Lapierre's funny CHESS PROBLEM [SPOILER]
Philippe Schnoebelen
Much thanks to Mario for posting this fun puzzle !
>>>>> "MV" == Mario Velucchi <velu...@cli.di.unipi.it> writes:
In article <3iahi1$u...@serra.unipi.it> velu...@cli.di.unipi.it (Mario Velucchi) writes:
MV> A. LAPIERRE
MV> Themes/64 1959
MV> 4th Hon. Mention
MV>
+---+---+---+---+---+---+---+---+
| | . | | . |*K | . | |*R |
+---+---+---+---+---+---+---+---+
|*P | |*P | |*P | |*P | |
+---+---+---+---+---+---+---+---+
| |*P | | . | P | . | | . |
+---+---+---+---+---+---+---+---+
| . | | . | | . | | . | |
+---+---+---+---+---+---+---+---+
| | . | | R | | . | | . |
+---+---+---+---+---+---+---+---+
| . | | . | | . | | P | |
+---+---+---+---+---+---+---+---+
| | . | | . | | P | | P |
+---+---+---+---+---+---+---+---+
| R | | . | | K | | . | |
+---+---+---+---+---+---+---+---+
Mate in 2.
This problem is a classic ! I find it much fun, not too complex, and quite
pedagogical. There is a (small) trick so that it is always a delight to
show this problem to your friends at the chess club (or wherever). Try it !
The solution is easy: after 1. Rad1!?, B can't prevent 2. Rd8 mate. Can't
he ? B can avoid being mated with 1 .. O-O.
Is this a valid defense? A rule (a convention) in chess problems is that
any side may castle provided it is not possible to prove (through a
retro-analysis) he has lost the right to castle.
Let's see. At first sight, nothing tells us that B may not castle
anymore. Same thing for W. But look at WRd4. If it is the original WKR,
then W can't castle anymore because the Ke1 had to let this R out or the SE
corner. Similarly, if it the original WQR, W can't castle anymore. Now if
it is a promoted R (why not ?) then B can't castle anymore because the BK
(or the BKR) had to move and let this promoted WR out of the 8th rank.
So that we conclude that B and W can't both still be allowed to castle, but
there is no way to tell which one did loose his castling rights. In this
specific case of "Mutually exclusive castlings", the general convention is
that whoever castles first is allowed, therefore forbidding castling by the
other side !!
This gives us the intended solution:
1. Rad1 ? O-O ! and no mate
1. O-O-O ! ... 2. Rd8# ( 1... O-O now forbidden)
Nice trick ! Highly illustrative of the Mutual Exclusion rule.
Now a question: do you know richer examples of this Mutual Exclusion
scheme ? E.g. a problem with all four castlings, involved in some threat or
some defense, but with some mutual dependencies.
--Philippe
Wei-Hwa Huang
a legal move, then White would have no mate in two. Is that allowed?
In that case, I think the question should have been "Given that White
can mate in 2, what is the mate?" I don't think the original
stipulation of "White to mate in 2" is sufficient.
Joost de Heer
Yes it is. Someone can castle in a composition UNLESS it can be proven that
he (or she?) can't castle. for instance : in the given position white may
castle. If white had 8 pawns on the board, white couldn't castle, because
the rook on d4 would have been an impossibility. Hope this makes things
clear. Btw another problem which includes castling:
--------
--------
-k-P----
--------
--------
--------
-r------
R---K---
White wins (forgot the name of the composer :( )
Joost de Heer
(joo...@sci.kun.nl)
--
------------------------------------------------------------------------------
He ... wat leuk ... een .sigje.
En hij is van joo...@sci.kun.nl! Da's fantaaaaastisch zeg
------------------------------------------------------------------------------
Philippe Schnoebelen
Wei-Hwa> In other words, the winning move is O-O-O because if that were
Wei-Hwa> not a legal move, then White would have no mate in two.
More precisely: the winning move is 1. O-O-O because this move forbids that
Black answers with 1 ... O-O.
And it forbids this because, in chess problems, castling rights are not
determined _exactly_ as in chess games.
A difference is that in chess problems you don't necessarily know the exact
history of the position, so that the rule says you can castle unless it can
be proved that you can't. And what can and can't be proved may vary
e.g. when the other side does castle, putting more constraints upon this
unknown history.
