Computer Chess Club Archives




Subject: Re: Two mate positions to solve.

Author: leonid

Date: 09:38:50 09/03/00

Go up one level in this thread

On September 03, 2000 at 11:32:09, Bert van den Akker wrote:

>On September 02, 2000 at 17:08:20, leonid wrote:
>>Please solve these two positions and say minimum number of moves that lead to
>> 1r2k2r/1b3pbp/1NPqN1n1/1P1Q4/1P2P3/P3B3/5PP1/R3KB1R White to go.
>> 6KQ/3B2Q1/QQQQQQQ1/5Nbn/qqqqq1q1/5qqk/8/8 White to go.
>>It will be good if you will indicate all main parameters of your trial. At
>>least, I see them this way:
>>Name of your program. Hash table size.
>>In what way (if you know) position was solved. Selective search or brute force
>>Time for each way of search.
>>CPU and its speed.
>>If you could come with some of your "crazy" (fantasy) position or normal one, it
>>will be good. When you will depose your, try to make them between 5 and 9 moves
>>mate. That deep you still can expect reasonable time to be solved by brute force
>>on our computers.
>>Thanks in advance,
>Hello Leonid,
>My program is basicly brute force but the extensions and null moves makes it
>also selective.

Interesting since it very different from my mate search.

>Depending on the positions mates till 45 plies can be
>solved in seconds. In this cases there are many checks and ony one reply on that
>Here are my the results of the second positions:
>[D] 6KQ/3B2Q1/QQQQQQQ1/5Nbn/qqqqq1q1/5qqk/8/8 w
>Name program: Bobo
>Hash table size: 262144 entries
>recursiv null move on

Do you use null move when solving the mate positions?

>extensions used:
>Check extension

Maybe, it is what I have. My program, when look for mate, search through
indicated number of moves. Those moves can be of every kind possible.

>Threat extension
>Single move out of check extension in normal search
>Single move out of check extension in quiet search
>(score= 1941) (depth= 1/1)
>(time = 19.9900000000052)
>(evaluations= 401986)
>(score= 1942) (depth= 1/1)
>(time = 26.5300000000061)
>(evaluations= 530496)
>(score= 1943) (depth= 1/1)
>(time = 32.0200000000041)
>(evaluations= 637640)
>(score= 1944) (depth= 1/1)
>(time = 40.2100000000064)
>(evaluations= 798971)
>Qa6xa4  Qb4xd6  Qa4xc4  Qd4xc4  Qc6xe4  Qd6xb6  Qe6xc4  Qf3xe4  Qc4xe4  Qg4xe4
>(score= 1954) (depth= 1/11)
>(time = 60.0300000000061)
>(evaluations= 1191561)
>(score= 1955) (depth= 1/1)
>(time = 65.2000000000044)
>(evaluations= 1291614)
>Qa6xc4  Qb4xd6  Qc6xd6  Qd4xd6
>(score= 1963) (depth= 1/4)
>(time = 76.9500000000044)
>(evaluations= 1523217)
>(score= 1964) (depth= 1/1)
>(time = 93.4300000000003)
>(evaluations= 1846206)
>Qc6xc4  Qd4xQ6  Qc4xe4  Qd6xe6+ Ld7xe6
>(score= 1965) (depth= 1/5)
>(time = 102.330000000002)
>(evaluations= 2013926)
>(score= 1966) (depth= 1/1)
>(time = 108.090000000004)
>(evaluations= 2122987)
>Qc6xe4  Qd4xQ6  Pf5xd6  Qc4xe6+ Ld7xe6
>(score= 1967) (depth= 1/5)
>(time = 118.040000000001)
>(evaluations= 2302593)
>Qc6xe4  Qd4xe4  Pf5xg3  Qe4xg6  Qf6xg6  Qg4xe6+ Ld7xe6+ Kh3-g2
>(score= 1946) (depth= 2/8)
>(time = 102.709999999999)
>(evaluations= 2114069)
>(score= 1947) (depth= 2/1)
>(time = 123.139999999999)
>(evaluations= 2482105)
>Qg6xh5+ Qg4xh5  Qh8xh5+ Kh3-g2  Qc6xe4  Qd4xe4  Qd6xg3+ Qf3xg3  Pf5xg3
>(score= 2309) (depth= 2/9)
>(time = 281.709999999999)
>(evaluations= 5391421)
>(score= 2537) (depth= 3/1)
>(time = 25.3699999999953)
>(evaluations= 513635)
>Qg6xh5+ Qg4xh5  Qh8xh5+ Qg3-h4  Pf5xh4+ Qc4xe6+ Bd7xe6+ Qe4xe6+ Qd6xe6+ Qf3-g4
>Qe6xg4+ Qd4xg4  Qh5xg4+ Kh3xg4  Qb6xb4+ Kg4-g3  Qf6xg5+ Kg3-h2  Qa6-e2+ Kh2-h3
>Qe2-f3+ Kh3-h2  Qf3-f4+ Kh2-h3  Qf4-f5+ Kh3-h2  Ph4-f3+ Kh2-h1  Qf5-h7+
>(score= 19971) (depth= 3/29)
>(time = 135.769999999997)
>Mate in 15 moves

The most probable is that your program is more that solid. This kind of position
could kill many good programs. It is very heavy. Average number of nodes in each
ply is 52. It was said by average nodes counter 8 plies deep. Occasional number
of nodes in ply reach beyond 100 moves.

I even don't know exactly number of minimum moves that lead to the mate in this
position. When tried to find the move, I only found that mate existe for sure in
11 moves. Expected that somebody else with better branching factor will say
final number. Since I could not solve this position by very quick selective
search, I never even tried to reach minimum number by brute force. It could be
too long. I know only that one selective search find mate in 11 moves in 8 sec.
Computer AMD 400M, llchess solver.


This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.