Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test - Pawn ending

Author: Sune Larsson

Date: 13:42:54 02/27/01

Go up one level in this thread


On February 27, 2001 at 16:03:18, John Merlino wrote:

>On February 27, 2001 at 15:37:03, Sune Larsson wrote:
>
>>
>>  [D]8/6k1/2p5/pp2p3/4P1K1/2PP4/P7/8 b - - 0 1
>>
>>  This is Damjanovic-Dvorecki, Vilnius 1978. Mistakes in the endgame are
>>  punished severely. Here black is on the move and has to decide if he
>>  wants to lose with 1.-c5??, thinks peace and draw is good enough with 1.-Kf6?,
>>  or goes for the whole cheese with 1.-Kg6! Oh, this is just tiresome
>>  calculation. No match for Kasparov, I guess. Computers?
>>
>>  The game went like this:
>>
>>  1.-Kg6!  [1.-c5?? 2.Kf5 b4 3.cxb4 cxb4 4.Kxe5 a4 5.Kd4+-]
>>           [1.-Kf6? 2.Kf3 c5 3.Ke3 b4 4.d4! cxd4+ 5.cxd4 a4 6.dxe5+ Kxe5 7.Kd3=]
>>
>>  2.Kf3  [2.d4 exd4 3.cxd4 b4-+]  2.-c5 3.Ke3  [3.d4 cxd4 4.cxd4 b4! 5.Ke3 a4-+]
>>  3.-b4 4.d4  [4.cxb4 cxb4 5.Kd2 a4! 6.Kc2 Kg5 7.Kd2 Kf4-+]
>>  4.-cxd4+ 5.cxd4 a4! 6.Kd3 b3 7.axb3 a3 8.Kc2 exd4 9.b4 d3+ 0-1
>>
>>  Test: 1.-c5?? loses. If you are a programmer please rewrite your program
>>                       totally from scratch. If you are a consumer just
>>                       throw the damned thing. ;)
>>
>>        1.-Kf6? draws. Ok, peaceful mind - but you just missed some glory.
>>                       Programmers: Have a look at the evals - the right move
>>                       was just around the corner.
>>
>>        1.-Kg6! wins.  Congrats! This is good stuff. Just hope the margin
>>                       to 1.-Kf6? was more than 0.02...
>>
>>  Sune
>
>I'm not so sure about this one. CM8000 sees 1...Kf6 as a win, as it would play
>3...a4! instead of 3...b4 (as in your line above):
>
>Time	Depth	Score	Positions	Moves
>0:00	7/8	0.58	10424		1...Kg6 2. d4 exd4 3. cxd4 Kf6
>					4. e5+ Ke7 5. Kg5 b4 6. Kf5 a4
>0:00	9/10	-0.43	61408		1...Kf6 2. Kf3 c5 3. Ke3 Ke6 4.
>					Kf3 b4 5. cxb4 cxb4 6. Ke3 a4
>0:00	10/11	-0.32	106547		1...Kf6 2. Kf3 Ke6 3. Ke3 b4 4.
>					cxb4 axb4 5. Kf3 c5 6. Ke2 Kd6
>					7. Ke3
>0:01	11/12	-0.43	192099		1...Kf6 2. Kf3 Ke6 3. Ke3 c5 4.
>					Kf3 Kd6 5. Ke2 b4 6. cxb4 cxb4
>					7. Ke3 a4
>0:02	12/13	-0.53	378998		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 b4 5. Kc2 Kg5 6. Kd2 Kf4 7.
>					Kc2 Ke3
>0:04	13/14	-0.61	604989		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					d4 exd4+ 5. cxd4 c4 6. Kf4 c3 7.
>					Ke3 b4 8. Kd3 Ke6
>0:07	14/15	-0.67	1011338		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					d4 exd4+ 5. cxd4 c4 6. Kd2 b4 7.
>					Kd1 Ke6 8. Kc2 b3+ 9. Kb2 bxa2
>					10. d5+ Kd6 11. Kxa2
>0:13	15/16	-0.98	1744296		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 b4 5. Kc2 Kg5 6. cxb4 cxb4
>					7. Kc1 Kf4 8. Kd2 b3 9. axb3 axb3
>					10. Kc3
>0:21	16/17	-0.99	2823338		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 b4 5. Kc2 Kg5 6. cxb4 cxb4
>					7. Kc1 Kf4 8. Kd2 a3 9. Kc1 Ke3
>					10. Kc2 Kd4
>0:39	17/18	-0.99	4886592		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 b4 5. Kc2 Kg5 6. cxb4 cxb4
>					7. Kc1 Kf4 8. Kd2 Kf3 9. Kc1 Ke3
>					10. Kc2 Kd4
>1:21	18/19	-1.92	9517380		1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 a3 5. Kc2 Kg5 6. Kb3 Kf4 7.
>					Kxa3 Ke3 8. d4 exd4 9. e5 dxc3
>					10. Kb3 Kd2 11. e6