Wei-Hwa> Is that allowed?
It's the convention used in chess problems. Lapierre's problem is a highly
illustrative example of what this convention implies. Really weirder
implications exist: witness the AP rule !
Wei-Hwa> In that case, I think the question should have
Wei-Hwa> been "Given that White can mate in 2, what is the mate?" I
Wei-Hwa> don't think the original stipulation of "White to mate in 2"
Wei-Hwa> is sufficient.
The original stipulation is sufficient only when you are familiar with the
(admittedly obscure) convention for mutually exclusive castlings in chess
problems. (You should be if you like retro-problems :-).
Now it is true that, for the average chess player, the problem looks like a
practical joke a la Sam Loyd. That's why I find it fun to show to the
friends at the chess club.
Hope all these messy explanations make some sense !
--
Philippe SCHNOEBELEN, LIFIA, 46 Av Felix VIALLET, 38000 Grenoble, FRANCE
email: p...@lifia.imag.fr
"Algebraic symbols are used when you do not know what you are talking about."
Joost de Heer
>In <3ifouq$4...@gap.cco.caltech.edu> whu...@cco.caltech.edu (Wei-Hwa Huang) writes:
>>In other words, the winning move is O-O-O because if that were not
>>a legal move, then White would have no mate in two. Is that allowed?
>>In that case, I think the question should have been "Given that White
>>can mate in 2, what is the mate?" I don't think the original
>>stipulation of "White to mate in 2" is sufficient.
>Yes it is. Someone can castle in a composition UNLESS it can be proven that
>he (or she?) can't castle. for instance : in the given position white may
>castle. If white had 8 pawns on the board, white couldn't castle, because
>the rook on d4 would have been an impossibility. Hope this makes things
>clear. Btw another problem which includes castling:
>--------
>--------
>-k-P----
>--------
>--------
>--------
>-r------
>R---K---
>White wins (forgot the name of the composer :( )
Found it. It's A. Selezniev, Tidskrift for Schack 1921. And the black
king is on b7 in original composition. But that doesn't change the idea.
Ian T Zimmerman
Philippe Schnoebelen <Philippe.S...@imag.fr> wrote:
>>>>>> "Wei-Hwa" == Wei-Hwa Huang <whu...@cco.caltech.edu> writes:
>
> Wei-Hwa> In other words, the winning move is O-O-O because if that were
> Wei-Hwa> not a legal move, then White would have no mate in two.
>
>More precisely: the winning move is 1. O-O-O because this move forbids that
>Black answers with 1 ... O-O.
>
>And it forbids this because, in chess problems, castling rights are not
>determined _exactly_ as in chess games.
>
>A difference is that in chess problems you don't necessarily know the exact
>history of the position, so that the rule says you can castle unless it can
>be proved that you can't. And what can and can't be proved may vary
>e.g. when the other side does castle, putting more constraints upon this
>unknown history.
>
>
> Wei-Hwa> Is that allowed?
>
>It's the convention used in chess problems. Lapierre's problem is a highly
>illustrative example of what this convention implies. Really weirder
>implications exist: witness the AP rule !
>
This is the easy answer. Things are not _that_ simple. First, there
are problems, labelled PRA (Partial Retrograde Analysis) where you're
told that there are several possible histories of the position and you
have to provide a solution for _every_ one of them. For example,
sometimes it can be proven that Black can castle either long or short,
but not both. (Trivially, this is the case when B has just Ra8, Ke8
and Rh8 and it's W to move.) A situation is even possible where W and
B castling is mutually exclusive, like in the present problem, _but_
if B can (and W can't) castle, something else in the game history
makes a solution possible. Example:
Luigi Ceriani, Europe Echecs 1960
r3k1S1/p1p3pS/R3P2p/1pP5/B6P/QP6/1PPP4/s3K2R
(for people who aren't British problemists :-)):
WKe1,Qa3,Ra6,Rh1,Ba4,Sg8,Sh7,Pb2,b3,c2,c5,d2,e6,h4;
BKe8,Ra8,Sa1,Pa7,b5,c7,g7,h6.
Mate in 2, PRA. Solution: Either W can castle, then 1.00, or else B
can castle but W can't, then 1.cb6 e.p.
Thus, our problem would be unsound if the stipulation included PRA, as
the case where B can castle doesn't have a solution.
