Author: Robert Hyatt
Date: 07:21:38 08/30/04
Go up one level in this thread
On August 30, 2004 at 02:54:25, martin fierz wrote: >On August 29, 2004 at 20:06:19, Robert Hyatt wrote: > >>On August 29, 2004 at 02:36:48, martin fierz wrote: >> >>>On August 28, 2004 at 18:40:43, martin fierz wrote: >>> >>>>aloha! >>>> >>>>after all the cloning talk, here's something to test your favorite engine with: >>>>i played a weakish IM today and got the following position after 56...Qg6+ >>>>57.Kf2: >>>> >>>>[D] 8/6kp/8/3p4/1P4q1/P4p2/5K2/4Q3 b - - 0 57 >>>> >>>>it took me a couple of seconds to see that after 57...Qh4+ 58. Kf1 Qxe1+ 59. >>>>Kxe1 Kf6 black is completely winning due to the floating square (it's nice to >>>>see that reading dvoretzky is some good!) and so i played that and won :-) >>>> >>>>interestingly, when i entered this game in chessbase running fritz7 in the >>>>background, fritz didn't seem to see this at all within a minute or so. so >>>>question to all: how long does your favorite engine need to see the line ...Qh4+ >>>>Kf1 Qxe1+ with a HUGE plus score (i.e. +9 or so - that is what my assessment of >>>>the position was after a couple of seconds...)? >>>> >>>>i don't want to claim that this is the only win, but it is by far the easiest >>>>one, and fritz 7 misses it completely. >>>> >>>>cheers >>>> martin >>> >>>kurt made the (unfortunately valid) point that my original position allows black >>>to win without exchanging queens. so i've modified it to exclude that >>>possibility. >>> >>>the point i wanted to make is that human experts don't treat an exchange into a >>>pawn ending as "yet another variation", instead they check such exchanges much >>>more carefully. which is why i see this Q-exchange variation immediately, and an >>>engine which either >>>1) doesn't check trades into pawn endings properly or >>>2) lacks knowledge about the square / the floating square >>>will have trouble coming up with the line i gave. which also means it would have >>>trouble in my next test position, which is a simple modification (i removed the >>>black pawn on f3 - if 6-piece tablebases with 4 pawns were available, one could >>>add another white pawn on a2/b2/b3 for engine confusion without changing the >>>result): >>> >>>[D] 8/6kp/8/3p4/1P4q1/P7/5K2/4Q3 b - - 0 57 >>> >>>black still wins with the Q-exchange, and he wins very easily. my fritz 7 on a >>>P4 1.4GHz is showing ...Qc4 (0.00)after 6 minutes (can anybody check out fritz >>>8?). >>> >>>so again: which engine sees this in a reasonable amount of time? to me it is >>>just very astonishing that a top engine like fritz can't handle this... i'm used >>>to analyze my games with fritz, and usually it finds errors i make, and not vice >>>versa :-) >>> >>>cheers >>> martin >> >> >>Crafty on my dual xeon: >> >> 1-> 0.02 -5.15 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 >> 2 0.02 -5.25 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 2-> 0.02 -5.25 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 3 0.02 -5.05 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 >> 3-> 0.02 -5.05 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 >> 4 0.02 -5.15 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 >> 4-> 0.02 -5.15 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 (s=2) >> 5 0.02 -5.05 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. Kf2 >> 5-> 0.03 -5.05 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. Kf2 (s=3) >> 6 0.03 -4.99 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 (s=2) >> 6-> 0.04 -4.99 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 >> 7 0.05 -4.75 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 6. b5 >> 7-> 0.08 -4.75 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 6. b5 (s=7) >> 8 0.11 -4.95 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 6. b5 d4 (s=6) >> 8-> 0.21 -4.95 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke5 5. a5 Kd6 6. b5 d4 (s=7) >> 9 0.36 -4.67 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke7 5. a5 Kd7 6. a6 Kc6 7. b5+ >> Kb6 (s=6) >> 9-> 0.92 -4.67 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke7 5. a5 Kd7 6. a6 Kc6 7. b5+ >> Kb6 (s=4) >> 10 1.69 -4.65 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke7 5. a5 Kd7 6. b5 Kc7 7. b6+ >> Kb7 8. Kf2 (s=3) >> 10-> 4.79 -4.65 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf6 >> 4. a4 Ke7 5. a5 Kd7 6. b5 Kc7 7. b6+ >> Kb7 8. Kf2 (s=3) >> 11 9.36 -4.57 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf7 >> 4. a4 Ke7 5. a5 Kd6 6. b5 Kc7 7. a6 >> Kb6 <HT> (s=2) >> 11-> 28.97 -4.57 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf7 >> 4. a4 Ke7 5. a5 Kd6 6. b5 Kc7 7. a6 >> Kb6 <HT> (s=3) >> 12 48.05 -4.61 1. ... Qh4+ 2. Ke2 Qxe1+ 3. Kxe1 Kf7 >> 4. a4 Ke7 5. a5 Kd6 6. Ke2 Kc7 7. Kd3 >> h6 8. Kd4 Kc6 (s=2) >> 12 48.05 2/30* 1. ... Qf5+ >> >>Takes 1 ply here. :) > >nice! >interesting how some (otherwise very good) programs have absolutely no clue >while others like crafty and hiarcs need no depth at all... > >cheers > martin If all your testing is against other computers, most of which don't know beans about this particular problem, you might never know you have the problem. If you test a lot against GM players (ICC for example) it will be highlighted very quickly. :)
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.