A Game of Runs – Measuring “Spurtability”

facebooktwitterreddit

May 27, 2015; Oakland, CA, USA; Golden State Warriors guard Shaun Livingston (34) and forward Draymond Green (23) and center Festus Ezeli (31) celebrate during the fourth quarter against the Houston Rockets in game five of the Western Conference Finals of the NBA Playoffs at Oracle Arena. Mandatory Credit: Kelley L Cox-USA TODAY Sports

It is often said basketball is a game of runs. Over the course of 48 minutes, there is an ebb and flow as either through adjustments or sheer chance teams string together a few buckets and consecutive stops.

The concept of “momentum” in sports is a hard one to capture at least in quantifiable terms. As of yet, it’s not clear if there is any real value to simply avoiding empty possessions on offense. The presence or absence of such an effect certainly has implications for the viability of a high-variance strategy relying heavily on threes. If multiple consecutive empty trips are a problem, that might tilt the calculus slightly more towards the surer thing two-pointer over the slightly higher expectation long three in some situations.[1. Of course, not even this rescues the contested midrange shot, so beloved of Basketball Fundamentalists.]

An extended run of points might be an example of this sort of momentum.  There is something inimitable about a team going on a run. As the points pile up, the opposition coach is forced to disgustedly burn a timeout, the crowd is either in a frenzy of exaltation or suffering anxiety about the home team’s plight.  The most feared teams in the NBA are the ones who can throw up a quick burst at any moment. No lead is safe, and a small gap can turn into an insurmountable chasm seemingly in an instant against a squad with this sort of “spurtability”. Of course, overall scoring margin accounts for much of this effect. Scoring in bunches and avoiding being scored on for lengthy periods sort of naturally leads to more tallies in the “for” than the “against” column.

Is there value in these runs of points over and above what might be expected? It’s a good question, and to take a first pass, I looked at play-by-play data from every regular season game this season.[2. Thanks as always to NC Data Guru Darryl Blackport.] Completely arbitrarily[3. And based on a question by Friend of the Program Jared Dubin.], I used runs of 7-0 or greater as a starting point[4. For easy of data analysis, I only looked at consecutive point scoring runs. Two 5-0 runs sandwiched around an opponent free throw probably has a similar effect or greater effect, but runs like that are harder to pull out of the data en masse, and I’m all about easy here. Longer runs only counted once, no matter how long, though the number of such runs decreased surprisingly rapidly as the minimum number of points increases.]. There was no particular analytic reason for picking this amount, other than one has to start somewhere, and a 7-0 or more run means scores on three straight, unanswered possessions. The initial results were startling.

NBA teams averaged 171.5 runs apiece of seven or more consecutive points over the course of the season. Teams averaged 2.6 of them in wins against 1.6 in losses. The team which had more of these runs won a staggering 82.2% of all games. Road teams in particularly struggled in games where they gave up more big runs than they scored, where in only 14.8% of the 553 visiting teams which were “outran” ended up winning the game.[6. Home teams won 90 of 413 games in which they gave up more 7+ point runs, a slightly more respectable 21.8%. Still not great, Bob.]

Unsurprisingly, various teams’ ability to go on and avoid runs tracked pretty strongly with overall winning percentage:

Also not a shock considering their two-way dominance and high pace, but Golden State crushes this any way you slice it. Most 7+ and 10+ runs scored[7. As the chart shows, even adding another three points to the “minimum” for a run greatly reduces the frequency, with only slight over 1/5th of all 7 point runs being extended out to 10 or beyond.], 4th fewest runs allowed, best by percentage (especially at home), and so on:

Of course, the above could just be neat-o[9. Sponsored by…nobody!] and not mean anything. But it certainly seems worth further investigation to see if this spurtability is distinguishable for regular old scoring margin, and if that difference (should it exist) has any additional explanatory power over team performance, either in the regular season or in terms of predicting playoff success.