In addition, some problemists, admittedly a minority (including
Ceriani :-)) feel or felt that the usual convention is wrong and that
PRA should be the norm. Hard to say, because there are amusing
problems based on either convention.
There's an exhaustive (and amusing) article on these matters in the
July 94 number of The Problemist (_Abominable Problems_ by Ronald
Turnbull). Things get really interesting, with more room for
controversy, when one introduces nonstandard rules (as, eg. "no piece
can move twice in a row").
--
Ian T Zimmerman +-------------------------------------------+
P.O. Box 13445 I With so many executioners available, I
Berkeley, California 94712 I suicide is a really foolish thing to do. I
USA <i...@rahul.net> +-------------------------------------------+
Greg Kennedy
: >In <3ifouq$4...@gap.cco.caltech.edu> whu...@cco.caltech.edu (Wei-Hwa Huang) writes:
: >>In other words, the winning move is O-O-O because if that were not
: >>a legal move, then White would have no mate in two. Is that allowed?
: >>In that case, I think the question should have been "Given that White
: >>can mate in 2, what is the mate?" I don't think the original
: >>stipulation of "White to mate in 2" is sufficient.
: >Yes it is. Someone can castle in a composition UNLESS it can be proven that
: >he (or she?) can't castle. for instance : in the given position white may
Okay, fine. But this is completely arbitrary, and certainly not obvious
to the casual reader of r.g.c.
A chess problem can also be something taken from a real chess game (which
uses FIDE or USCF rules). In that case, it would be wrong to _assume_
anything about the position which is not actually known.
Why not simply stipulate:
"White can castle either way, Black cannot. No en passant possible."
After typing in maybe two dozen lines of ASCII to represent the
chessboard, how tough can it be to add one more of text so the position
is complete....
Note that the threadname is not composition, but chess problem.
Greg Kennedy
Norbert Petsch
: > Btw another problem which includes castling:
: >--------
: >--------
: >-k-P----
: >--------
: >--------
: >--------
: >-r------
: >R---K---
: >White wins (forgot the name of the composer :( )
: Found it. It's A. Selezniev, Tidskrift for Schack 1921. And the black
: king is on b7 in original composition.
: But that doesn't change the idea.
OH NO, that's essential. With king on c6 Rd1 would do.
With king on b7 you need 0-0-0 to win a tempo in order to avoid that the
b king can block the w pawn.
_________________
Norbert Petsch
pet...@mi.uni-erlangen.de
Noam Elkies
It works with Kb6 too: 1.Rd1? Rh2! 2.d7 Rh1+ etc. Likewise 1.0-0-0? Ra2!
--Noam D. Elkies (elk...@math.harvard.edu)
Dept. of Mathematics, Harvard University
Joost de Heer
>Ok, let me try to join two threads together here (clue!).
>The following idea just occurred to me:
> -----B--
> --------
> --------
> --------
> -------R
> -k------
> -------R
> ----K---
>White to play and mate in two. I'll post the solution in a couple
>of days. Yes, there is a solution.
>
> Phil
Okay .... stupidity rules in this solution :)
One of the rooks in a promoted one. White has given odds the rook on a1.
White does the famous "ghost castle" Ke1-c1. Black has to do Kb3-c3 and
white mates with Rh2-h3#
PS :) This is only to show my incapacity of solving problems :)
--
This is my rifle. There are many like it, but this one is mine.
My rifle is my best friend. It is my life. FEAR
I must master it as I must master my life. FACTORY
Without me my rifle is useless. Without my rifle I am useless.
P R Watson
: In article <3j295...@uni-erlangen.de> pet...@mi.uni-erlangen.de (Norbert Petsch) writes:
: >Joost de Heer (joo...@sci.kun.nl) wrote:
: >: In <D4GMA...@sci.kun.nl> joo...@sci.kun.nl (Joost de Heer) writes:
: >
: >: > Btw another problem which includes castling:
: >
: >: >--------
: >: >--------
: >: >-k-P----
: >: >--------
: >: >--------
: >: >--------
: >: >-r------
: >: >R---K---
: >
: >: >White wins (forgot the name of the composer :( )
: >
Ok, let me try to join two threads together here (clue!).
P R Watson
: >Ok, let me try to join two threads together here (clue!).
