Is there any position, where moving a pawn to last rank but not promoting it to any piece would be the best move?

By the way, I know this is an illegal move. Assume it is legal.

Edit: Now we have a nice answer from Dag Oskar which ends with stalemate. That's good. But I wonder, if there is winning position, winning by not promoting, and the only winning move is not promoting.

Edit #2: I think a winning position is impossible, as this seems to be occurable when opponent stalemates with promotion. If opponent is not stalemating, then there is no point of not promoting anyway. But if opponent stalemates with a knight, then bishop or rook doesn't stalemate, and etc. It's impossible to have a position where knight, bishop and rook stalemates. So, in every such position, there will be a correct promotion. Not promoting won't be the best move anyway.

Therefore, I'm accepting the answer.

Edit #3: Rosie F's answer demonstrates that it's possible to win with a 'not promoting move'. It's very entertaining to see it's possible. Thanks for the effort.

up vote 14 down vote accepted

The Matt Bengtson problem Prof. Elkies mentions is:

[Title "Matt Bengtson, Chess Braintwisters (Burt Hochberg), no. 103. White to move & draw."]
[fen "4kn2/3p1pPp/4pPpK/6P1/8/2p5/1b6/8 w - - 0 0"]

1.

The start of the solution-line can't be shown in PGN as White's 1st move is illegal: 1. g8=Black queen Qg7+ (anything else stalemates White) 2. fxg7 c2 3. g8=Q c1=Q 4. Qxf7+ Kd8 5. Qe8+ Kc7 6. Qc8+ Kb6 (Kd6 7. Qc5+) 7. Qc6+ Ka7/Ka5 8. Qb7/Qb5, drawing. White's queen is then taken for stalemate. Solution line taken from Chess Braintwisters.

(This is in answer to a comment by Prof. Elkies to Dag Oskar Madsen's answer, but, it seems, the syntax that activates the PGN-viewer in answers doesn't do so in comments.)

A similar oddity from the same book:

[Title "John Beasley, EBUR, 1996. White to move & win."]
[fen "5r1K/2k1P3/3N1Bqp/N2Q1B1P/8/8/8/1R6 w - - 0 0"]

This problem was reprinted as no. 107 in Burt Hochberg's Chess Braintwisters.

White is in check, and his only reasonable move is to capture on f8 (1. Qg8? Qxg8#). But 1. exf8=R allows stalemate after 1. ... Qg7+ because the new rook guards d8 after 2. Bxg7. Similarly, 1. exf8=N guards d7 allowing 1. ... Qh7+ and stalemate, and 1. exf8=Q/B guards d6 allowing 1. ... Qg8+ and stalemate. The answer is 1. exf8! (no promotion). Now 1. ... Qg7+ 2. Bxg7 releases d8; 1. ... Qh7+ 2. Bxh7 releases d7; 1. ... Qg8 2. Qxg8 releases d6. 1. ... Qxf6+ destroys the stalemate at once, and White wins. Solution line and discussion taken from Chess Braintwisters.

  • This cannot be the position. The question is not about promoting to the wrong color, it's about not promoting at all. – Dag Oskar Madsen May 27 '16 at 16:17
  • 5
    The second position seems to be the real deal, though. Amazing that it's possible! – Dag Oskar Madsen May 27 '16 at 17:36
  • 1
    @DagOskarMadsen I don't know if it's true but I read somewhere that even though "everyone knew" you couldn't promote to the wrong colour, it wasn't officially written into the rules until that puzzle was created and pointed out the error – undergroundmonorail May 28 '16 at 11:07
  • @undergroundmonorail Yes, there are some other questions on the site dealing with promotion to the wrong color, see for instance chess.stackexchange.com/questions/8291. This question however is about leaving the pawn unpromoted on the 8th rank. – Dag Oskar Madsen May 28 '16 at 11:13
  • 1
    @undergroundmonorail Thanks for the chess.SE reference. "Until modern times the St Petersburg laws were perhaps the only ones that stated that the promoted piece must be of the same colour as the pawn, and some problemists took advantage of the oversight." - David Hooper & Kenneth Whyld, The Oxford Companion to Chess, 1st ed. (1987) pub. OUP. I don't know of any such problem composed before "modern times". It intrigues me what these problems & their composers could be. – Rosie F May 28 '16 at 12:31

Here is an example:

[fen "8/4P3/8/4p1p1/2p3Pp/p4p1K/k1p2P1P/8 w - - 0 1"]

1. e8 leads to stalemate next move, while all legal promotions lose to 1... c1=Q followed by 2... Qf1#.

