[xkeoyxuh] Chess annotations

A proposal for the meaning of chess move annotations.

Question marks attempt to objectively assess the move:
??? Gives away a full point, i.e., loses in a won position.
?? Gives away half a point
? Seems a bad move but the annotator is not sure if it actually loses at least half a point.
Nothing indicates no change in evaluation.
Additionally, a decision to resign or draw may also be annotated with question marks.

With the power of computers, annotators may be graded on their ability to correctly assess a move.

The big change in philosophy comes with exclamation points.  Exclamation points (one or more) unashamedly subjectively assess how "excited" the annotator is about the move, for whatever reason (probably should be explained in comments).  Previously (Hübner's approach), this tendency was defined cynically or derisively.  However, in this age of computers, playing objectively good but boring moves is no longer that exciting.  Moves with exclamation points may win special prizes (a finer gradation than game prizes), and players who play lots of exclamation points are more likely to be invited back next year or to other tournaments.  (Players who play lots of question marks are more likely just to lose.)

An AI task is for a computer to assess exclamation points similar to a human annotator, or for a chess engine to prefer such moves.  It's interesting the human-computer duality with question marks; they switch roles of being ground "truth".

It is possible for a move to get question marks and exclamation points: objectively bad but still exciting.  Question marks are written first for this case.

The annotation !? therefore has a special meaning.  The annotator believes the move is exciting, but the player found it by dumb luck, rather than any Tal-like personality trait or desire to play an interesting move in that situation.  (Annotation is of course always a sport of arrogance.)  Such a move should not be basis for inviting a player back.

One might have the longest annotation ???!!?, a move which objectively loses, but is very exciting (sets a beautiful trap which recovers the win?), but the player probably didn't realize it.

I'm not completely happy with this system, because there's no way to distinguish between objectively bad moves and blunders, the latter being a subjective evaluation of whether such a mistake is acceptable for that player.  The problem occurs nowadays in tablebase endgames, where ground truth may be known (evaluating a move ?? or ???) but it is beyond the ability of the annotator, or subjectively, "any human", to understand it or explain it (and the game continued with the opponent unable to capitalize on the mistake), so probably should not be considered a blunder.