[Retros] MDR / postulates / hierarchy

Rol, Guus G.A.Rol at umcutrecht.nl
Tue Mar 4 04:10:41 EST 2008

Great reply, Roberto! If every composer upholds your standards we can
progress in a manner that is not only artistic but scientifically sound
as well. Sorry if I missed an earlier presentation of the MDR
definitions; these will really do the trick of making the issues
decidable. In the future I may try to shoot some holes in them (which
you won't mind) but I guess I will really have to zoom in the details
for that.

Best, Guus.

-----Oorspronkelijk bericht-----
Van: retros-bounces at janko.at [mailto:retros-bounces at janko.at] Namens
raosorio at fibertel.com.ar
Verzonden: maandag 3 maart 2008 23:44
Aan: retros at janko.at
Onderwerp: [Retros] MDR / postulates / hierarchy

(2nd try)

Hi Guus,

Many thanks for your thoughts. This interaction is being very fruitfull
to adjust the ideas.
I take your "decidable". concept, very clear.
Let me present again the MDR basic definitions in a more organized way.
Retroanalyzable Deviation from theRules: is the deviation resulting from
a move that does not respect the article 3 definitions (one or more)
SOLELY in terms of,

* Opportunity ( checks, rights to castle and to capture e.p.)
* Completeness (the move does not make all its parts)
* Sequence (the move does not respect the established order of its

PARTS: I mean here entire parts, i.e, showing a changed diagram after
it's done (a photo, not a partial desplacement. Qa1-a8 to capture a
piece on a8, not a fraction as Qa1-a5)

Note that these moves are geometrically correct (pieces move from a
square to another according article 3). This definition produces 16 well
defined cases (4 related to full moves illegal due to checks or rights,
4 related to underway castling and captures, 3 to underway straight
promotion and 5 to underway capturing promotion). An underway move, even
if it's legal, is treated here as a deviation.

After this, MDR is a convention trying to organize all the animals in
the zoo in order to be a clear presentation for solvable problems.

Minimal Deviation from the Rules (MDR): convention that ALLOWS to
present position with Retroanalyzable Deviation from the Rules,
establishing that the correct solution of the problem is the one showing
the minor deviation according to the following hierarchical order: 1)
the move is complete 2) the move is legal 3) the move is the less
incomplete one.

"The move is the less incomplete one" means the one showing the higher %
of its parts done. MDR is a retroanalysis convention so all the genre
conventions apply here. With this, the definition and articles 1 to 5 of
the Laws of Chess (Codex), the following corollaries apply,

1) The deviation can't be produced after mate, DR, stalemate, ...
positions, due to the automatic application of the rules that stop the
2) Only the last move can produce deviation (same reason as before)
3) In case of illegality, 7.4.a retraction applies and the forward play
is conditioned by article 4.
4) In case of underway legal move, article 4.6 applies, forcing to
complete the move.
5) In case of multiple solution alternatives, the correct solution is
determined according to the following preference: a) legal full move b)
illegal full move c) the less incomplete legal move d) the less
incomplete illegal move.

It should be noted that MDR does not force the deviation, it just alert
the solutionist on this possibility that has to be proved by
retroanalysis. This makes posible the deviation /non deviation

I think that after this , positions of the MDR type are clear. The
hierarchy is just to isolate the correct solution and it's apropiated
for a wide range of problems,including the BG retro type and even
Finzer's. However, Finzer's type problems have a different solution
under MDR since this applies the 7.4.a retraction in the cases, for
instance, of illegal sequence promotions (show in one of our examples),
instead of the tolerance asked by Lothar.
The "minimal" deviation issue: of course it doesn't try to be an
absolute concept. It means minimal according to the given preferrence,
which has been conceived to be a poper frame for problems dealing with
illegal moves. Please note that after the Retroanalyzable Deviation
definition the cases are well defined. So, we are not dealing with every
type of illegalities. The 16 cases are the list you asked for (these
cases can be deduced form the definitions, one or two perhaps with some
polemics). Of course, they will be presented.
Thanks again for this. Any further comments very wellcome.

More information about the Retros mailing list