Computers and Chess Annotation

Most chess programs also offer an "analysis" facility which given a game compares the moves played with the best move possible, displaying the corresponding scores and move continuations. This is useful to dedicated players but isn't like the annotated games seen in books and newspapers. Annotation styles vary greatly, ranging from the textless analysis of international journals like Informator, to the wise-cracking style seen in some Sunday papers. We are hoping to produce automated annotations that in style are between these 2 extremes.

What to comment on

• Good and Bad moves - Whatever the annotation style, good and bad moves (denoted by ! and ?) are always commented upon. Bad moves are easily identified - they're moves which are significantly worse than the best move. However, good moves aren't so easily recognised - the best move (even if it's far better than the next best move) doesn't always deserve a !. For instance, recapturing a Q is usually best because it's forced rather than because it's clever. To merit a ! a move needs to be in some way unobvious.

  Turning points (especially the last, fatal mistake) also feature in annotations. By keeping a note of position scores one can comment on fluctuating fortunes.

• Phase Changes - When the game reaches a new phase (the endgame, for example) there's often a review of the position.
• Traps - One also sees comments of the form Not Nf7? because of Qh4!. At any stage in a game there can be many moves than will lose straight away, so why do annotators pick on certain moves? Again, the notion of "obviousness" is useful.
• Threats - Another common type of comment is of the form Threatening the fork Nf7. A threat is a move that quickly leads to a significant advantage.
• Explanations - Players like to know why certain moves are played. "Fritz" (a commercial program) has an "explain all moves" option which might help beginners. It's rather difficult to produce explanations that satisfy better players though.
• Feature description - Sometimes the annotator takes a look at more general features - White's Queen-side pawn majority might be useful in the endgame, White's dynamic potential compensates for his pawn weaknesses, It's a double-edged position, etc. These features may not be assessed in the evaluation routine and even if they are, their impact on the overall score is unknown.

Using existing evaluation functions

• width - It's useful to look at the best 8 (say) moves and how good they are. This will help characterise the position. For example
  • there may only be one good move
  • there may be several moves that are equally good
  De Groot suggested that on average there were 2 good moves per position on average (often 1, seldom more than 5) but initially players (especially bad players) might consider far more moves.
• depth - It's useful to see how the goodness of a move varies according to the depth of analysis. A move that shallowly is good could be considered "obvious" but if deeper analysis shows it to be bad then it's a tempting mistake. A move that shallowly is bad but turns out good on deeper analysis might be worth a !. So if the shallow and full-depth goodness of the best few moves were known for each position then already some interesting conclusions could be drawn

  For example, suppose a histogram of a position's moves came out like this

            Goodness      Obviousness
            xxxxxxxxxx    xxxxxxxxxx
            xx            x
            x             x
            x             x
            x             x
            x             x
  

  (i.e. there's only one good move, but it's an obvious move). This is an easy position! Alternatively, consider this situation

            Goodness      Obviousness
            xxxxxxxxxx    x
            xxxx          x
            xxxx          x
            xxxx          x
            x             xxxxxxxxx
            x             xxxxxxxxx
  

  Here there's one good, 3 average moves and 2 bad moves. Unfortunately, the bad moves are the most obvious ones to make. This is a difficult, tricky position.

Making annotations readable

Experts and beginners might prefer a different balance between comments and hard analysis. There may also be a need to treat games between beginners differently to games between grand-masters.

Whole-game annotating

Whole-game annotating may also make possible the retrospective identification of plans and motives.

New evaluation procedures

• Use of an alternative evaluation function that returns an array of factor-specific values
• Pins, forks and skewers could be identified.
• The ability to create a list of positional strengths and weaknesses would be useful.
• Sometimes pattern-matching is a good way to identify "obvious" moves. For example, if there are black pawns on c5 and d6, then black plays e5, white might instinctively play Nd5.
• Unclear or sharp positions are important to identify. Some players will steer towards them, others will accept a slightly inferior position rather than have to face them. A routine to map out the 2D space of the move-tree in terms of "turbulence" might be useful.
• Investigate "chessmaps" (July 2000 issue of Artificial Intelligence, p.235-250).

Playing

Implementation and Examples

Definitions

Moves

• [Objectively] best move - the move which (as far as can be determined) leads to the best possible position
• Good move - a move which (as far as can be determined) is close to being the objectively best move
• Obvious move - a move which using an n-ply analysis (where n is small) is good
• Easy move - a move which using an n-ply analysis (where n is small) is close to the best move using an n-ply analysis
• Risky move - a move which leads to a much more unclear position
• Double-edged move - leading to a complicated position
• Quiet move - a move which doesn't lead to a more unclear position and doesn't have an immediate threat.
• ! - a move which is
  • much better than the alternatives
  • not obvious
• !! - a move which is
  • much better than the alternatives
  • not at all obvious
• ? - a move which is much worse than at least one alternative
• ?? - a blunder: a move which is much worse than at least one alternative, converting a win into a draw/loss or a draw into a loss
• !? - not objectively best, but good for pragmatic reasons.
• ?! - far from objectively best, but maybe good for pragmatic reasons.
• Forced move - a move (usually obvious) that is necessary in order to maintain the status quo.
• Trap - a move that has an obvious reply which is a bad move.
• Sacrifice - a move that immediately loses material, but for a sufficiently large n leads to an improved position according to an n-ply analysis
• Swindle - same as a trap?
• Passive move - ???
• Aggressive move - ???
• Provocative move - ???
• Pragmatic move - A move which isn't objectively best but may lead to a better result than alternatives do. For example, if the player prefers quiet positions and the opponent doesn't, then a quiet move may be a good idea.
• Prophylactic move - ???
• Accurate move - ???
• Threat - a move which will win quickly

Positions

• Unclear position - a position where few obvious moves are good.
• Drawish position - a position where all the possible paths consisting of obvious moves lead to a draw.
• Wild position - ???
• Solid position - ???
• Tactical position - ???
• Strategic position - a position where there are several similarly-good moves
• Complicated position - a position where some of the obvious moves are good and some bad.
• Sharp position - a position with traps and few good moves.

Terms

• Pressure - ???
• Initiative - ???
• Counter Attack - ???
• Gaining/losing a tempo - ???