In the past offseason I undertook a study that has always interested me—assessing the value of a timeout.
Timeouts are the only tangible resource a coach has at his disposal. They only have four of them (though they can only carry three into the second half should they choose not to call one in the first) and while there is a lot of conversation within the flow of the game regarding timeouts there hasn’t been a lot of analytical research, more so just announcers proclaiming “this team needs a timeout!” when an opposing team is going on a run or the home crowd is getting fired up and giving their squad some positive momentum.
I wanted to know how effectively Mike White has been using timeouts for the Gators. I had noticed, and then confirmed with the data that he is someone who doesn’t like to call many timeouts in the flow of the game and doesn’t seem to be a believer in using timeouts to stop runs or change momentum, he wants to save his timeouts for the end of the game where he can set his defense or draw up an offensive play call.
To evaluate Florida’s timeouts I looked at the clock when each one was used and divided them into two categories. The first category was momentum-shifting timeouts, or timeouts used in the first 37 minutes of a game. Timeouts called during the bulk of game action are almost always used when a team isn’t playing well and a coach wants to reiterate something in the huddle or change a game plan so I wanted those to be different than the next category I used, crunch time timeouts.
Crunch time timeouts were the timeouts used in the final three minutes of a game. Everyone knows these timeouts because they’re the stoppages that make the final 2:15 of a game take 20 minutes as coaches flex their muscles and manipulate the X’s and O’s.
I evaluated momentum timeouts and crunch time timeouts differently to see their value. For momentum timeouts, I looked at the previous three minutes before each timeout and noted who was on a run and how much that run was. For example, if Florida’s opponent was on a 7-2 run in the three minutes leading up to the timeout I would mark it as a -5 run. Then, I looked at the three minutes following the timeout to see what run occurred next, i.e. if the Gators went on an 8-2 run it would be marked as a +6 run.
For crunch time timeouts I looked at the possession immediately following the timeout. If White called a possession after the Gators scored to set the defense, I’d then take note of the outcome of the following possession. If they called a timeout to draw up an offensive play, I’d record how that offensive possession went.
When it came to last year’s team you can read the statistical findings here. But, allow me to summarize here with a direct quote from my article:
“… the Gators were successful or had positive outcomes on 87% of their timeouts in the first 37 minutes of a basketball game. That is a staggering number that shows just how absurdly valuable timeouts are and how they can change the momentum of a game.”
87% of the time Florida used a momentum changing timeout it was effective, which is a spectacular number when you think about it. But what about those timeouts at the end of the game? Were they equally effective?
“The Gators were at 0.83 points per possession on plays coming after timeouts and that was actually worse than their average 0.869 PPP on half court offensive possessions.”
“They had 9 positive outcomes, 11 negative outcomes, and 16 equal outcomes.”
So really, the effects of the crunch time timeouts were negligible.
That was last season’s numbers and now that we’re deep into the 2019-20 season it was time to do the same study with the statistics so far. I used the same methods to evaluate the timeouts, and I was really interested to see if the numbers looked similar to last year’s or if the data sample that showed using timeouts to change momentum was much more valuable than at the end of games was just a fluke. Here is what I found.
Mike White’s Timeout Philosophy
As I mentioned earlier Coach White spent the 2018-19 season largely hoarding his timeouts for the end of games. Now that we’re in 2019-20, he’s largely doing the same. We’ve seen a few earlier timeouts in the second half but he is yet to use multiple timeouts in the first half, despite some pretty devastating scoring droughts. For example, Baylor went on a 13-2 run in the first half to get a lead the Gators ultimately couldn’t overcome without White calling a timeout, something we saw as well as when White didn’t call a first half timeout during a 14-2 Alabama run or against Utah State where a nearly 9-minute stretch without a field goal didn’t warrant a timeout call from him. For White it’s all about keeping his timeouts for the end of the game, something that the data showed didn’t work well last year, but this is a fresh season and a fresh sample size so let’s see what we get.
Florida has called 51 timeouts this season and 31 of them fall into the category of momentum timeouts in the first 37 minutes of a game. In all but 4 of these timeouts the Gators called one while on a negative run, which makes sense since it’s negative runs that make you want to call a timeout to change up momentum.
The best outcome you could ask for out of a momentum-changing timeout is entering the timeout on a negative run and completely changing momentum to a positive run. The Gators were able to accomplish this on 16 of their 31 timeout calls, or 52% of the time.
When you think of it, that’s a tremendous number. Remember most of the time these timeouts were called the Gators were bleeding out, either unable to get a stop or going through an offensive dry spell. It would be one thing if Florida’s timeout simply stopped the other team from continuing to dominate, but just over half the time the Gators went from being on the wrong side of the run to going on a run of their own.
