Hey, NFL Coaches: Stop Wasting Your Timeouts

Early in the second quarter of the Los Angeles Rams’ wild-card playoff game against the Arizona Cardinals, Rams head coach Sean McVay faced a third-and-2 at the Arizona 33-yard line with a 7-0 lead. Running back Cam Akers had just gained 3 yards rushing up the middle on the previous play, and he substituted out of the game as time ticked off the play clock. With 16 seconds left, quarterback Matthew Stafford took a step toward the huddle but abruptly stopped, throwing his arms up slightly in exasperation. A close-up shot of McVay reading from his play sheet revealed the reason why: Stafford still hadn’t gotten the play call from his coach.

The Rams finally broke the huddle with eight seconds on the play clock. Stafford walked to the line, quickly looked over the defense, glanced up at the clock and shook his head, seemingly realizing he didn’t have time to execute the pre-snap motion the play called for. He called time and stalked to the sideline, leaving the Rams with just one timeout left in the half.

As it turns out, burning through timeouts unnecessarily is something of a McVay calling card. According to the NFL’s Michael Lopez, no coach wasted more timeouts on a per-game basis from 2018 through Week 9 of 2021. Those “operational” timeouts — timeouts due to play clock mismanagement, the wrong personnel on the field, failed challenges, or simply the look the defense is giving the offense — are largely avoidable and can be quite costly. After all, timeouts come in handy at the end of each half and help a team keep drives alive late in the game. 

This led us to wonder: Is McVay’s play-calling good enough to overcome the lost value of “wasted” timeouts? And not just McVay — is any NFL coach good enough to spend timeouts frivolously and still come out ahead?

To find out, we used the nflfastR package and looked at every play teams ran during the 2021 season (including the wild-card round) on offense after calling an operational timeout. For this analysis, we defined operational timeouts as low-leverage timeouts taken with four or more minutes left in the half — in other words, timeouts taken when the game probably isn’t on the line. After identifying those plays, we summed them up by team and calculated the expected points added (EPA) per play and the win probability added (WPA) per play. Finally, we calculated the total points added for each metric as well.

AFC North rivals Baltimore and Cleveland battled for the fewest operational timeouts in the NFL this season with just three for Baltimore and four for Cleveland. Even better, the Ravens and the Browns came in second and third in win probability added per play on the plays they called after those timeouts, indicating that perhaps they weren’t wasted after all — they actually helped set the team up for success. 

Another AFC North team, the division champion Cincinnati Bengals, also makes an appearance, ranking fifth in WPA per play on a surprisingly high 20 plays after operational timeouts. But no team in the league benefited more from pausing the action midgame than the Miami Dolphins.

Not only did the Fins lead the league in WPA on plays after operational timeouts, they called eight of them over the course of the regular season, accumulating 51.1 percentage points of total win probability added. Brian Flores truly did nothing wrong.

At sixth, we find McVay’s Rams firmly in the black, averaging a point and a half of WPA on offense after “wasting” a timeout. Not too shabby! So does this mean that McVay has silenced the critics, and the nattering nabobs of negativism have been hoisted by their own petard? 

Well, maybe not. We still need to account for the value of those timeouts. Back in 2014, Brian Burke, now at ESPN, conducted an analysis that estimated the value of a timeout as being worth 3.1 points of win probability. If we use his finding and subtract it from the WPA per play in the chart above for each team, we can calculate a new metric that will tell us if the timeouts were worth the cost. We call the new metric Timeout Win Probability Over Expected, or TOWPOE, and it paints a pretty bleak picture.

Only two teams — the Dolphins and the Ravens — managed to extract surplus value from their operational timeouts. Whatever benefit the other 30 teams received from stopping the clock and regrouping before the end of the half (or game) didn’t come close to covering the costs of the timeouts they spent to purchase those precious extra seconds. 

Just look at the Chicago Bears. Matt Nagy hemorrhaged an average of 7.3 percentage points of win probability each time he called an operational timeout this year, according to TOWPOE. Not to be outdone, Matt Rhule and the Carolina Panthers called 23 probably terrible timeouts. Rhule accrued an astonishing total win probability debt of 96 percentage points, most in the league.

Even if Burke’s estimate of the value of a timeout is a little too high, Miami, Baltimore and Cleveland still were the only teams to bank more than 2 points of WPA per play after an operational timeout. Unless you believe timeouts are worth less than 2 points of win probability, which seems dubious, the conclusions really don’t change. 

For McVay and the Rams in the wild-card round, everything worked out in the end. On the play immediately following the Rams’ timeout, wide receiver Odell Beckham Jr. lined up in the slot in a trips-bunch formation, something Arizona was apparently unprepared for. The Cardinals secondary blew the coverage, and Beckham ended up wide open outside the numbers on the left side of the field. Odell caught a 15-yard pass from Stafford and rumbled another 16 yards before being pushed out at the 2. The Rams would go on to score a touchdown and beat the Cardinals handily, 34-11.

McVay wasted a timeout, and it didn’t seem to matter. But he shouldn’t count on that if he wants his team to get back to the Super Bowl. The takeaway for teams this year is clear: Calling a play — any play — is likely better than wasting a timeout in search of the perfect call. 

Check out our latest NFL predictions.