[Retros] Fastest selfmates in 1?

[Elkies, Noam]

Rewan Demontay <rewandemontay at gmail.com> wrote:

> NIce save Noam, turning my 6.5 into a 6.0!  [ . . . ]
> And yes, the switchcback is a neat little gimmick.

Thanks.  With your permission I'd like to quote it in my eventual
follow-up to that Chess Stack Exchange question (authored by
"Rewan Demontay, version by Noam D. Elkies").

> There surely must be thousands, if not tens of thousands,
> of possibilities in 5.5 to be found in a Labelle search.

On the order of a thousand *games*, yes; but I'd have be surprised
if thousands of *positions* turned up.  Now that F.Labelle has taken up
this challenge, it seems quite unlikely that there'd be hundreds of
different positions now that the first 10% of his search turned up
only my new pair. It could even be that all 5.5-move positions
have already been discovered between the Krabbe diary entry
(including the Kalajdzievski variations) and my pair.

> The aforementioned 5.5 that saves a ply is such an example,
> provides for numerous 5th Black moves for numerous more solution[s].

I count only 8 choices for that extra ply: the single and double moves of
pawns a7,b7,c7 and the two moves of Nb8.  (Moving d7, e7, or
any Kingside piece would destroy the selfmate.)  Each of those works with
either 1 e3 or 1 e4, and likewise each of Blijlevens' positions has a variant
with 1 e3 or 1 ... e6; I see no other options, and thus count 22 known
positions so far: Blijlevens 2+2, Kalajdzievski 2*8, and my 2.
I expect that Francois will soon tell us if we've missed anything else.

NDE