re: Can the Pelicans become a top defensive team?

quote:

---

For a long time we’ve used blocks as a proxy for rim protection, and that actually made sense since there is a definite relationship and they were the only data poins that could be realistically collected by score keepers. There is a strong relationship between blocks and attempting to defending the rim. Last year there was a 0.662 R2 (.813 correlation) between blocks per forty minutes and contests of opponents’ shots at the rim per forty minutes, for all players that saw at least 500 minutes of action. As shown below:

---

quote:

---

For those players significantly away from the trend line the question is whether they are very efficient at getting blocks per contest at the rim, or are blocking shots away from the rim. And even more importantly how good of a proxy blocks are, overall, for defending the rim?

SportVU, of course, also provides the percentage opponents score on each player when defending the rim. Below is a chart of blocks per 40 minutes played and opponent’s field goal percentage at them rim when the player contested a shot last year, the R2 in this case is only 0.29.

---

quote:

---

An interesting thing in analyzing this data is that with the frequency and effectiveness of rim protection included, the value of blocks is negative 1. There are a number of reasons high frequency blockers might be less valuable rim protectors, from collecting blocks away from the rim to a propensity to get out of position, being over eager and falling shot fakes. But, that is not the kind of term I am comfortable including in a model based on one year of data without a better understanding of the complete long-term relationship. In any case, we don’t have SportVU data in any league outside of the NBA and there for only one year plus at this point, so we’ll have to keep using blocks as a proxy, but with perhaps an added caveat or two

---