Here is a new thread to discuss statistics on players. I've been downloading data automatically from the bwf website: I wish there was more data recorded, but there's already enough to obtain interesting tidbits. See Kento Momota (桃田賢斗) and Kento Momota (桃田賢斗) to see what kind of data can be extracted. For each year from 2010 to 2019, I looked at the top 10 players for the final weekly ranking of the year, and stored their statistics for the year: in the following scatter plots, each point represents one of these top 10 players, so for each year, you can see ten points. From the data, one can derive trends which must be folkore for professionals. The first one is: The game is much more physical. Matches, games and points last longer: here are their average duration. In 2010, the average game duration of any top-10 player was less than 21 minutes, and less than 19 minutes for half of them. In 2019, none of these averages is less than 19min30s, and half is more than 21 minutes. In 2010, the average match duration of any top-10 player was less than 48 minutes. In 2019, it was always higher than 46 minutes, with half being ≥ 50 minutes.

It is interesting to see the stats I would like to see the time broken down to show total time spent in a rally if the stats are available. I assume that mins/pt is showing the average match time divided by the number of points? I was under the impression there is more time wasted/resting between rallies and the actual rally length is still approximately the same - but certainly more physical and greater speed.

No, this info is not available: the number of strokes is completely unknown. I computed the total number of minutes (on the bwf website, the duration of each match is given) and divided it by the total number of points. With a bit of extra work, it is possible to extract the sequence of points: tournamentsoftware says who scored at each point, from 0-0 to the end of each game, but I don't know what to do with it. Anyone interested with that?

i love those graphs. if i didn't watch the match i use it to see when the turning point of a game was, if there was one. did someone lose a big lead? who likely had control of the game, or lost control. it can be used to count, and calculate the % of, game points/match points converted. it would let us know who can close out a close game with a high degree of reliability and who tends to choke.

Here are two tables, giving the stats for the top-10 ranked players in respectively 2010 and 2019 (including the latest China Open).

How would you define "clutchness"? To extract data, we need robotic non-ambiguous definitions. For example, for each game won, one could count the number of game points needed. And for each game lost, one could count how many game points were saved. But maybe there are better data. One could also compute %Point to restricted periods, like the end of a game, but how do we define precisely the "money time"?

that's a good word... or clutch-rate. game points won vs. saved/lost hmmm... i think the graph tells you everything you need to know for the stat. clutch-rate/choke-rate: 90%+ conversion rate makes you a clutch player, less then 70% conversion and you're a choker. crunch-time: occurs when a player reaches 18 w/ a lead of 2 points or less.