Author: Tony Werten
Date: 01:11:12 08/11/02
Go up one level in this thread
On August 10, 2002 at 21:56:48, John Merlino wrote: >On August 10, 2002 at 21:53:49, John Merlino wrote: > >>Here is a position from a game that Slater and I discussed in a different >>thread: >> >>[D]3R4/8/8/8/2K5/6p1/3p1rk1/8 b - - 0 78 >> >>In that game, Black was obviously using EGTBs and played 78...d1=Q, leading to >>an EGTB mate in 22 moves (including the promotion). However, the more natural >>looking move, Kf1, leads to a mate in 14, as Chessmaster 9000 shows on a >>PIII-733: >> >>Time Depth Score Positions Moves >>0:00 1/5 -8.88 14450 78...Kf1 79.Kc3 g2 80.Rd4 g1=Q >> 81.Rxd2 Rxd2 82.Kxd2 >>0:00 2/6 -9.00 26438 78...Kf1 79.Kc3 g2 80.Rxd2 Rxd2 >> 81.Kxd2 g1=Q 82.Kd3 Qc5 83.Ke4 Ke2 >>0:00 3/7 -9.00 74581 78...Kf1 79.Kc3 g2 80.Rxd2 Rxd2 >> 81.Kxd2 g1=Q 82.Kd3 Qc5 83.Ke4 Ke2 >>0:02 4/8 -11.19 264563 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Re2 82.Rg8 d1=Q 83.Kc5 Kf1 >>0:05 5/9 -13.17 520117 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Re2 82.Rd8 g1=Q 83.Rxd2 Kxd2 84.Kd5 >>0:15 6/10 -13.30 1610302 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Re2 82.Rd8 g1=Q 83.Rd5 Qg4+ 84.Kc5 >> Qc8+ 85.Kb5 d1=Q 86.Rxd1+ Kxd1 >>0:46 7/11 -13.35 4752680 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Re2 82.Rd8 g1=Q 83.Rd5 Qg4+ 84.Kc5 >> Qc8+ 85.Kb5 Qb7+ 86.Kc5 Qc7+ 87.Kb5 >> d1=Q 88.Rxd1+ Kxd1 >>3:06 8/12 -15.25 19391217 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Kf1 82.Rd8 g1=Q 83.Rd4 Qg8+ 84.Kc5 >> Qg5+ 85.Kc4 Rf4 86.Kb3 Rxd4 >>9:45 9/13 -Mate14 62576834 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ >> Re2 82.Rd8 g1=Q 83.Rd6 Qg8+ 84.Kb4 >> Qb8+ 85.Kc4 Qc7+ 86.Kb5 Re5+ 87.Kb4 >> Qc5+ 88.Kb3 Re3+ 89.Rd3 Rxd3+ 90.Kb2 >> Qc1+ 91.Ka2 Ra3# >> >>Note that the first capture is 23 ply away, making it very difficult >>(impossible?) to find it via tablebases. Can anybody else find the mate faster? >> >>So, SPEAKING STRICTLY AESTHETICALLY, what do you think of EGTBs when used in >>game play? When we decided to implement them for Chessmaster 9000, our Producer >>(who was a class A player before he gave up taking the game "seriously") stated >>that he was, aesthetically speaking, strongly against them. He even went so far >>as to say that it was grossly unfair to use them in matches against humans, for >>obvious reasons. He also stated that something to the equivalent of "chess >>engines and opening books are an art form, but endgame databases are just plain >>boring -- and chess should be artistic and exciting". What say all of you? >> >>As for me, adding support for them was very useful in giving the user more ways >>to improve his/her game, and that meant much more to me (as a game designer) >>than just waiting for the game to get down to enough pieces for the EGTBs to >>kick in and quickly decide the game. But, then again, the first time I saw "mate >>in 35", I thought it was pretty cool.... :-) >> >>For reference, the other thread (with an interesting sacrifice) can be seen >>here: >> >>http://www.talkchess.com/forums/1/message.html?244901 >> >>jm > >Correction. As I was typing the above, I was leaving the analysis running, and >CM9000 just improved to Mate in 12, with the first capture 15 ply away: > >23:47 10/14 -Mate12 159363014 78...Kf1 79.Kc3 Ke1 80.Kc4 g2 81.Re8+ > Re2 82.Rd8 g1=Q 83.Rd6 Qg8+ 84.Kb4 > Qb8+ 85.Kc5 Qxd6+ 86.Kxd6 d1=Q+ > 87.Kc5 Rc2+ 88.Kb4 Qb1+ 89.Ka4 Ra2# > >Makes little to no difference to the discussion, though.... The ugly things that happen with egtb is when the engine jumps out of analyses after finding a mate. (obviously cm doesn't do that ) I have seen an example were without egtb a mate in 8 would be found but with egtb a mate in 78 ( or something ). Starting with mate in 78 and then improving to 8 is ok of course. Tony > >jm
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.