Act VI · Watch
Nine hundred games worth watching
Not a database. A library — filtered hard, so that opening it at random is a good idea rather than a gamble.
There are databases with ten million games in them, free, a search away. Brass Pawn does not try to be one, and the reason is that a database is a place you go with a question. A library is a place you go without one.
Nine hundred is small enough that somebody could get through it, and every one of them earned its place against three rules.
1
Decisive
A draw can be a masterpiece and almost never teaches one. Every game here ends with somebody winning, so there is always a moment where it went wrong and always a question worth asking about it.
2
Two named players
Game scores from published collections of the players’ own games. You can look somebody up, watch four of their games in a row, and start to hear an accent in how they play — which is not something a pile of anonymous games can give you.
3
Short, or famous
Either finished inside twenty-five moves or well known enough to have a name of its own. A ninety-move technical grind between two people you have never heard of is a library nobody opens twice.
The part that is not watching
The move you disagree with is the whole point.
Watching a strong game from beginning to end is pleasant and teaches less than people hope. What teaches is the moment — and there is always one — where you look at the board and think I would have taken there, and then the player you are watching does something else and wins anyway.
At that moment you can stop being a spectator. Take the position over and carry on against the engine from exactly the square where you disagreed. Play your move. Find out.
Usually you find out that your move was fine and the game was still lost, which is its own lesson: the mistake was three moves earlier and you walked past it.
Where they come from
Game scores are matters of record. Annotations are not.
The scores come from published collections of the players’ own games. A record of which moves were played in a public tournament game is a fact, and facts are not copyrightable — which is why a library like this can exist at all. What is copyrightable is somebody’s commentary on those moves, so none is included: no annotations, no borrowed analysis, no exclamation marks inherited from a book.
What the app adds instead is its own — the same coach that grades your moves will grade the ones on screen if you ask it to, from the same engine, with the same win-probability scale as everywhere else. It is not a grandmaster’s opinion of a grandmaster’s move. It is an engine’s, and it says so.
Judging the players is a different mode.
Watch is for games worth seeing. Guess the Elo is for ordinary rated games where the question is not “what happened” but “how strong were these two” — sixteen hundred of them, both players within a hundred and fifty points of each other.