Let's Talk Assists

The cornerstone of every box score line:  Points, Rebounds, Assists.  They are all treated with (more or less) equal respect.  The problem is that assists just aren't that big of a deal.


John Stockton disagrees


Hang on, before you get all upset, I didn't say they are meaningless.  Of course they matter.  It's just that they aren't as important as rebounds, turnovers, steals, or made (and missed, in a negative sense) field goals.  Looking at the formula for Wins Produced, you might notice that assists appear to be weigh only about two-thirds as much as those factors (slightly more than blocked shots and made/missed free throws).  Why is this?  Well, there are a few answers to that.  The stathead's answer is basically: "We ran a regression over dozens of seasons and thousands of games, and found that assists don't correlate to wins as much as those other factors do."  Of course, this isn't a very satisfying answer.  I mean, I could watch these for hours and never get tired of it.  I mean, would Kurt Rambis have scored ANY points in the NBA without Magic?  So what follows is an attempt to break down the why's of this.  And all in a (mostly) non-quantitive manner, so that some of you folks can stop bitching about all the stats.

Unfortunately, unlike points and rebounds, the problem is that not all assists are created equally.  Or in other words, what we are measuring in the box score column under "assists" isn't really consistent.  Some assists are John Sally making a ridiculous pass to a wide open Dennis Rodman for a sure slam, resulting in a basket that otherwise never happens.  And some assists are Dwight Howard passing out of a double team to an open wing man, who makes a ho-hum swing pass to a guard at the top of the key, who shoots a ho-hum 20-foot jumper that hits about 50% of the time.  Which is the better play?  Well, if the jumper goes in, they both count as an assist, even though the first is clearly the better play.

Proponents of Wins Produced believe in a couple of things:  that gaining (and keeping) possession of the ball matter a lot, and that scoring efficiently matters a lot.  This means rebounds are a big deal, because you gain possession of the ball (any ball up for grabs to rebound is, by definition, a loose ball) and steals and turnovers are a big deal, because you gain (lose) a possession.  Assists, on the other hand, well, they fit in here in a rather nebulous manner.  It should be rather intuitive that assists are not obvectively measured.  Let's take a look at how.

The "Home Assist" Advantage

You can find all kinds of evidence that scorekeepers tend to be overly generous when handing out assists to the home team with a little googling (some stats analysis here, but my favorite anecdote is here).  Play around a little with splits on basketball-reference.com and it's fairly easy to see.  There are a few reasons for this, but the primary one is that there isn't really any ethical barrier to a scorekeeper doing this, so he or she doesn't have a lot of incentive to be extra dilligent about eliminating any bias (or monitoring himself for a sub-conscious bias).  In other words, when a team scores a basket, it's two points regardless of whether it was assisted or not; the scorekeeper won't have any influence on the game's outcome if he gives Rajon Rondo an extra assist, even if Glen Davis took four dribbles after a pass.  Which leads me to...

The Team Would Have Scored Anyway

Sometimes an assist "creates" a basket.  But sometimes an assist is awarded in a situation where a player was going to score anyway.  An example is Michael Jordan and Scottie Pippen in any 2-on-1 situation.  MJ's driving to the bucket.  If the defender commits to him hard, he'll dish and Pippen will score.  If the defender doesn't commit hard enough, Jordan will score unassisted (this is MJ, so even if the defender commits hard, he might get posterized for an unassisted dunk and 1).  Basically, those two points were going on the board the minute the 2-on-1 play emerged, and whether an assist happens has no influence on the outcome of the 2-on-1 (or the game).  The real assist here (if there is one) belongs to whomever it was that made that 2-on-1 happen!  Which leads me to...

Assists Don't Always Capture the Real Work

I was playing pickup ball once, and held the ball on the wing.  The low post player went over to the weak side, and set a bone-crushing pick for the other low-post player, who curled around it while the picker's defender wasn't paying attention.  I bulletted a pass into the wide-open post for a lay-up.  I'm sure I'd get an assist for this, and sure, it's commendable that I was watching the play develop (as opposed to pointing my eyes to the floor and commencing my dribble as soon as I catch the ball), but come on, how hard is it to hit a guy who is wide open?  The real assist here belongs to the guy who set that pick, but he'll never get that credit.

But lest you think I am implying that assists are all meaningless, my next article will talk about why they do matter!  In the meantime, use the comments to talk about your favorite assists, or suggest other situations that you see come up where the assist metric doesn't quite capture what it should.