I also looked at improvements from a timeout. Let’s say the Gators were on negative run of -6, they called a timeout, and then in the subsequent three minutes the teams tied and it was 0. That wasn’t an exact momentum change like we discussed in the previous paragraph but it still totally stopped a run and that makes the timeout call extremely effective there too.
A post timeout improvement or a stop of opponent momentum happened on 10 of their calls. These aren’t as incredible as totally stopping an opponent’s run and shifting it to a run of their own, but effectively stopping an opponent’s run is a great use of a timeout.
So, that was 10 timeouts that stopped an opponent’s run and 16 timeouts that incredibly shifted the momentum from Florida’s opponents to the Gators’ favor. That’s 26 positive outcomes on 31 momentum timeout calls.
That’s positive outcomes on 84% of momentum timeout calls.
Not only is that an extremely positive number but it’s quite close to what last year’s data stated where Florida had positive outcomes on 87% of their momentum timeout calls. Considering this data looks similar to the sample size from last year it’s looking like these numbers aren’t a fluke, it really seems like most of the time when you call a timeout in the first 37 minutes of a game to try to swing momentum, it works.
61% of Florida’s timeouts have been used as momentum timeouts but that leaves a big 39% chunk of timeouts used in crunch time so let’s see how effective they were.
Crunch Time Timeout Usage
Since I was looking at possessions immediately following timeouts and seeing how Florida’s offense or defense faired following a timeout the best way to gauge effective is by looking at the points per possession numbers.
When Florida called a timeout to run an offensive play in crunch time they were at 0.8 points per possession.
When they called a timeout to set their defense they allowed 0.77 points per possession.
For some context, the Gators are at 0.901 points per possession in the half court offensively this season. That means that coming out of timeouts they are actually scoring worse than their average. In all fairness you have look at some other factors here. At the end of games, players put in much more effort defensively, make no mistake about it. Additionally, refs are less likely to call fouls and are likely to let players be physical which makes it tougher to score as well.
Even with that being the case… 0.8 points per possession isn’t great coming out of timeouts. When I did this same study about last year’s season they were at 0.83 points per possession so you can see they haven’t been great heading out of timeouts.
Defensively the Gators have been allowing 0.827 points per possession in the half court so when they’ve set their defense with a timeout they have seen an increase in their effective allowing 0.77.
What To Do With These Numbers
Timeouts are a finite resource in the game of basketball and you want to use them as effectively as possible so let’s look at the best bang for your buck.
When Florida used timeouts at the end of the game the offensive impact seemed negligible. They have only been at 0.8 points per possession and while you can’t know what they would be at if they ran those same possessions without a timeout the fact they are at 0.9 points per possession generally in the half court you have to wonder if calling a timeout to draw up a play is really that effective. The flip side of calling a timeout to set up a play means the other team also has a timeout to put proper personnel on the floor and discuss what they want to do, and with modern video scouting it’s entirely likely the opponent knows what’s going to be run so they can scheme accordingly.
Florida’s crunch time defense coming out of a timeout has been better than their regular half court possession but you’ve got to wonder if they’d have similar results without the timeouts anyways given the fact that they’d be selling out and giving their all at the end of the game. An improvement of 0.057 PPP defensively is really solid, don’t get me wrong, but you’ve got to wonder how much the timeout has to do with that.
On the other hand, timeouts used to change momentum have shown to be 84% effective for the Gators and that seems to be a much better usage of timeouts. I’m not saying that Mike White shouldn’t save any timeouts for the end of the game, but carrying 2 or 3 timeouts into the final five minutes of the game like he has done so many times simply doesn’t appear to be the best option given the data we’ve seen.
There have been too many bad stretches for the Gators this year where timeouts haven’t been used to stop momentum, whether it’s been Florida’s struggles to string together stops or the ice cold scoring streaks they have gone on. Knowing that if a timeout is used there is an 84% chance of a positive outcome I think White should be much more liberal with his use of early timeouts.
Once again, let’s look at the Baylor game. White held onto timeouts as Baylor stretched out a double-digit lead, presumably in case the game got close so he’d have timeouts available at the end.
Well, timeouts don’t do you much when there is a couple minutes left and you’re down 14, so using the timeouts earlier to prevent the initial lead they had in the first place would have been a better strategy.
A lot of basketball thinking and coaching philosophy is rooted in tradition and most coaches have been taught to hold on to timeouts until the end of the game but at least in Florida’s case the data doesn’t suggest that’s a wise move. Five-minute stretches of poor basketball have killed the Gators this year and using timeouts earlier and often could be a great way to keep these droughts from happening.