[This post will be updated as more 2021 season data becomes available].
I use the shot zones defined by the WNBA, as shown on a league graph presented a third into the 2019 season (see here for article, and below for the shot zones we are using):
The WNBA tracks shot locations, and let us consult them but unfortunately does not provide the raw data (i.e. the exact coordinate for each shot) that would allow us to reconstruct and customize such maps. I use data scrapped from ESPN, thanks to Saiem Gilani and Geoff Hutchinson's WeHoop R package. ESPN provides shot location for each shot, and I am able to assign each shot to a zone based on distance to the basket and location on the court.
The first thing I look at are evolutions across time. We know about the 3 points revolution in the NBA, and we already know it's reached the WNBA. In 2002, the WNBA was attempting 20.4FGA a game. In 2020, this number was 30.9. The other type of shots on the rise in recent years are shots in the paint (non restricted area). Two types of shots are being sacrificed and have decreased: shots in the restricted area and mid-range shots. 3FGA have replaced long midrange FGA, in line with the 3 points revolution across basketball. More surprising, shots in the paint looked like they are converging toward shots from the restricted area. Interestingly, the percentage of 3FGA from corner shots has not increased over the years, and if anything else has decreased. Given that the shot percentages in corner 3s are at least as good, if not better, than those for behind the break 3s and other non-restricted area shots, this feels like an unexplored area for the WNBA. This is all consistent with Seth Partnow's findings for The Athletic.
Building on these findings, I look at two things: (i) whether and how much 2020 and the 'Wubble' season was an outlier as opposed to a confirmation or acceleration of recent trends and (ii) how we can explain the increase shots in the paint, seemingly at the expense of restricted area shots.
To try and answer the first question, I add data from the 2021 season (updated on June 4th). It is a small sample, with only 44 games played out of the 192 scheduled to be played. I will update the data throughout the season, but the early data seems to indicate that 2020 might have been an outlier. The dramatic decrease in shots in the restricted area and increase in shots in the paint in 2020 appears to have been an outlier. The slow trend might still hold (with shots in the paint converging toward shots in the restricted area), but is not as steep as 2020 data could suggest and will need to be confirmed with future data. I do not have a convincing explanation for why the restricted and paint areas numbers were so off in the Wubble. They could be a mere statistical outlier, but how far the data points are from the lines, in addition to the symmetry between shots in the paint and shots in the restricted area, makes me think something more systematic was at play. The zone determination is based on distance to the hoop, not any manual coding, and therefore the Wubble numbers cannot be attributed to the idiosyncrasies of specific human coders. I do not know how ESPN measures shot distance, maybe something was mechanically wrong with their measurement tool in the Wubble. If measurement was not a problem and actual shot patterns in the Wubble were indeed different, the specific Wubble conditions (no travel, no fans) could hardly explain these findings.
Comments