>4:11	19/20	-2.84	26611495	1...Kf6 2. Kf3 c5 3. Ke3 a4 4.
>					Kd2 b4 5. Kc2 Kg5 6. Kc1 Kf4 7.
>					Kd2 b3 8. axb3 axb3 9. c4 b2 10.
>					Kc2 Ke3 11. Kxb2 Kxd3 12. Kc1 Kxe4
>
>Also, here's what it sees after 3...b4 4.d4, from your 1...Kf6? line above (note
>4...exd4+ rather than 4...cxd4+, which IS a draw):
>
>Time	Depth	Score	Positions	Moves
>0:00	7/8	-0.67	22773		4...c4 5. Kd2 b3 6. Kc1 exd4 7.
>					cxd4 bxa2 8. Kb2 Ke7 9. Kxa2 c3
>0:00	8/9	-0.74	53183		4...c4 5. Kd2 b3 6. Kc1 exd4 7.
>					cxd4 Ke6 8. axb3 cxb3 9. Kb2 a4
>0:01	9/10	-0.50	116802		4...c4 5. cxb4 axb4 6. d5 Ke7 7.
>					Ke2 Kd6 8. Kd2 c3+ 9. Kd3 Kc5 10.
>					Kc2
>0:01	9/10	-1.42	150625		4...exd4+ 5. cxd4 a4 6. dxc5 b3
>					7. axb3 a3 8. c6 Ke6 9. c7 Kd7
>					10. Kd4 a2
>0:02	10/11	-3.09	301015		4...exd4+ 5. cxd4 a4 6. Kd3 b3
>					7. axb3 a3 8. Kc2 cxd4 9. Kc1 Ke5
>					10. b4 Kxe4
>0:03	11/12	-3.08	446245		4...exd4+ 5. cxd4 a4 6. Kd3 b3
>					7. axb3 a3 8. e5+ Ke6 9. d5+ Kxe5
>					10. Kc2 Kxd5 11. Kb1 Kd4 12. Ka2
>0:06	12/13	-4.22	774563		4...exd4+ 5. cxd4 a4 6. Kd3 b3
>					7. axb3 a3 8. e5+ Ke7 9. Kc2 cxd4
>					10. Kc1 a2 11. Kb2 d3 12. b4 a1=Q+
>					13. Kxa1
>0:12	13/14	-6.64	1418541		4...exd4+ 5. cxd4 a4 6. dxc5 b3
>					7. axb3 a3 8. c6 Ke6 9. Kd3 a2
>					10. Kc4 a1=Q 11. b4 Kd6 12. Kb5
>					Qe5+ 13. Kb6 Qxe4
>0:19	14/15	-6.98	2235356		4...exd4+ 5. cxd4 a4 6. dxc5 b3
>					7. axb3 a3 8. c6 Ke6 9. Kd3 a2
>					10. Kc4 a1=Q 11. b4 Kd6 12. Kb5
>					Qe5+ 13. Kb6 Qd4+ 14. Kb5 Qxe4
>0:58	15/16	-7.72	6315957		4...exd4+ 5. cxd4 a4 6. e5+ Ke7
>					7. Kd3 b3 8. axb3 a3 9. dxc5 a2
>					10. Kc4 a1=Q 11. Kb5 Qxe5 12. b4
>					Kd7 13. Ka6 Kc6
>
>The results are very similar if White responds to 4.d4 to 3...a4, rather than
>3...b4.
>
>But, at least we all agree that 1...c5?? loses! ;-):
>
>Time	Depth	Score	Positions	Moves
>0:00	7/8	3.87	11148		2. Kf5 Kf7 3. Kxe5 c4 4. dxc4 bxc4
>					5. Kd5 Ke7 6. Kxc4 Kd6
>0:00	10/11	3.82	68824		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 b3 6. axb3 axb3 7. Kc3 Kf6
>					8. Kxb3 Ke5
>0:01	11/12	4.08	128116		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 Kf6 6. Kc4 b3 7. axb3 axb3
>					8. Kxb3 Ke5 9. Kc4
>0:01	12/13	4.37	212689		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 b3 6. axb3 axb3 7. Kc3 b2
>					8. Kxb2 Kf6 9. Kc3 Ke6 10. Kd4
>0:02	13/14	3.15	375757		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 Kf7 6. Kc4 b3 7. axb3 a3
>					8. Kc3 Ke8 9. d4 a2
>0:05	14/15	3.14	668868		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 Kf7 6. Kc4 b3 7. axb3 a3
>					8. Kc3 Ke8 9. d4 a2 10. Kb2
>0:09	15/16	3.10	1234305		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 Kf6 6. Kc4 b3 7. axb3 a3
>					8. Kc3 Ke5 9. b4 Kd6 10. Kb3
>0:22	16/17	5.08	2802226		2. Kf5 a4 3. Kxe5 b4 4. cxb4 cxb4
>					5. Kd4 Kf6 6. Kc4 b3 7. axb3 axb3
>					8. Kxb3 Ke5 9. Kc4 Ke6 10. d4 Kf7
>					11. e5 Ke7 12. d5
>
>jm

 Oops, looks like GM Dvorecki has made some mistakes in his comments...
 The variations I added were strictly his and at first glance they looked
 okey. But I cannot refute CM:s analysis at this stage. Looks like 1.-Kf6
 also wins, given the moves of CM. Fantastic! But then, these pawn endings
 are tricky and maybe some talent around here can find something new?!

 Sune



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.