by Andrew Mooney
The post-Home Run Derby slump in baseball is well-documented; as the theory goes, a slugger’s power numbers decline on average after participating, whether due to bad swing habits picked up during the event or just an inability to sustain the momentum that got him into the derby in the first place. As the NBA season heads into All-Star weekend, and players participate in equally gimmicky competitions, should they be worried about their performance the rest of the season? Could the same sort of slump exist in basketball?
Though the dunk contest might be more interesting to examine, no readily accessible data exists that splits dunk totals before and after the All-Star break. That leaves us with the 3-point contest; might sharpshooters participating in the event suffer a similar regression following the All-Star break?
The data suggest that their 3-point shooting percentage does indeed drop. I analyzed the shooting statistics of 3-point contestants from the last five seasons before and after All-Star weekend. My results are shown below.
This representation is somewhat misleading, given that, in a couple cases, second half shooting performance was determined by a very small sample size. For example, in the second half of 2007, Damon Jones had only 13 shot attempts, making four of them.
All together, these 30 players made 3056 out 6964 3-point attempts prior to the All-Star break, good for 43.9 percent. In the second part of the year, the players made 1302 out of 3256 attempts, equal to 40.0 percent, or a drop of 3.9 percent. Evaluated this way, the decrease is highly significant, with a p-value of 3.45 x 10^(-6). However for each individual player this is not the case. If the average player had the same percentage drop (divide all results by 30) the result is not statistically significant (p-value: 0.226).
Below is the dataset with which I worked. Who will be the next to fall to the 3-point curse? Not the mighty Ryan Anderson!
Sorry if these are dumb questions–
Is it possible that players who entered the 3-pt contest were selected due to shooting unexpectedly high (compared to their career average) pre-contest? If that is true, and if it’s likely they will shoot their career average in any given period (so, post all star break), is a decline expected (Higher than average pre-break vs average post break)?
Another question I have–is league-wide 3-pt % constant over the course of a season? If there is a league wide decline (tired legs?) or increase (more practice?) that may also put this data in context.
What Matt said. I’m surprised the original post made no mention of regression to the mean..
Yeah, this is what I was thinking. Instead of looking at their second halves compared to their first halves, look at their second halves compared to their careers.
If you look at the top 10 in any stat category in any sport from one season, and look at that same group of players again the next season, as a group they tend to regress as much as 15 or 20%.
http://sportsstatsanalysis.com/2011/01/13/laws-of-regression/
How do these numbers compare to other NBA players? Most of this drop can probably be explained by players being more tired thoughout the 2nd half of the season than they were in the first.
“The post-Home Run Derby slump in baseball is well-documented”
It is? Where?
http://www.hardballtimes.com/main/fantasy/article/do-hitters-decline-after-the-home-run-derby/