933954 +4.66 Rb6
15/27 00:09 8579389 937023 +4.55 Rb6 h5 c8=Q Rfxc8 Rxc8 Ra3 Rf8 Kh7 Ke6 Rxg3 Rb7 Re3+ Kd5 Kg6 Rff7 Kf5 Rxg7
16/24 00:09 8978672 938896 +4.48 Rb6 h5 c8=Q Rfxc8 Rxc8 Ra3 Rf8 Kh7 Ke6 Rxg3 Rb7 Re3+ Kd5 Kg6 Rff7 Rxh3 Rxg7+ Kf5 Kd4
17/25 00:11 10491637 948353 +4.48 Rb6 h5 c8=Q Rfxc8 Rxc8 Ra3 Rf8 Kh7 Ke6 Rxg3 Rb7 Re3+ Kd6 Rd3+ Kc5 Kg6 Rff7 Rxh3 Rxg7+ Kf5 Kd4
18/31 00:13 12574484 958055 +4.54 Rb6 h5 c8=Q Rfxc8 Rxc8 Ra3 Ke7 Rxg3 Rxf6+ Kh7 Rd6 Rxh3 Rc5 Re3+ Kd7 Rh3 Rc7 Ra3 Ke6 h4 Rdd7 h3
19/31 00:16 16370262 964715 +4.57 Rb6 h5 c8=Q Rfxc8 Rxc8 Ra3 Ke7 Rxg3 Rxf6+ Kh7 Rf7 Rxh3 Rc6 Rd3 Ke6 h4 Rcc7 Re3+ Kd6 h3 Rxg7+ Kh6 Rg4 Kh5
19/31+ 00:23 22500791 978295 +4.93 Rb1
19/31 00:28 27882371 976718 +4.71 Rb1 Ra7 Re1 Rfa8 Re8 Rxc7+ Rxc7 Ra3 g4 Rxh3 Rg8 Kh7 Rf8 h5 gxh5 Rxh5 Rb8 Kg6 Kd6 Ra5
19/31+ 00:39 39082014 989793 +5.32 Rb2
-
Sven
- Posts: 3576
- Joined: Thu May 15, 2008 7:57 pm
- Location: Berlin, Germany
Post
by Sven » Thu Mar 04, 2010 9:35 pm
rvida wrote:hgm wrote:
In the short run that gains him a Knight, which might be valued more than a 7th-rank Pawn..
Indeed, this is the cause. The passed pawn is valued only 3.26. After promotion it is: -1.00 for loosing a pawn, +3.25 for gaining a knight, +0.87 bonus for being a minor piece up for a total of 3.12. After the knight gains some mobility and threatens black king, its value outweights the value of the passer.
After tweaking/increasing passed pawn evaluation now it switches to a different move much faster:
You can easily and quickly go into a wrong direction by
a) tuning evaluation based on one single position,
b) trying to increase passed pawn eval such that the engine prefers "infinitely" having a 7th rank passer on the board over actually promoting it.
It may be right that underpromotion to a knight is not the major goal of having a passer since usually it will be promoted to a queen. But I doubt that a very high static bonus for 7th rank passers is always adequate, since you really don't know whether the pawn can promote and win a huge amount of material or not. To decide about this is more tactics than positional evaluation IMO, so the search should find that out.
Sven