Website founded by
Milan Velimirović
in 2006

16:09 UTC
ISC 2022



Remember me

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

Rating lists


MatPlus.Net Forum General Promotion to knight
You can only view this page!
(1) Posted by Geoff Foster [Friday, Jun 11, 2021 11:29]

Promotion to knight

In chess problems, is promotion to knight regarded as an underpromotion?
(Read Only)pid=21248
(2) Posted by James Malcom [Friday, Jun 11, 2021 16:09]

To my knowledge, yes, knight promotion is generally regarded as an underpromotion.
(Read Only)pid=21249
(3) Posted by Joose Norri [Saturday, Jun 12, 2021 00:47]

I can't think that way. What component of the queen's powers is deducted to get knight's powers?
(Read Only)pid=21252
(4) Posted by Andrew Buchanan [Saturday, Jun 12, 2021 10:25]

"Underpromotion" is one of those useful but imprecise words where different people have different ideas, varying over formats too. Everyone is right. It can't be given a single meaning. My own arbitrary preference is that it's most meaningful for studies and direct mates. The knight uncertainty is just one aspect.

That suggests that "underpromotion" be a derived rather than foundational term.

A problem database can offer a neutral base keyword e.g. "promotion" (potentially also differentiating retro play, diagram & forward play). Then people can craft their own queries to filter on the particular flavour that interests that day.
(Read Only)pid=21254
(5) Posted by Hauke Reddmann [Saturday, Jun 12, 2021 10:51]

If I would maintain a database...

- Ideal would be keywords that could be automatically added (in principle).
In this case it would be promoQ,promoS,... (even joke promotions could
be tagged this way) plus promoting side.
- Mind usefulness for the end user. ("en passant" should be split into black
and white, but rather not into line where it happens.)
- The Albrecht is quite satisfactorily in this matter (if I would *want* to start
bitching, I wouldn't end that fast, but then, it wasn't me who put in years
of hard work...)
- Such databases are prone to evolution; if you find out ten years later your
design decision was crappy, tough luck. (Who here types on a Dvorak? :-)
(Read Only)pid=21255
(6) Posted by Dmitri Turevski [Saturday, Jun 12, 2021 17:27]


All very good points. The solution I came up with, and that is basically already implemented in yacpdb, is that an entry has two independent sets of keywords: user assigned and auto assigned. One can search for one or the other or combine criteria.
(Read Only)pid=21256
(7) Posted by Andrew Buchanan [Sunday, Jun 13, 2021 02:39]

The YACPDB split is great. At the system risk of making things more complicated (but I think quite doable), I think there are 4 buckets. Also keywords need parameters.
1) keywords automatically populated by analysis of problem + solution (e.g. promotion(QQQRqbS) )
2) set of published rules indicating other derived keywords (e.g. promotion(QQQRqbS) => promoS, Ceriani-Frolkin promotion => Prentos, what criteria make a problem in the PDB fairy, or retro.)
3) shared controlled vocabulary to allow for non-computable things (e.g. task attempted) also over-ride over automated stuff (e.g. an Add Pieces problem may look like a miniature but it's not.) So the users set these up and manage them, but there is an expectation that they are a shared asset so people can edit others work.
4) personal free-form keywords anyone creates and uses like Twitter tags (what can possibly go wrong, heh?)
(Read Only)pid=21257
(8) Posted by Joost de Heer [Sunday, Jun 13, 2021 12:01]


In chess problems, is promotion to knight regarded as an underpromotion?

To push this even further: In fairy chess, should promotion to a rook or a bishop be considered an underpromotion? 'Under' implies that the motivation for the promotion is based that a bishop or rook is less than a queen because it only has partial queen movement, but in a lot of fairy variants (e.g. Madrasi or anticirce) the promotion is not based on the fact that a queen is too strong, but that a bishop has different properties.
(Read Only)pid=21258
(9) Posted by edderiofer [Sunday, Jun 13, 2021 15:31]

To push this even further, is a promotion to *any* non-queen piece underpromotion in fairy chess? What if it's a promotion to a piece that's "stronger" than a queen, like an amazon?
(Read Only)pid=21259
(10) Posted by Andrew Buchanan [Sunday, Jun 13, 2021 16:55]

Good points! And indeed there is also Dummy Pawn. Is that promotion at all? If it is, it’s an under-under promotion. This all points to the relativism of this jungle term under-promotion. We need to stick to the garden facts of what objects did what where, and allow users to apply their own favourite mappings to non-standard terms
(Read Only)pid=21260
(11) Posted by Hauke Reddmann [Monday, Jun 14, 2021 11:03]

...which is what happens when chess terms are taken over to
problem chess, sometimes it works, sometimes it doesn't.
(We could compile a list here :-)
(Read Only)pid=21263
(12) Posted by Andrew Buchanan [Monday, Jun 21, 2021 09:13]

This is a perfect case study for the garden/jungle perspective.

"Underpromotion" is a common jungle term, and as such does not have a single meaning. Fortunately, it's currently only a descriptive term, and does not impact any conditions. So we back off and try to define a solid data model for the garden. Then people will be able to define their own meaning, and filter problems sets accordingly.

It's immediately clear that "underpromotion" has nothing to do with the enjoyable activity of promotion! It's based on the relative capabilities of officers, particularly their movement. Even for the case of just 4 orthodox officers, we as a community have managed to get in a complete twist. Based on progress so far, our chance to do this successfully for the spectacular zoo of fairy creatures seems laughably slender. The fairy community deserve better.

We will have a table to indicate the capabilities of creatures and creature types, since most come from potentially infinite families. We can then derive various relations between them, such as Camel is a functional subset of Gnu, Mao is a decomposed Knight. People can then slice/dice this table at will. Cases where one creature is notionally outperformed by its functional inferior are paradoxical and inherently interesting to we problemists. But there is no way that this should be a total ordering for all situations. At some universal level, queen and knight are incomparable, although endgamers may nevertheless find the superiority of queen significant.

The relationship between a leaper and its many possible decompositions is provocative, and potentially hugely fruitful compositionally in all genres. Decomposition allows for fractal relationships between pieces. E.g. three knight moves can combine chirally to make a knight move on a board scaled by a factor of 5: isn't that cool?
(Read Only)pid=21286

No more posts

MatPlus.Net Forum General Promotion to knight