Website founded by
Milan Velimirović
in 2006

7:19 UTC
ISC 2020



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum General Nature of promotions
You can only view this page!
(1) Posted by Andrew Buchanan [Saturday, Jan 9, 2021 17:56]

Nature of promotions

(moved from another thread, as distracting there)

Phoenix is where an officer is captured, and *then* a friendly pawn promotes to that same kind of officer. But what if the pawn promotes *before* the officer departs? The term "phoenix" doesn't seem appropriate, since there is no rebirth, so several years ago the new term "prenix" was invented and accepted in the Retro Mailing List. In many situations, it can be difficult to tell which of these two events happened first. And is it really interesting anyway? Maybe the generic term "nix" can be used for "phoenix/prenix". *Some* term is needed, because our vocabulary has a significant hole, which the neighbouring term "obtrusive" tends to bleed into. The word "nix" means to eradicate, to "zero": thus the nixing of the original officer nixes the excess count of officers of its particular kind.

What's the bigger picture? The V&V Encyclopedia is pretty good here in identifying disjoint cases, but needs a little expansion. For any composition, we are supposed to examine the history leading up to the diagram, either as a brief retro check of legality, or more rigorously because there is retro-content to be understood.

Suppose a pawn has been promoted.

(1) It may have been captured later, before the diagram is reached.

(1.1) If we know what kind of officer ephemerally existed, we call this kind of promotion "CERIANI-FROLKIN".

(2) Otherwise the promoted pawn is still somewhere on the board.

(2.1) If no original officer of that kind has been captured, we must logically have excess officers of that kind. This is termed "NON-STANDARD MATERIAL", and means 3+ knights or rooks, or 2+ queens, dark-square bishops or light-square bishops. This is often regarded as a defect, depending upon the context.

(2.2) Otherwise some original officer must have been captured to keep the number standard.

(2.2.1) It may be trivially obvious from looking at a part of the diagram that a some specific kind of officer on the board is promoted: e.g. wBa8 and any of: wPe2g2 or bPb7 or bKb7. An example of the other main pattern is wRa8b8 wBc1 wPa2c2d2. In this case the promotion is called "OBTRUSIVE". This again is often regarded as a defect, depending upon the context.

(2.2.2) Or more careful retroanalysis may indicate a past promotion into a specific kind of officer still on the board. There has never been a single accepted term for this. "Phoenix/prenix" is accurate but clumsy, therefore the term "NIX" is proposed. Even for a non-retro problem, I don't think such a rare situation is regarded as a defect these days, and could be regarded as adding to the content.

The stories of multiple pawns of both players can interact, complicating the application of this simple taxonomy In many (most) cases we cannot infer all details of promotion history. The terms we have are intended to cover important cases when definitive conclusions can nonetheless be reached.

Comments/corrections/improvements welcome.
(Read Only)pid=20420
(2) Posted by Andrew Buchanan [Saturday, Jan 9, 2021 17:58]

There are more also specific terms. Schnoebelen promotion is a kind of Prentos promotion (= capture by officer) which is a kind of Ceriani-Frolkin promotion. Being Schnoebelen doesn't stop the promotion *also* being Prentos & C-F. There are many other thematic properties like Pronkin, Anti-Pronkin, Donati etc to classify. Joy.

However, before diving into the themes, I felt the need for a simple foundation for promoted material as a whole. Particularly, "obtrusive" is a vague and derogatory term, and it's unfair to composers & PDB curators for it to be used in the current way. Even V&V encyclopaedia is inconsistent about the term. There have been a number of complaints by composers and solvers about the current usage in PDB, and it needs fixing. Nixing :)

Composition has changed since 1900s when obtrusion was invented as a concept: retroanalysis has emerged as a respectable discipline in its own right. Sir Jeremy Morse, writing much later, was an apologist for the old ways: what to him are "obtrusive" (horrible usage) and "obviously obtrusive" (horrible phrase) I feel can be more reasonably termed "nix" and "obtrusive". And Morse was only one writer: others have different perspectives.

Beyond "obtrusive", there are also issues about incomplete knowledge, especially when one consider multiple pawns. Does it matter if we can't track which original pawn promoted on a certain square? Probably not, but what are we identifying? How to match the promoted units against the captures of the same kind of officer? Also when are we making our observations? I think we should focus on the problem diagram, rather than looking at e.g. units which were phoenix, and later got captured so would also be C-F. Or "future C-F" where promoted officers are captured in the forward play. Later can maybe have a model which allows for examination through past and future, but the visual and judgemental nature of many of these ideas suggests that we should start with the diagram, where the aesthetics are primarily being applied.

But I haven't got Pronkin solved. To take your example, suppose in a proof game, one player loses both rooks, and then promotes one pawn to another rook, which visits both corner squares and then is captured. What combination of Phoenixes, Pronkins & C-F should one award? I think in this case it's 1 C-F, 0 Phoenixes, and 2 quasi-Pronkins (= a core term I've just invented, meaning a promoted unit visits a valid start square for its kind, irrespective of the fate of its original occupant.)

Does it really matter that the original rooks have been captured in order to celebrate the visitation of a promoted unit on a1 & h1? If not, then "quasi-Pronkin" is the core achievement being logged". Then "Pronkin" can be seen as a derived term which means quasi-Pronkin + the fact that original wRa1 had already been captured? (Although does it matter that Ra1 had *already* been captured? Suppose it was captured *after* the visit, as a subset of Prenix? All we really care about here is that we achieve standard material for diagram purposes.) When you look at it like this: quasi-Pronkin seems a sensible achievement to track; Pronkin appears now cumbersome and charged with an unnecessary constraint, but I certainly wouldn't ask to change it. Defining it as a derived achievement of other primitives is an elegant solution

If anyone has a nice name to give the primitive "quasi-Pronkin", I would be grateful.

I don't want to lose the context in which historical problems were conceived, I enjoy the charm of old language, and I don't want to confuse people. Therefore while I'm happy to fill in holes, I am unwilling to change the meaning of any existing term, or the term of any existing meaning EXCEPT where what we have is already inconsistent or somehow offensive. Then it's fair game, imho.
(Read Only)pid=20421
(3) Posted by Joost de Heer [Saturday, Jan 9, 2021 18:43]

But I haven't got Pronkin solved. To take your example, suppose in a proof game, one player loses both rooks, and then promotes one pawn to another rook, which visits both corner squares and then is captured comes close (the rook isn't captured).
(Read Only)pid=20426
(4) Posted by Hauke Reddmann [Sunday, Jan 10, 2021 10:58]

@Andrew: if in doubt, make a graphic. In this case, a space-time
diagram (even on the danger that it begins to look like :-)
(Read Only)pid=20434

No more posts

MatPlus.Net Forum General Nature of promotions