jump to navigation

An old self-mate September 6, 2007

Posted by dorigo in chess, games, personal.
7 comments

Disappointing 97% of my readers, my production today is a post about chess. And not only that: not even about real chess, but rather a chess problem. Worse: it is a rather uncommon problem stipulation, one according to which “White moves and forces black to mate white in x moves“. X=5 in the case of the position below. I found the diagram while putting order in my library - expect more finds in the near future!

Self mate in 5 - T.Dorigo, January 9th, 1999

I composed this problem in 1988, but revised it eleven years later to make it prettier. A look at the position is enough to show that, according to ordinary parameters, white is winning hands down, his material superiority being ridiculous. But that is of no interest here: what matters is the mechanism by which black is driven by force to actually win.

So a second glance suffices to see that in five moves, only the black bishop in d8 has a chance to be the piece administering mate to the white king. But how can white force it ? And how does white traps his own king in order to have no escape out of the long diagonal a1-h8, where a check by the black bishop is the only chance ?

The solution is rather pretty and simple, once one understands the required moves of the mechanism: the black bishop has to be lured on the fifth rank in order to parry a discovered check by the h5 rook, administered by the white f5 bishop moving to b1, where it takes crucially away one white square from the white king. Then, it will have to capture the white bishop in c3, once the b4 knight moves away giving another discovered check.

Here is the solution, which is twofold:

1. Na8!

with two variations - black in fact has only two possible moves:
A) 1…., Bc7 2.Bb1+ (luring the B on the fifth rank) 2….,Be5 3.Rd6!! (the first finesse: white prevents black from the option of escaping the e file while leaving the long diagonal) 3…,exd6 4.d5! (opening the diagonal and closing the rank) 4….,B moves (d4, f6, g7, h8 are all possible, while Bxc3 wins one move earlier) 5.Nc2+ (the second discovered check, which forces the capture) 5…,Bxc3++ and white is mated. Note that no other move of the Nb4 would have worked in this variation. Nd5 defends the B, Nc6 gives mate, and Nd3 allows 6.Nb2!

B) 1…., Bxb6 2.Bb1+ (as before, but this time the black bishop is lured to c5!) 2….,Bc5 3.Qg5!! (the second finesse - white needs to disentangle the black bishop from the 5th rank pin before he can lure it to b4 and c3) 3….,hxg5 (black has no other move and is thus forced to take and unpin its own bishop) 4.Nc2+ (white now forces black to parry the check) 4….,Bb4 5.Ne1 (but almost any move is ok now!) 5….,Bxc3++ - black has no other move with the bishop, which is pinned in the c3-a5 diagonal, and so can’t help taking, and mating white).

There is one additional subtlety in the second variation which, in my opinion, makes this problem one of the nicest I ever composed. After 2….Bc5 why can’t white just play 3.Nd5?!!, unpinning the bishop and simultaneously giving check, forcing the unpinned bishop to move to b4 at once, and thus winning a tempo with respect to the original variation ?

The answer is simple: white has no way to then remove the Nd5, or he will be giving black a check again with the Rh5! So after 3.Nd5 Bb4 he cannot force black to mate him, since the Nd5 will capture the offending black bishop once it takes on c3…

There is also, alas, a flaw in the second variation, i.e. the multiple knight moves that white can make at move 4 and 5 to get mated. Oh well…

I know that maybe only one or two readers have been able to read to this point. I hope they, at least, have enjoyed this composition… And I hope the problem is correct! One never really knows, there are always things one may overlook when one composes chess positions. So this is a dry run before I submit it to a chess composition contest!

UPDATE: my last words above were sort of prophetic… The moment I posted the problem, I saw it is flawed. Not in a major way, but for a purist like me it is still not acceptable. The first move could just as well be 1.Ne8! In fact, the knight does not participate in the following events in any way. So hereby I propose a better version, see below:

Self mate in 5 - T.Dorigo, September 6th, 2007

Nothing much has changed, but for the absence of the Nc7 and the white queen having been placed in e1 rather than in h4. This works better! First of all, there is now only one possible initial move to solve the problem, namely 1.Qh4!! (otherwise in variation B white has no means of forcing an interference in the fifth rank). Second, the key move is quite hard to find now!

UPDATE 2: Oh, DUH! Wait. Now it simply does not work anymore: the white Rb6 has to be defended, otherwise the first move by black is 1….Kxb6 and things do not work anymore. So the original version only has one key, namely 1.Nc7-a8, defending the Rb6. Fine. I still prefer the cryptic manouver 1.Qe1-h4 though, so by adding a knight in a8 to the second diagram, we get to a reasonable -and working- position:

 

Self Mate in 5 - T.Dorigo, September 6th, 2007

Ok, I would now go back to work, so I really hope this works… 1.Qe1-h4!! is the key, with the followup I already explained above.