: >The following idea just occurred to me:
: > -----B--
: > --------
: > --------
: > --------
: > -------R
: > -k------
: > -------R
: > ----K---
: >White to play and mate in two. I'll post the solution in a couple
: >of days. Yes, there is a solution.
: >
: > Phil
: Okay .... stupidity rules in this solution :)
: One of the rooks in a promoted one. White has given odds the rook on a1.
: White does the famous "ghost castle" Ke1-c1. Black has to do Kb3-c3 and
: white mates with Rh2-h3#
: PS :) This is only to show my incapacity of solving problems :)
: --
Very good indeed Joost! Well done.
I suppose I'm trying to satirise the 'problem mentality' (gently).
Thanks for taking it in good part.
Phil
P R Watson
: Joost de Heer (joo...@sci.kun.nl) wrote:
: : In <3jfjfs$e...@columbia.acc.brad.ac.uk> prwa...@comp.brad.ac.uk (P R Watson) writes:
: : > -----B--
: : > --------
: : > --------
: : > --------
: : > -------R
: : > -k------
: : > -------R
: : > ----K---
: : >White to play and mate in two.
: : >
: : One of the rooks in a promoted one. White has given odds the rook on a1.
: : White does the famous "ghost castle" Ke1-c1. Black has to do Kb3-c3 and
: : white mates with Rh2-h3#
: Very good indeed Joost! Well done.
Mark Crowther has told me he's seen a problem by the great Vassily Smyslov
using the same theme. So it seems I'm not so original after all. I'd
really like to see the Smyslov problem - can anybody help?
Phil
SCP
"Not so original" is putting it mildly. The new president of the FIDE PCCC
(Permanent Commission for Chess Composition), Bedrich Formanek, in a mini-
lecture at the Belfort (1994) meeting presented an 1863 problem (Arnold
Pongracz, Schachzeitung xii/1863) with the ghost castling trick. But it does
not take much scholarship to realize that an idea involving odds games must
date from the era in which such games were still popular. In fact, I suspect
there are examples earlier than 1863.
Although it is by no means a great problem, I quote the position below for
its historical interest. (Note: I have not tested it, so it may well be
cooked.)
Selfmate in 6; 16/1Q6/p3p3/2p1B3/2k2P2/2S5/2B1K3 (6+4)
Intention: 1.Be3 a4 2.0-0-0!?! a3 3.Sa1 a2 4.Bd4+ ed4: 5.Bc2 d3 6.Bd1 d2#
As there was some rude reference to "problem mentality", I should add that
it is player and not problemist mentality that is amused by such things.
S. Pantazis
US Problem Bulletin
Joost de Heer
[snip]
>Selfmate in 6; 16/1Q6/p3p3/2p1B3/2k2P2/2S5/2B1K3 (6+4)
Something seems to be wrong here. There are only 7 lines here, and the 1st
set of numbers seems to miss a letter. Please repost this problem! I'd like
to see it!
>Intention: 1.Be3 a4 2.0-0-0!?! a3 3.Sa1 a2 4.Bd4+ ed4: 5.Bc2 d3 6.Bd1 d2#
> S. Pantazis
> US Problem Bulletin
Joost de Heer
(joo...@sci.kun.nl)
--
SCP
>
>In <3jko59$l...@news.acns.nwu.edu> cpiz...@merle.acns.nwu.edu (SCP) writes:
>
>[snip]
>
>>Selfmate in 6; 16/1Q6/p3p3/2p1B3/2k2P2/2S5/2B1K3 (6+4)
>
>Something seems to be wrong here. There are only 7 lines here, and the 1st
>set of numbers seems to miss a letter. Please repost this problem! I'd like
>to see it!
>
Nothing wrong; 16/ = 8/8/ as you have probably realized by now. If I had
written 32/ would you assume that I had forgotten three rows and three
chessmen, instead of crediting me with meaning what I wrote, namely that
there are 32 empty squares?
>>Intention: 1.Be3 a4 2.0-0-0!?! a3 3.Sa1 a2 4.Bd4+ ed4: 5.Bc2 d3 6.Bd1 d2#
>
>> S. Pantazis
>> US Problem Bulletin
>
>Joost de Heer
>(joo...@sci.kun.nl)
>
>[tasteless sig deleted]
SP/USPB