Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: White to move and win

Author: John Merlino

Date: 12:02:57 04/24/02

Go up one level in this thread


On April 24, 2002 at 14:55:57, Roy Eassa wrote:

>On April 24, 2002 at 14:03:30, John Merlino wrote:
>
>>On April 24, 2002 at 13:48:12, Roy Eassa wrote:
>>
>>>
>>>Does ANY program see the whole line (3 underpromotions) up front?
>>>
>>>[d] 1b6/4P3/1P2PN2/8/8/P1k5/P1p2P2/K5B1 w - - 0 1
>>>
>>>1.Ne4+
>>>
>>>   [1.Nd5+? Kd2 2.Kb2 c1Q+ 3.Kb3 Bd6! 4.Nb4 Bxe7 5.b7 Qc7 -+]
>>>
>>>1...Kd3 2.Nc5+
>>>
>>>   [2.Kb2? Be5+ and Black mates]
>>>
>>>2...Kc3 3.Nb3 Be5 4.f4! forced
>>>
>>>   [4.Bh2? Bf6 5.e8N Bh8 and Black mates!]
>>>
>>>4...Bf6 {or ...Bg7} 5.e8N! forced
>>>
>>>5...Bh8 6.f5! forced
>>>
>>>6...Be5 7.Bh2! forced
>>>
>>>7...Bxh2 8.b7! forced
>>>
>>>8...Be5! 9.b8B!
>>>
>>>   [9.b8Q? Kc4+ 10.Qxe5 c1Q+ 11.Nxc1 stalemate]
>>>
>>>9...Bxb8 10.Nc7! forced
>>>
>>>10...Bxc7 11.e7 forced
>>>
>>>11...Be5 12.e8R!
>>>
>>>   [12.e8Q? Kc4+ 13.Qxe5 c1Q+ 14.Nxc1 stalemate]
>>>
>>>12...Bg7 {or ...Bf6}
>>>
>>>13.Re6 followed by f6
>>>
>>>
>>>Problem attributed to Mark Libiurkin, circa 1933.
>>
>>There is a problem with the above line. Here is CM's analysis from the initial
>>position (note the first seven moves agree with your above line):
>>
>>Time	Depth	Score	Positions	Moves
>>0:00	2/3	9.94	3081		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Kc4
>>					4.Kb2
>>0:00	3/4	7.65	6474		1.Ne4+ Kd3 2.Nc5+ Kd2 3.Nb3+ Kd1
>>					4.f4 Bxf4 5.Bd4
>>0:00	4/5	9.82	13420		1.Ne4+ Kd3 2.Nc5+ Ke2 3.Kb2 Kd1
>>					4.Nd3 Be5+ 5.Kb3
>>0:00	5/6	12.01	27685		1.Ne4+ Kd3 2.Nc5+ Kd2 3.Nb3+ Kd1
>>					4.f4 Bxf4 5.Bd4 Be3 6.Bxe3
>>0:00	6/7	12.07	55694		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.Bh2 Bf6 5.e8=N Be7 6.Be5+ Kc4
>>0:01	7/8	12.01	105397		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.f4 Bf6 5.e8=N Bh8 6.f5
>>0:02	8/9	12.23	199081		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.f4 Bf6 5.e8=N Bh8 6.f5 Be5 7.f6
>>0:05	9/10	9.27	469863		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.f4 Bf6 5.e8=N Bh8 6.f5 Be5 7.Bd4+
>>					Bxd4
>>0:09	10/11	9.37	865339		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.f4 Bf6 5.e8=N Bh8 6.f5 Be5 7.Bh2
>>					Bxh2 8.Nf6
>>0:21	11/12	9.37	1872831		1.Ne4+ Kd3 2.Nc5+ Kc3 3.Nb3 Be5
>>					4.f4 Bf6 5.e8=N Bh8 6.f5 Be5 7.Bh2
>>					Bxh2 8.Nf6
>>
>>However, in your line, 8...Be5 loses outright. Here is Chessmaster 9000's
>>analysis after 8.b7, which says that 8...Kd3 is the best move, and leads to a
>>game where White has a queen and three pawns, and Black has a queen and a
>>bishop:
>>
>>Time	Depth	Score	Positions	Moves
>>0:00	2/3	7.65	917		8...Be5 9.b8=Q Bxb8
>>0:00	3/4	7.61	2731		8...Be5 9.b8=B Bxb8 10.f6
>>0:00	4/5	7.73	7396		8...Be5 9.b8=B Bxb8 10.f6 Bf4
>>0:00	4/5	6.00	9191		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nf6
>>0:00	5/6	5.44	12675		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nf6 Kd2
>>0:00	6/7	4.47	21350		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.f6 Kb1 13.f7 c1=Q
>>					14.f8=Q Qc3+ 15.Ka4 Kxa2
>>0:00	7/8	3.61	40916		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nf6 Kb1 13.b8=Q Bxb8
>>0:00	8/9	2.92	71437		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.b8=Q Bxb8 13.f6
>>					Kb1 14.f7 c1=Q 15.f8=Q Qc2+ 16.Kb4
>>					Kxa2
>>0:01	9/10	2.91	129596		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.f6 Kb1 13.f7 c1=Q
