[Retros] Promotion: legal order of partial actions

[raosorio at fibertel.com.ar]

Hi Guus,

Good analysis.  You wrote,

"Let's assume that the FIDE only ratifies (partial) acts which are legal immediate successors to previously ratified acts"

This a solid interpretation.  On the other hand, I've already suggested the following interpretation of the
article 4 spirit:

"Article 4 is not a penalties one, but an article trying to minimyze the deviation between illegal moves
and their consequent forced legal moves (You touched that piece?, then you have to move it. You
made half castling?, then, complete the castling, etc). In other words, article 4 tries  to keep the game 
as clean as posible with no penalties but just restrictions".

Based on this interpretation, after a promotion made in an illegal (formally due to the order or 
completeness) way, that move would be forced after retraction 7.4 in order to "minimyze the deviation".

Honestly, I'm not that interested in the rules repecting the real game but in their application to problems.
This promotion issue is being an obstacle to close the definition of "Minimal Deviation from the Rules"
I'm working on. I'm trying to keep MDR, far away from being fairy, as an "ultraortodox convention" since it
wold apply the "Laws of chess" in a wider scope than the traditional convention.

Of course I could restrict the partial promotions to the "legal order", but then we would loose a very
rich field. Let's interpretate a promotion move as a sequence of partial actions (VISIBLE IN THE DIAGRAM
after each part), and identify the incomplete move cases,

* Straight promotion: compossed by two parts, a) to remove the pawn from the board  b) to install
   the promoted piece on the promotion square. There are two incomplete move cases: YES-NO /
   NO-YES.

* Capturing promotion:  composse by three parts: a) to remove the opponet's piece from the board
   b)  to remove the pawn from the board  c) to install the promoted piece on the promotion square.
   There are three (four?) incomplete move cases: YES-NO-NO / YES-YES-NO / YES-NO-YES.

The point here is that these incompleteness cases allow nice diagrams with good retrocontents.
I've tested some examples with some of our retrocolleges and the ideas are easily taken despite
of the "order". A simple example,

R4bB1/p1pppPpp/Pp4p1/1N6/8/P7/kPP1P3/1RK5
How does this end?
  b) wRa8 to h8 

The twin(b) (YES-NO-NO) would be correct while twin (a) (YES-NO-YES) would be not, unless the
alternative article 4 interpretation were taken. I addition, the Finzer's problem published in 
Die Schwalbe would be incorrect.

Some suggestion to keep all the incompleteness cases included and still be framed by the Laws?

Roberto Osorio


Hi Roberto, 

Some decades ago Tim Krabbé wrote his famous "Schaakcuriosa" booklets in which he made a similar observation - though the rules have changed a bit since then. In a position where Queening a Pawn would stalemate the game, his devilish personae would take a Queen out of the box, place it innocently on the promotion square (not releasing it), and then with sudden insight withdraw it and place the winning Rook instead! To be absolutely sure of the support of the referee he would secretly stir his coffee with that Rook during the proceedings so he could always claim that he touched the Rook first! 

This in itself funny story also offers a lesson to learn. Reading the players intention to make the Queen promotion is complicated by actions outside the direct environment of the chess board such as stirring the coffee. We therefore need to restrict which types of acts, spaces and objects can be part of the rule evaluation process. I dislike to include "touching the Queen or any other off-board promotion piece" as acceptable evidence, but "placing and releasing the piece on a possible promotion square" could be reasonably allowed in. On that point I agree with you that the rules look somewhat illogical. 

Nevertheless I can also see a logical principle underlying the current FIDE approach. Let's assume that the FIDE only ratifies (partial) acts which are legal immediate successors to previously ratified acts. Another way to put it is by saying that you cannot assemble legal bits and pieces that are scattered over the time track in order to deduce the intended (or enforceable) move. In your cases a) and b) the required bridging act of moving up the Pawn to the promotion rank is missing and the Queen-placing act is therefore taken out of the equation as an isolated event. Compare that to a regular example like: white Qd1 is picked up and dropped on h6; on the way back, the players hand touches a black Bishop on g1. What to do? The first legal act was: touching and picking up the Queen. It was followed by the illegal act of dropping the Queen on h8, which must therefore be retracted. The subsequent touching of Bg1 is a legal immediate successor to the previous legal act, and imp
lies that the Bishop must be captured. Resulting action: Qd1xBg1. 

Whether or not this is the actual FIDE philosophy I do not know. It is just the one that comes to mind when reading the FIDE articles. 

Best, 
Guus Rol.