Any chess mavens out there that can give me another pair of eyes on this viz? I was interested in when the best chess players tend to castle and to which side. So I'm wondering if this viz is clear and accurate (I don't play much chess but am curious about this move!). Thanks!
@Allitorban Very interesting. I don’t play much either but I truly enjoyed the documentary about Carlson on Netflix. Your viz is very elegant, the conclusion at the top is good and I like that you explain what castling is. Maybe one thing I would like to know is whether the player who castled won the game or not (not just Carlsen).
@Allitorban this leaves me with the question, "how common is castling?"
I can see that castling, when it occurs, is more common in the early game (a trend that is surely influenced by the requirements for castling), but I don't know if it is actually a common move as this could be only a small fraction of total games. You could consider changing your dependent variable from count to proportion, e.g. out of all "Move #10" in the tourney, what percent were castling?
@Allitorban this is really neat.
Question: how are the participants ordered within a row? I can see, for example, that Carlson castled king side 3 times on his 8th move, but they’re spread out and I don’t know why. If you consolidated them at either the beginning of the line or the end, it’d be easier to compare when he castled most frequently.
@Allitorban something else I think might be interesting is whether the player was playing black or white, since white usually has an advantage.
It could also be interesting to see when castling occurred by only one or both players in a game. Since you’re showing players, I assume it’s possible (perhaps likely) that a pair of those dots occurred in the same game. This might, however, over-complicate what is a wonderfully simple visualization.
All that said, I love the viz & thought it clear 😁
@Allitorban Looks awesome with a clean story! Not necessarily a problem, but the fact that 'time' is plotted bottom-to-top took me a moment to parse, as it runs against convention. Also "move #" might be confusing since "number of moves" is what the chart is about (distinguish moves from turns?), but that's not what's on the y-axis.