[Retros] Accidental retro
mri_two at t-online.de
Wed Feb 8 09:14:43 EST 2006
>> ... Lajos Szász (1906-1929).
> 1906-29 -- died at 23!?
These data are taken from Giffard/Biénabe "Le Guide des Échecs",
Chapter 'Dictionnaire des Problémistes', so I assume them to be
> It would be nice if the book also listed sources for
> the various positions so that one could check the original.
The book lists the sources, but without reference to the problems.
>> ... "Illegal position. Black has no last move."
> I wasn't aware that computer programs can now check this.
I do not claim that my program can do that. The only thing I can
state, is, that under some circumstances my program is able to
detect illegal positions.
In #2940, it is easy to see (i.e without deeper analysis, simply
by the fact, that none of black's pieces has a square, from
which it could have moved to its current place), that black has
no last move.
> Is yours the first such program?
The first retroanalytical program I'm aware of was mentioned in
an article in the 'International Journal of Man-Machine Studies',
1988, 29, pp.97-112.
The authors B.E.P. Balden and M.A. Bramer describe 'RETRO', an
"expert system for solving retrograde-analysis chess problems by
means of heuristic methods".
To illustrate the abilities of their program, they used some problems
of Smullyan's Sherlock Holmes book, e.g. the program was able to
But this program was only a (university) research project (?).
The first publicly available retro-program was 'Retractor 1.0',
written by Theodore Hwa and Chad Whipkey (around 1998).
Nowadays their might be more programs, I know of one ('Pacemaker')
for such an exotic thing as Anti-Circe Proca defensive retractors.
> Does it also recognize illegality of Castling in #1747 or #1942?
Yes, but that's easy in both cases:
#1942 directly matches a simple 'castling-preventing' pattern,
in #1747 the same pattern occurs, if all last black moves are investigated.
More information about the Retros