Engine test: tricky 3-move mate

Discussion of chess software programming and technical issues.

Moderator: Ras

Chessnut1071
Posts: 313
Joined: Tue Aug 03, 2021 2:41 pm
Full name: Bill Beame

Engine test: tricky 3-move mate

Post by Chessnut1071 »

[fen]3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 1 [/fen]

Some engines have trouble with this puzzle. I'll give the solution later. It tests ent passant and discovered check. enjoy.
amanjpro
Posts: 883
Joined: Sat Mar 13, 2021 1:47 am
Full name: Amanj Sherwany

Re: Engine test: tricky 3-move mate

Post by amanjpro »

Chessnut1071 wrote: Sat Oct 16, 2021 5:06 am [fen]3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 1 [/fen]

Some engines have trouble with this puzzle. I'll give the solution later. It tests ent passant and discovered check. enjoy.
Can you please learn how to put FENs?

Code: Select all

position fen 3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 1
string info invalid FEN notation 3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 0 1, castling part is not correct --
string info invalid FEN notation 3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 0 1, en-passant part is not correct '0'
User avatar
flok
Posts: 606
Joined: Tue Jul 03, 2018 10:19 am
Full name: Folkert van Heusden

Re: Engine test: tricky 3-move mate

Post by flok »

Chessnut1071 wrote: Sat Oct 16, 2021 5:06 am [fen]3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w -- 0 1 [/fen]

Some engines have trouble with this puzzle. I'll give the solution later. It tests ent passant and discovered check. enjoy.
Invalid FEN.
abulmo2
Posts: 479
Joined: Fri Dec 16, 2016 11:04 am
Location: France
Full name: Richard Delorme

Re: Engine test: tricky 3-move mate

Post by abulmo2 »

A space is missing between the castling right and en passant fields. Morever the diagram said that black is to move while the fen said it is white turn. The correct fen could be:
[fen]3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w - - 0 1[/fen]

For uci engines, the line to enter is

Code: Select all

position fen 3R3N/2pR1p2/4k3/N1P5/1q4PK/2B4B/8/8 w - - 0 1
Amoeba finds it easily (or luckily?):

Code: Select all

info depth 1 seldepth 1 score cp 3537 time 0 nodes 36 nps 155642 pv c3b4 
info depth 2 seldepth 2 score cp 3537 time 0 nodes 77 nps 270935 pv c3b4 e6f6 
info depth 3 seldepth 3 score cp 3537 time 0 nodes 149 nps 445574 pv c3b4 e6f6 h8f7 
info depth 4 seldepth 7 score mate 4 time 1 nodes 3032 nps 2187590 pv d7d6 c7d6 d8d6 e6e7 c3f6 e7e8 d6d8 
info depth 5 seldepth 7 score mate 4 time 2 nodes 7756 nps 2732236 pv d7d6 c7d6 d8d6 e6e7 c3f6 e7e8 d6d8 
info depth 6 seldepth 7 score mate 3 time 4 nodes 12882 nps 2995187 pv c5c6 b4b8 d8b8 f7f6 b8e8 
bestmove c5c6
Richard Delorme
User avatar
flok
Posts: 606
Joined: Tue Jul 03, 2018 10:19 am
Full name: Folkert van Heusden

Re: Engine test: tricky 3-move mate

Post by flok »

Code: Select all

info depth 1 score cp 2433 time 36 nodes 76 nps 2111 pv c3b4
info depth 2 score cp 2554 time 55 nodes 239 nps 4345 pv c3b4 e6f6 d7c7
info depth 3 score cp 2567 time 155 nodes 4180 nps 26967 pv c3b4 f7f6 h8f7
info depth 4 score cp 2692 time 182 nodes 6264 nps 34417 pv c3b4 f7f5 g4f5 e6e5 d7c7
info depth 5 score cp 9999 time 365 nodes 24556 nps 67276 pv c3b4 f7f5 b4c3 f5f4 g4g5
info depth 6 score cp 9999 time 559 nodes 32794 nps 58665 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 7 score cp 9999 time 1078 nodes 111978 nps 103875 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 8 score cp 9999 time 1276 nodes 127240 nps 99717 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 9 score cp 9999 time 2428 nodes 348137 nps 143384 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 10 score cp 9999 time 5493 nodes 1396344 nps 254204 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 11 score cp 9999 time 11927 nodes 2959739 nps 248154 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 12 score cp 9999 time 49473 nodes 18422763 nps 372380 pv c5c6 b4c3 d8e8 e6f6 d7f7
bestmove c5c6
Hmmm something goes wrong here.
User avatar
flok
Posts: 606
Joined: Tue Jul 03, 2018 10:19 am
Full name: Folkert van Heusden

Re: Engine test: tricky 3-move mate

Post by flok »

flok wrote: Sat Oct 16, 2021 8:15 am

Code: Select all

info depth 1 score cp 2433 time 36 nodes 76 nps 2111 pv c3b4
info depth 2 score cp 2554 time 55 nodes 239 nps 4345 pv c3b4 e6f6 d7c7
info depth 3 score cp 2567 time 155 nodes 4180 nps 26967 pv c3b4 f7f6 h8f7
info depth 4 score cp 2692 time 182 nodes 6264 nps 34417 pv c3b4 f7f5 g4f5 e6e5 d7c7
info depth 5 score cp 9999 time 365 nodes 24556 nps 67276 pv c3b4 f7f5 b4c3 f5f4 g4g5
info depth 6 score cp 9999 time 559 nodes 32794 nps 58665 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 7 score cp 9999 time 1078 nodes 111978 nps 103875 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 8 score cp 9999 time 1276 nodes 127240 nps 99717 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 9 score cp 9999 time 2428 nodes 348137 nps 143384 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 10 score cp 9999 time 5493 nodes 1396344 nps 254204 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 11 score cp 9999 time 11927 nodes 2959739 nps 248154 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 12 score cp 9999 time 49473 nodes 18422763 nps 372380 pv c5c6 b4c3 d8e8 e6f6 d7f7
bestmove c5c6
Hmmm something goes wrong here.

