[Retros] Fastest selfmates in 1?

[Elkies, Noam]

Rewan Demontay <rewandemontay at gmail.com> writes:

I didn't know or forgot that you're on this mailing list too!
The reason I haven't yet given your answer the "accepted" bonus,
nor posted my own solution, was that I had hoped to include a definitive
statement that there's no 5.0, and perhaps a complete list of 5.5's.

> I still cannot find your 5.5 Noam, blind I seem,
> but my Krabbe find was indeed a swirl of luck.

My 5.5 is somewhat tricky, and perhaps lucky . . .
Of course if it's within the range of F.Labelle's data
then it will be there, together with the three in Krabbe #153
and any others that might exist.

> And you say that a unique PG is beyond the power of a
> computer search? Then bwahahaha, for I am now your new overlord!

Beyond *exhaustive* computer search, like many other tasks
of this flavor such as the shortest help-stalemate.  But perhaps not,
seeing that your nice construction is a half-move longer than necessary:
it remains sound with the last two plies replaced by 6 . . . Qg5+
<FEN: rnb2bnr/pppp1ppp/4pk2/6qK/4P1Q1/8/PPPP1PPP/RNB2BNR>
(adding a Black switchback)!  Possibly the capture-free constraint
brings it within the range of F.Labelle's tables, and if not then
a complete list of 5.5's would still show that 6.0 is minimal for
a selfmate with a unique proof game (assuming none of the 5.5's
have that property).

NDE