  • 4
    Good one. So, it makes sense because promoting enables moving, but a pawn on last rank can't move, so introduces stalemate ideas. I wonder if theres a position wins by not promoting. – ferit Jan 5 '16 at 22:53
  • Yup, this works. Same idea with 7 units instead of 13: move bKa2 to g1 and you can remove two White pawns (f2,h2) and four Black (a3,c4,e5,f3). 1 c8Q c1Q 2 Qb5(f7,f8) stops Qf1# but 2 . . . Qe3+ mates (and even 2 . . . Qf1+ still wins). – Noam D. Elkies Jan 7 '16 at 1:06
  • @NoamD.Elkies Can you confirm that a position where not promoting would be the only winning move is impossible? – Dag Oskar Madsen Jan 7 '16 at 1:26
  • 3
    Actually constructing such a position is very difficult but not impossible. Your argument shows that at least some of the promotions cannot result in immediate stalemate, but stalemate can still be forced later. The first example I know was constructed about 15 years ago by Matt Bengtson and published in Hochberg's Chess Braintwisters. I can't find the position immediately but can ask Matt for it. – Noam D. Elkies Jan 7 '16 at 3:36
  • @NoamD.Elkies That would have been great. I'm really curious about how such a position looks like. – Dag Oskar Madsen Jan 13 '16 at 12:36

Here's another problem where White to move loses, but moving a pawn to the 8th rank and not promoting would (if it were legal) draw.

"By the 19th century [...] it was thought that promotion should give the maximum possible advantage. It was claimed that in some cases the best move would be to remain a pawn, and Kling composed a position to show this:

[Title "Josef Kling"]
[fen "r7/1Pp5/2P3p1/8/6pb/4p1kP/4P1p1/6K1 w - - 0 0"]

If 1 bxa8=any piece, 1 ... gxh3 2 any h2#, but if the pawn on a8 remains a pawn then 1 ... gxh3 is stalemate." - David Hooper & Kenneth Whyld, The Oxford Companion to Chess, 1st ed. (1987) pub. OUP, s.v. promotion (slightly adapted). Dag Oskar Madsen's problem has the same tempo (not sure of the correct term here) as Kling's: White's only mobile unit is the pawn; promotion to any piece enables Black to #2, whereas by making the pawn immobile White immediately stalemates himself.

  • 1
    What about 1. bxa8 Kf4, what does white do? – Dag Oskar Madsen May 28 '16 at 13:38
  • I've edited to clarify that White to move is to draw (not win). Granted, Sabot's OP's 1st edit wondered if there was such a problem with White to move and win. But I felt that this problem was pertinent, especially as it is old and is part of the history of the law regarding promotion. Ack. 2. Kh2. Good point. So perhaps Kling's example was unsound after all. (My diagram is as in H&W.) – Rosie F May 28 '16 at 15:20
  • Ack. 2. Kh2. (Then perhaps 2. ... Bf2 3. h4 g1Q# or am I wrong?) – Rosie F May 28 '16 at 15:29
  • Nice example, thanks! – ferit May 30 '16 at 10:52
  • 1
    @Saibot You're welcome. But note Dag Oskar's 1. bxa8 Kf4 which allows 2. Kh2 -- unless I've missed something, this busts Kling's problem, so perhaps it wasn't as nice an example as I'd intended. – Rosie F May 31 '16 at 7:14

protected by Phonon yesterday

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.