-- ON JAN. 11, Long Beach State trailed UC Davis 44-40 at halftime. It was a disappointing 20 minutes for the heavily favored 49ers, but they would quickly forget about the sluggish start. LBSU's Dan Jennings opened the second half by making a jumper, and when Tyler Lamb stole the ball, Mike Caffey converted a bucket. After another empty UC Davis possession, Caffey banged home a three-pointer. It was a 7-0 run, and Aggies coach Jim Les took a timeout to stop the bleeding.
It didn't work. The Beach went on to score the next 10 points, extending its streak to 17-0 before another UC Davis timeout. But all the timeouts in the world couldn't slow the 49ers. After a lengthy media break, LBSU picked up where it left off, extending its run to 30-0 before UC Davis scored again -- the longest run in a conference game this season. A four-point deficit had become a 26-point lead, and the remainder of the second half was virtual garbage time as Long Beach cruised to a 99-74 win.
So was the 49ers' can't-stop, won't-stop scoring proof of momentum? Not quite. Because, well, it's just not that simple.
ALL THINGS BEING EQUAL, A RUN DOESN'T PREDICT ANOTHER RUN
Undoubtedly, the players, coaches and especially the fans cheering on the home team at the Pyramid experienced what they felt was momentum. In the real world, though, finding evidence of momentum in a basketball game has always been a difficult task. It's not a mistake to say that basketball, in general, is a game of runs, but past work (most notably, a 2012 paper titled "Random Walk Picture of Basketball Scoring," by Alan Gabel and Sidney Redner of Boston University) has shown it's a mistake to think that a scoring run -- the definition of momentum we're using -- is useful from a predictive standpoint.
A team that goes on a 10-0 run, for example, is no more likely to enjoy success over the next few possessions than a large-scale view would suggest. In other words, if we thought two competing teams were similar in strength, the presence of a lengthy scoring run shouldn't change that expectation over the next few possessions; we should still expect the teams to play fairly evenly. I won't completely discount the existence of momentum as a predictor, but even using the fanciest research techniques, it's difficult to detect predictable momentum in scoring because of so many outside variables.
SURE, DEFENSE SHOULD HELP
There is one particular aspect in basketball that should lend itself to more scoring runs, though: It's easier for an offense to score after it has held its opponent to an empty possession, as opposed to allowing a score. Over the past five seasons of play, teams average 1.013 points in the possession after a stop compared with 0.989 after a score. This is not a difficult riddle to solve. When a team scores, it's much easier for it to set its defense, thus making it slightly more difficult for the opponent to score on the ensuing possession.
BUT MOST LEADERS ARE JUST LOOKING TO MAINTAIN THEIR LEAD
However, there are other, stronger forces in play that tend to work against the presence of scoring runs. You can trace these forces back to a general tenet of the game: The object of basketball is not to outscore the opponent by as many points as possible but just to simply outscore the opponent. Period. Thus, a run that results in a team increasing its lead might cause that team to make strategic decisions that are more conducive to preserving the lead for the remainder of the game, rather than increasing it in the short term.
An obvious example of a first-half strategy that would work against a long scoring run is a coach who, when his team is leading, benches key players in foul trouble -- or even just rests a key player regardless of his foul situation. It's also possible that the players on the court become less aggressive, choosing to conserve energy while in the lead. (It's not that players are lazy, but it's pretty natural to try less hard with a 10-point lead than a 10-point deficit.)
You can also see this in end-game situations, when teams will tend to take more time on offense to reduce opponents' possessions. This often makes the leading team's offense less efficient -- and less likely to continue a scoring run. But it's a worthy sacrifice given the benefits of limiting the opponent's ability to score points in bulk.
AND ACTUALLY, THE TRAILING TEAM IS MORE EFFICIENT ON OFFENSE
Long before the end game begins, however, there's evidence that a team in the lead sacrifices efficient offense, while the trailing team plays with more urgency. This is illustrated by looking at offensive efficiency as a function of scoring margin (see chart on previous page). College basketball teams tend to be more efficient when trailing than when leading. Likewise, their defense tends to be more porous when leading. This effect is fairly strong even when excluding the last 10 minutes of the second half or any overtime action.
Yet the differences we see in those two stats are not huge -- we're talking a few hundredths of a point per possession. It's not something you'll see with the naked eye, and you'd need to analyze hundreds of possessions to see the effect. (In this case, I'm looking at every game since the 2009-10 season, so we're talking tens of thousands of possessions.)
One notable result, though, is that this effect is intensified in "toss-up" games -- those in which the underdog had at least a 40 percent chance of winning. The difference in offensive efficiency between a team with a 10-point deficit and a 10-point advantage is about 0.10 points per possession, whereas for what I'll call merely "competitive" matchups -- in which the underdog had at least a 20 percent chance of winning -- the difference is about 0.07 points per possession.
RESULT? MOMENTUM IS ELUSIVE FOR LEADERS, AND THAT'S A GOOD THING
Something long known about basketball is that final scores are closer than they would be if teams' scoring were random from possession to possession. This effect isn't nearly as strong as it is for other sports, most notably baseball, in which the effectiveness of a team's offense isn't related to the size of its lead or deficit. But in hoops it's a big deal, and it's the main reason momentum has very little, if any, predictive value. This sounds depressing, but I suspect that even knowing this won't dampen your enthusiasm during your team's next run.
Even with this effect working against the likelihood of scoring runs, 51 percent of the 5,308 toss-up games played since the 2009-10 season have contained a run of at least 10-0 by one of the competing teams. And 15 percent of those games have featured multiple such runs. But based on what we know about trailing teams playing better than leading teams, it shouldn't be too surprising that a team trailing by double digits is 58 percent more likely to go on such a run than a team leading by double digits.
So a scoring run -- momentum, if you want to call it that -- is not all that unusual, but it's more elusive for the team in the lead. Still, the general lack of willingness for leading teams to put the gas pedal to the floor might be the optimal approach to winning. It's maintaining a lead while conserving resources. And that result has probably played a large role in the popularity of the game. In the NCAA tournament, we'll marvel at how many games aren't decided until the final minute. At least part of the reason that happens is that in close games, you'll find momentum where you find motivation: on the losing side.