[Retros] shortest proofgames of castling switchback

Noam Elkies elkies at math.harvard.edu
Tue Jul 17 18:29:35 EDT 2007

I wrote:

> Argh -- so the extra move b3 at the end can't be entirely random.

> Let's try Kf1 instead:

Unfortunately that's not enough, because Mario's cook allows
Black an extra move of slack too:

1.a2-a4 g7-g6 2.Ra1-a3 Bf8-g7 3.Ra3-h3 Bg7-d4
4.d2-d3 Ke8-f8 5.Qd1-d2 Kf8-g7 6.Qd2-h6 Kg7-f6
7.Qh6-f8 b7-b6 8.Rh3-h6 Bc8-a6 9.h2-h4 Qd8-c8
10.Rh1-h3 Qc8-b7 11.Rh3-g3 Qb7-f3 12.Rg3-g5 Qf3-h5
13.g2-g4 Ba6-b5 14.Bf1-g2 Nb8-a6 15.Qf8-b8 Kf6-g7
17.Bg2-b7 Ng8-f6 18.Bb7-c8 Kg7-f8 18.Bc1-f4 Bd4-e5
19.Ke1-f1 Kf8-e8

Sorry about the wild goose chase.

There's still the White 0-0 attempt (as modified with Re8 instead of Rf8),
which seems less vulnerable to such cooks because the bBh3 prevents Kg2
even after Qf1-b1. Any hope of computer checking for that one?

--Noam D. Elkies

More information about the Retros mailing list