AW: [Retros] Re: Massacre SPGs - Stalemate SPG

[Gerd Wilts]

Francois, thank you for your interest in the PDB!


> Before asking I did my homework and searched for:

>

> K='proof game' AND K='stalemate'                  [0 PROBLEMS]

>

> Today I checked more carefully and 'stalemate' isn't a keyword and 'unique

> proof game' returns more matches than 'proof game', so these are probably

> the searches I should have done:

>

> K='unique proof game' AND K='retro stalemate'     [0 PROBLEMS]

> K='unique proof game' AND COMMENT='stalemate'     [0 PROBLEMS]


The property "stalemate" is currently not indexed in the PDB, so you can
indeed not search for PGs ending in stalemate. I will think about adding
such a search criterion. The most general way to solve this problem would
be to store each move of the solution including all interesting move
properties in the database. Then it would even be possible to search all
problems where the 5th move of Black was an ep capture, etc.


> - Is the database limited to "best problems" only?


No, Hans-Peter Reich adds as many published help mate problems as possible,
and I do the same for retro problems.


> - Wouldn't it be more useful to limit it to "publishable problems"

> instead, but then to have a "best problem" keyword or something like

> "QUALITY >= 3" for those who want a search limited to the best problems?


I am thinking about a feature which lets the user assign a judgement for
each problem. Then it would be possible to see how others liked the problem.


> - What's the difference between K='unique proof game' and K='proof game'?


I use "Unique proof game" for problems which have one (or more) completely
determined solutions (each move of the solution and the order of all moves
are completely determined). Regarding the length of the proof game only the
number of moves as given in the stipulation is considered: when a proof game
has a longer or shorter solution which is not unique, the proof is still
considered unique if the solution in the stipulated number of moves is
unique.

"proof game" is used for all other proof games.



> If every "unique proof game" is also a "proof game", then why do I get:

> K='unique proof game' AND NOT K='proof game'     [1057 PROBLEMS]


The two keywords are indeed not used consistently. When I started to assign
keywords to proof games, I assigned both keywords "proof game" and "unique
proof game"
to each proof game with one or more completely determined solution.
Meanwhile
I changed my mind and I assign only the keyword "unique proof game" to such
problems.

Best,

Gerd