Can Timeouts Change the Outcome of Basketball Games?

From Serguei Saavedra, Satyam Mukherjee, James P. Bagrow.  Here is the paper, here is the abstract:

In basketball, timeouts are believed to reverse the momentum of a game. However, here we show timeouts have no significant effect on the final outcomes of games. Moreover, we find that the timeout factor only appears to reinforce the game of dominant teams, meaning that only the most successful teams can find any positive benefit. We find no association with team payrolls, suggesting that richer teams are not particularly better at capitalizing on timeouts. Our findings support that strategic breaks have little impact on workplace performance and productivity.

For the pointer I thank Michelle Dawson.


Game time coaching involves a lot of "in the moment" decision making. If an opposing team is mounting a rally, it has to be stopped and a timeout is seen as an effective way to disrupt the momentum. In a situation like that, a coach will use the timeout to respond to what is currently happening and won't really think if the timeout will or will not affect the outcome of the game later.

Does this pass for statistical analysis in their field (physics)?

They essentially create a placebo by randomly changing the time of a timeout within a given quarter. So if there were 3 timeouts in a quarter, they compare the actual change in score (from the timeout to either the next timeout or end of quarter) to the fake change in score generated by randomly changing the time of the timeout, but preserving the number of timeouts per quarter.

They don't do any robustness, discuss the pros or cons of their approach. If an undergrad did this, they'd get a B.

As a physicist, let me just say none of these authors appears to have an affiliation associated with physics. However it is sad to see they are all affiliated with Northwestern U.

The timeouts might not affect the outcome of the game but the number of free throws does. With few exceptions, the team that shoots the most free throws wins the game. Pretty exciting.

"Few exceptions"? What percentage? Like 30%?

On the basis of a statistical analysis of the last 16 NBA seasons by Wilberforce and Cobbett, if the home team shoots the most free throws, which occurs 88.4% of the time, they win 83.3% of the games.

The team that is behind in the late fourth quarter starts using intentional fouls.

It would be nice to know if this practice had any effect on the outcome of a game, and if it is found not to, we could maybe see an end to the most annoying thing in sports.

Yup, it's a classic case of correlation causality, unless they somehow were able to leave the intentional fouls out of their counts.

Our findings support that strategic breaks have little impact on workplace performance and productivity.
Is this a joke?

Yeah...this paper is investigating point-differential not "productivity" like, e.g., offensive efficiency. Also a giant reach to apply findings from professional basketball to the workplace.

Right, that generalization is absurd. Stipulating their result, There are two teams out there competing. The result says nothing about absolute performance, only that neither benefits more than the other.

And they got financial support for the paper! Economics really is a great business, isn't it.

I also wonder if their data drops the scheduled TV timeouts.

These statistical analyses of sports always seem to me to be sloppy.

In the present case: it is natural to expect a score differential to increase with time. So, if score differentials are unchanged, that might indicate that the timeout did provide an advantage, it just wasn't enough. More importantly -- and this is in the vein of how I think these studies tend to be sloppy -- teams might call timeout *in anticipation* of circumstance where they expect a rapid increase in score differential. In that case, maintaining score differential would be a huge success.

Stopping the other team's momentum is but one reason that NBA coaches call timeouts. Giving star players additional non-game time rest, making player substations, and organizing plays near the end of the game are all reasons coaches call timeouts that have nothing to do with momentum, per se.

According to analysis reported by ESPN, calling a timeout on offense to set up a play near the end of a close game is highly counterproductive. When the team with the ball calls the time out, they are much less likely to score than when they don't call a timeout. Did the authors compensate for this? Did the authors even know about this phenomenon?

There are causality questions with that result. If a team grabs a defensive rebound, it has a choice of calling a timeout or not. Well if you've got a 2-on-1 break going, you're not going to call timeout -- you're going to run down the court and score. But if the defense is already in position, you might want to call a timeout to raise your probability of scoring from say 40% to 45%. The non-timeout possessions can have a higher scoring percentage (>90% in the fastbreak situation), yet it is still correct to call a timeout when the situation calls for it.

I'd like to see the same study done with college basketball as the subject. I would think that momentum and timeouts might have a different effect for unpaid amateurs than for seasoned pros.

This study reminds me of Keynesian economics. "But for the Keynesian bailout, things would have been worse". But it's hard to tell if this claim is true--it's the Heisenberg Uncertainty Principle. In the basketball case, one could argue that had a timeout not been called, a blowout would have resulted (against the team calling the timeout) even worse than whatever point differential actually occurred post ante.

Time-outs work great with kids, though!

I believe Phil Jackson mentioned this in Sacred Hoops or one of his other books. He was known for letting his teams play through tough stretches.

John Wooden hated timeouts and rarely used them. He did alright.

Yes. they turn the game into a hockey game or the news. Two key variables are resting your star player, and coaching. Got to correct for good coaching (set plays and how to decide which coahes are best) and tired star players. I guess your best five (and Ginobili was a bench player), if they are better than worst five, moreso than other team. I like the +/- basketball stat. So teams that have biggest spread between best 5 and rest...gotta analyze that. And injuries and age.

If the TO reinforces dominant teams, frig, that's exactly the correlation you were looking for!! But you also said no effect. The Greeks wouldn't have been happy with this shallow analysis. If it reinfocres dominant team it means yes, don't take your TO if you suck. In hockey there is one TO. It is the fastest game on Earth.

Making FTs a 3pt (shoot from anywhere but give 3pts if you make it and they don't ) "horse" shot would be sweet. The correct BBall stregy is to knee or trip superior players early in a series. I think only the Pistons used it.

A Slovak just smoked Erat, a decent Czech sniper, in the head with late hit when refs turned away. Slovaks up 3-1. @#$%, I would've had Slovaks, Russia and Verlander. Who schedules a hockey game at 6am?!

How it worked when my team won the city championships, is we didn't use many subs and the crappy player parents were happy with that at the end of the year. When I was rested, the TOs didn't matter. When I was lactic-acided, I couldn't jump as fast or recover as fast, without a T.O.
The team we played was equally good. They had a player at 12 who could swerve his corner kicks right into both top corners. We played them in the regular season and I was just barely able to get to his kicks. Because they didn't score, he had concave kicks instead, the championship game. Phew. Them and the dirty team we drew, were the only ones who scared me. One playoff game my knees were so bloody I wanted to pass out. But in the zone last game. Their star got an early breakaway and I smothered his soft low shot to the corner. I didn't even know I had it I was too in the zone. I couldn't feel the soccer ball right under me.
I played them years later in the B division and my crappier team was one of only two teams to get a goal on them. In high school the grade 12 friends of the team goalie got most of the games. I got the best team when I went to watch a game with German Measles. Instead I played; used looseleaf in shins to fool ref. We had great talent but no teamplay. My teammates were booting me in the shins before the game started and I held them to 10 goals. We never practised 3 on 0's...Next year the guy behind me was Provincial team gold medal winner. @#$%.

What you could do is look at team performance after both team's time outs have been exhausted.
...we had a player who looked suspiciously like the star of a knocked out team. At least our ringer wasn't 18 and able to hoof it in from centre. That guy was not on their team earlier in the year. Some say our GWG was offside...I guess we'll never know. Slide tackled that kid years later incessantly. Publicly subsidized sports build leadership.

Yeah, the paper actually says the hardest workers and the laziest workers benefit. Again, without timeouts, both teams would suck. Athletes know this. Kobe with his 43 minutes knows this. Translated to the workplace, that would mean all workers would suck, which isn't as good as all workers being rested. Get the athletes in the Universities to write these, plz.

Comments for this post are closed