>>					14.f8=Q Qc3+ 15.Ka4 Qc6+ 16.Kb4
>>					Qxb7+ 17.Kc5 Kxa2
>>0:03	10/11	2.62	316378		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nc7 Bxc7 13.b8=Q
>>					Bxb8 14.e7 Kd1 15.e8=Q c1=Q 16.Qxb8
>>					Qc2+ 17.Kb4 Qxf5
>>0:07	11/12	2.28	637636		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nc7 Bxc7 13.e7 Kb1
>>					14.e8=Q c1=Q 15.Qe2 Ka1 16.Qg2
>>					Qb1+ 17.Kc3
>>0:15	12/13	2.27	1366434		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nc7 Bxc7 13.e7 Kb1
>>					14.e8=Q c1=Q 15.Qe4+ Ka1 16.Qd4+
>>					Kb1 17.Qd3+ Ka1 18.Qc3+ Qxc3+ 19.Kxc3
>>					Kxa2 20.a4 Ka3
>>0:42	13/14	2.35	3685053		8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nc7 Bxc7 13.e7 Kd2
>>					14.e8=Q c1=Q 15.Qd7+ Ke3 16.Qd5
>>					Kf4 17.f6 Qe3+ 18.Kc4 Qc1+ 19.Kb5
>>					Qxa3
>>1:57	14/15	2.40	10277553	8...Kd3 9.Nc1+ Kd2 10.Kb2 Be5+
>>					11.Kb3 Kxc1 12.Nc7 Bxc7 13.e7 Kd2
>>					14.e8=Q c1=Q 15.Qd7+ Ke3 16.Qd5
>>					Kf4 17.f6 Qe3+ 18.Kc4 Qxa3 19.Qd2+
>>					Kf3
>>
>>Here is the analysis after 8...Be5 (note the remaining two underpromotions and
>>the much higher eval for White):
>>
>>Time	Depth	Score	Positions	Moves
>>0:00	2/3	0.00	453		9.b8=Q Kc4+ 10.Qxe5 c1=Q+ 11.Nxc1
>>0:00	2/3	7.61	680		9.b8=B Bxb8 10.f6
>>0:00	3/4	7.73	1862		9.b8=B Bxb8 10.f6 Bf4
>>0:00	4/5	7.47	5160		9.b8=B Bxb8 10.Nf6 Be5 11.Nd7
>>0:00	5/6	7.92	9869		9.b8=B Bxb8 10.Nc7 Kc4 11.e7 Ba7
>>					12.Kb2
>>0:00	6/7	10.09	16706		9.b8=B Bxb8 10.Nc7 Kc4 11.e7 Ba7
>>					12.Kb2 c1=Q+ 13.Kxc1
>>0:00	7/8	6.96	31443		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re7
>>0:00	8/9	7.32	47963		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Rg8 Bf6
>>0:00	9/10	7.73	82525		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bf6 13.Re6 Bg7 14.Rc6+
>>					Kd3+ 15.f6
>>0:01	10/11	7.82	147748		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 Kd3+ 14.f6 Bf8
>>					15.Kb2
>>0:02	11/12	8.16	271525		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 Kc4+ 14.f6 Bh8
>>					15.Kb2 Kd5 16.Re8 c1=Q+ 17.Kxc1
>>					Bxf6
>>0:04	12/13	8.20	479706		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 Kc4+ 14.f6 Bh8
>>					15.Rd6 Kb5 16.Kb2
>>0:09	13/14	8.44	924526		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 Kc4+ 14.f6 Bh8
>>					15.Rd6 Kc3 16.Rc6+ Kd3 17.Kb2
>>0:19	14/15	9.24	1948702		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 c1=Q+ 14.Nxc1
>>					Kc2+ 15.f6 Bxf6+ 16.Rxf6 Kxc1 17.a4
>>					Kc2 18.a5 Kd2 19.Kb2 Kd3 20.a6
>>0:53	15/16	9.80	5290251		9.b8=B Bxb8 10.Nc7 Bxc7 11.e7 Be5
>>					12.e8=R Bg7 13.Re6 Bf8 14.Rc6+
>>					Kd3 15.Kb2 Ke4 16.f6 Ke5 17.f7
>>					c1=Q+ 18.Rxc1
>>
>>jm
>
>
>Nothing to disagree with except the word "problem" -- White wins in all
>variations, but the main line I chose is the one where White is forced to
>underpromote three separate times (to a knight, bishop, and rook).  The other
>line may evaluate less badly for Black, but it's a more straightforward win for
>a human IMHO.
>
>To my analysis above, I should have added the following variation to Black's
>8...Be5:
>
>   [8...Kd3 9.b8Q Bxb8 10.Nc7 Bxc7 11.Nc1+ Kd2 12.Kb2 Be5+ 13.Kb3 Kxc1
>      (13...Bf6 14.Ne2 +-)
>   14.e7 +-]

I agree with everything, except that you probably cannot expect a prog to find
this triple-underpromotion line, because Black has a better response than the
one that will force White to follow it.

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.