Code: Select all

...
info depth 9 score cp 9995 time 2845 nodes 363708 nps 127841 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 10 score cp 9995 time 7023 nodes 1420511 nps 202265 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 11 score cp 9995 time 15023 nodes 3046144 nps 202765 pv c5c6 b4c3 d8e8 e6f6 d7f7
Well, scoring fixed. Still has problems realising it found the right move.
Joost Buijs
Posts: 1646
Joined: Thu Jul 16, 2009 10:47 am
Location: Almere, The Netherlands

Re: Engine test: tricky 3-move mate

Post by Joost Buijs »

Nightmare has no problem with this either.

Code: Select all

info depth 1 score cp 1628 time 1 nodes 52 nps 52000 tbhits 0 pv d7d6
info depth 2 score cp 1628 time 2 nodes 1067 nps 533500 tbhits 0 pv d7d6 c7d6
info depth 3 score cp 1566 time 3 nodes 8704 nps 2901333 tbhits 0 pv d7e7 e6e7 c3b4
info depth 4 score mate 4 time 3 nodes 20502 nps 6834000 tbhits 0 pv d7d6 c7d6 d8d6 e6e7
info depth 5 score mate 4 time 4 nodes 35087 nps 8771750 tbhits 0 pv d7d6 c7d6 d8d6 e6e7 a5c6
info depth 6 score mate 3 time 6 nodes 76169 nps 12694833 tbhits 0 pv c5c6 b4d6 d8e8 e6d5 e8e5
info depth 7 score mate 3 time 7 nodes 95088 nps 13584000 tbhits 0 pv c5c6 b4d6 d8e8 e6d5 e8e5
info depth 8 score mate 3 time 9 nodes 138009 nps 15334333 tbhits 0 pv c5c6 b4d6 d8e8 e6d5 e8e5
info depth 9 score mate 3 time 10 nodes 171031 nps 17103100 tbhits 0 pv c5c6 b4d6 d8e8 e6d5 e8e5
User avatar
Ajedrecista
Posts: 2134
Joined: Wed Jul 13, 2011 9:04 pm
Location: Madrid, Spain.

Re: Engine test: tricky 3-move mate.

Post by Ajedrecista »

Hello:

There is a bunch of variations once the key move is played, like in many three-movers. This fact resembles the choose your own adventure approach. All the variations can be found at YACPDB (Yet Another Chess Problem Database):

https://yacpdb.org/#209270

It is another Howard's problem, which is the source that Bill told us that he would use in three-movers and moremovers.

The featured problem is found with an introductory comment in Howard's book How to Solve Chess Problems, Second Revised Edition (1961), page 108, problem No. 66:

https://books.google.es/books?id=TAr0Vr ... &q&f=false

Image

(Scanning courtesy of Google Books).

Regards from Spain.

Ajedrecista.
User avatar
Ras
Posts: 2703
Joined: Tue Aug 30, 2016 8:19 pm
Full name: Rasmus Althoff

Re: Engine test: tricky 3-move mate

Post by Ras »

CT800 V1.43:

Code: Select all

info depth 2 seldepth 2 score cp 2452 time 50 nodes 674 nps 13480 hashfull 0 tbhits 0 pv c3b4 f7f6
info depth 3 seldepth 4 score cp 2479 time 51 nodes 949 nps 18607 hashfull 0 tbhits 0 pv c3b4 f7f6 g4g5 e6e5
info depth 4 seldepth 7 score mate 4 time 51 nodes 1644 nps 32235 hashfull 0 tbhits 0 pv d7d6 c7d6 d8d6 e6e7 c3f6 e7e8 d6d8
info depth 5 seldepth 7 score mate 4 time 52 nodes 2071 nps 39826 hashfull 0 tbhits 0 pv d7d6 c7d6 d8d6 e6e7 c3f6 e7e8 d6d8
info depth 6 seldepth 7 score mate 4 time 53 nodes 2709 nps 51113 hashfull 0 tbhits 0 pv d7d6 c7d6 d8d6 e6e7 c3f6 e7e8 d6d8
info depth 7 seldepth 7 score mate 3 time 54 nodes 3741 nps 69277 hashfull 0 tbhits 0 pv c5c6 b4c3 d8e8 e6f6 d7f7
info depth 8 seldepth 8 score mate 3 time 55 nodes 4435 nps 80636 hashfull 0 tbhits 0 pv c5c6 b4c3 d8e8 e6f6 d7f7
Rasmus Althoff
https://www.ct800.net
JVMerlino
Posts: 1404
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: Engine test: tricky 3-move mate

Post by JVMerlino »

The King finds it in less than 1600 positions searched:

Code: Select all

Time	Depth	Score	Positions	Moves
0:00	1/3	Mate03	1598		1.c6 Qxc3 2.Re8+ Kf6 3.Rxf7#
0:00	1/4	Mate03	2839		1.c6 Qxc3 2.Re8+ Kf6 3.Rxf7#
0:00	1/5	Mate03	6521		1.c6 Qxc3 2.Re8+ Kf6 3.Rxf7#