For A Few Defensive Contests More – Rim Protection Redux

January 5, 2015; Oakland, CA, USA; Oklahoma City Thunder forward Serge Ibaka (9) shoots the basketball against Golden State Warriors forward Draymond Green (23) during the second quarter at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports
January 5, 2015; Oakland, CA, USA; Oklahoma City Thunder forward Serge Ibaka (9) shoots the basketball against Golden State Warriors forward Draymond Green (23) during the second quarter at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports /
facebooktwitterreddit
January 5, 2015; Oakland, CA, USA; Oklahoma City Thunder forward Serge Ibaka (9) shoots the basketball against Golden State Warriors forward Draymond Green (23) during the second quarter at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports
January 5, 2015; Oakland, CA, USA; Oklahoma City Thunder forward Serge Ibaka (9) shoots the basketball against Golden State Warriors forward Draymond Green (23) during the second quarter at Oracle Arena. Mandatory Credit: Kyle Terada-USA TODAY Sports /

After Wednesday’s Deep Dive (and also the full league stats update here) into revised Rim Protection methodology, I got some feedback indicating the switch from scaled (i.e. as compared to positional average) to raw points saves was a bit confusing. Rudy Gobert saves an estimated 9 points per game, is that good? Obviously, since it’s Gobert and protecting the rim is kinda what he do, we can assume it probably is. But still, we’re not quite sure how good. Which brings us back to the need for an average, a baseline level of comparison.

As I mentioned last time, the problem is average for what? Enes Kanter is an above average rim protector relative to all 450+ players in the NBA, but one of the worst, if not the absolute worst among big men. At the same time, players like Danny Green or KJ McDaniels can provide an extra bit of value with their unusual ability to contest shots effectively at the rim even while playing primarily on the outside even though the same level of rim protection would be pitiful from a player with more direct responsibility in that area. In today’s NBA, traditional positions are often considered passe, but for the purpose of defensive role they can serve an important categorization purpose:

There are two (or more) things going on there. First bigs tend to be, well, bigger than other players. Extra height, wingspan, size and so on all naturally help bother shots at the rim because geometry. To put it more directly, big men are more skilled rim defenders than small guys. The other aspect apart from sheer size is proximity to the rim, which naturally puts players in better position to contest shots at the rim, as the following illustration of Contest %[1. Shots contested at the rim / opponents’ shots at the rim while player is on the floor, per SportVU data.] by position:

It doesn’t make much sense to normalize the size/skill element by position. Five centers would be better at defending the rim than five point guards because it’s harder to shoot over five 6’10+ guys than five 6’5 and under, and to measure the rim protection of those lineups, we wouldn’t want to obscure those real skill differences. However, the proximity to the rim is worth controlling for. As I discussed earlier this summer[1.  And to briefly update that post, here are PF/C comparisons for a number of players who frequently swung between the two big positions last year:

PF C comp specific bigs
PF C comp specific bigs /

Again, Greg Monroe might be fine in Milwaukee, and I remain concerned about LaMarcus Aldridge defensively in San Antonio.] with respect to players who swing between the 4 and the 5, their responsibilities on defense can shift greatly depending on who they are playing with. Which in turn affects who they are guarding.

For example image a contest between Detroit and Golden State last year. Assuming the Warriors were playing their usual starting tandem of Andrew Bogut at center and Draymond Green at power forward, Greg Monroe’s defensive assignment would be very different if he was playing next to Andre Drummond, or next to[2. Pre-catastrophic personnel move, I’m never letting that one go…] Josh Smith. If Monroe is matched up with Bogut he is naturally going to be more likely around the rim to bother shooters than if he’s chasing Green all over the perimeter.

In other words, the amount of rim protection “work” Monroe or any other player should be expected to do depends on the interaction between the two lineups. That becomes unwieldy rather quickly, however, so I’m going to instead make an assumption that these responsibilities largely depend on just the lineup of the player’s own team. Generally speaking, for any given lineup the “center” has one mix of defensive responsibilities, the power forward another, slightly less interior-focused set and so on. Those roles will of course differ across teams, but for the purposes of a league-wide analysis, I’m going to go with the averages for each position as discussed in Wednesday’s article:

Using a weighted average of Basketball-Reference.com’s positional playing time estimates[7. As we start to edge in to more “on/off” analysis like Mika’s from the other day, a better set of position estimates might become available, but that’s a whole other big project and B-Ref will do for now.]  to create an “expected” points saved number, I can then compare to a player’s actual number, and voila:

That certainly matches up with the intuitive understanding of who is good or not at defending the basket. I particularly like how it demonstrates the increased importance of rim protection as you move “up” the positional latter. The difference between the best and worst among point guards is pretty “so what?” The difference between Gobert and Kanter though?

It’s not perfect, especially because there are some players BRef’s algorhythm’s miscatergorize (for example, Draymond Green tended to get assigned SF minutes while Harrison Barnes got saddled with PF minutes which combined to make Green look like a much stronger rim protector in this analysis than he really was. He certainly didn’t embarrass himself, but being credited with 52% of his minute at SF[1. And the attendant 3 points/36 lower expectation in terms of rim protection.] skews the analysis. Simply swapping his playing time estimates of those with Barnes would lead to Green rating at about +0.1/36 in terms of point saved over expectation. For a guy “playing up” the whole time who brings Dray’s playmaking small forward skills to the offensive side of the table is a huge net win for the Dubs. But Serge Ibaka he isn’t.

Further, as long as this is a known issue[6. For example, the next time a small ball four like Green jumps off the page like this, a quick look at the position estimates can show the discrepancy between the estimate going into the data and how the player is actually being used on the floor.], this comparison to an expected baseline is probably a more intuitive way of understanding the value a player is adding or subtracting via rim protection while on the floor.