[Retros] Variables

[Joost de Heer]

> That makes things more interesting Joost. There is more to Stephens

> comments than just the "Variable Piece". He also adjusts the notation

> of moves by writing down "Vc4-b3" instead of "Vc4xb3" or "Vc4xb3

> e.p." as would be usual. That introduces new unknowns (captures)

> besides the Variable Piece. One would need to peel off the layers to

> see if there is anything AP hidden inside (which I doubt). Note that

> the e.p. capture needs no justification since the preceding double

> step move is a 100% certainty. The playing options of a Pawn are

> therefore as clearly defined as those of any other piece and need no

> justification.



> By the way, I've encountered PbyD (prove it by doing it) before and

> I've always wondered if that could mean something different from

> PF-strategy. If it included arbitrary e.p. captures as well, what

> would they need AP-logic for?


In the given diagram, Vb2-b4 Vc4-b3 could be Rb2-b4 Bc4-b3 too (and if 
the position was totally different, it could also mean Pb2-b4 Kc4-b3, or 
Qb2-b4 Qc4-b3).

To quote from the article:

"In fact, Tadashi[ Wakashima, the inventor of Variables, JdH]'s 
intention is that the moves should be written as 1. Vb2-b4 Vc4-b3. As 
this is compatible also with 1. Rb2-b4 Bc4-b3, it cannot simply be 
asserted that capture has taken place; only departure and arrival 
squares can be claimed. If c4 was a P (or a V that has already been 
shown to be a P), then c4-b3 is consistent only with ep capture."
[...]
"The same justification must be made for promotions; in C [wVe1 wVh1 
wVh7 bKg7] one cannot claim to play h7-h8=S, but instead 1. Vh7-h8 only 
(note that this shows the wVh7 to be a pawn); the particular promotion 
must be justified by later play. And Ve1-g1 is not sufficient to justify 
this move as OO; but if followed by Kg7-g6 the alternatives Re1-g1 and 
Qe1-g1 are ruled out leaving only OO, so we know that g1=K and Vh1 is 
now on f1 and is a rook."

As I said before, don't confuse Variables AP logic with retro AP logic. 
For variables, later moves can be used to rule out possibilities, or to 
(re)move pieces that aren't directly affected by the variable move 
itself. The AP terminology is used in the latter sense.

Joost