clock menu more-arrow no yes

Filed under:

Andy Reid may not be as bad with timeouts as you think, and there’s proof

New, comments

New research from our sister blog Acme Packing Company paints a very different picture of Reid’s behavior

Kansas City Chiefs v Seattle Seahawks Photo by Abbie Parr/Getty Images

When people complain about how Kansas City Chiefs head coach Andy Reid manages his use of timeouts, I have often remarked that, in the 26 years I have been covering the team, not a single one of the seven head coaches of the Chiefs had gone without being accused of poor timeout management.

“What’s more likely?” I always ask. “Is it that every single Chiefs head coach has been bad at managing timeouts, or that most fans focus on this when it turns out another timeout is needed at the end of a close game?”

One Wednesday, two different sources pointed us to some new research conducted by Paul Noonan at our sister SBNation blog Acme Packing Company. This data suggests that Reid’s management of timeouts might not be as bad as you think.

NFL Network’s show Good Morning Football brought up Noonan’s research on the air — and on Tuesday, I also got an e-mail from Arrowhead Pride reader Ian Buckley, who said he’d originally seen it mentioned on NBC Sports Pro Football Talk. Thanks for that tip, Ian!

Back in October, Noonan first sought to figure out if Green Bay Packers then-head coach Mike McCarthy was wasting more timeouts than other NFL head coaches (Sound like something you’ve thought about Andy Reid?). To figure this out, he downloaded NFL play-by-play data, counting the number of early timeouts for every team.

I defined “early” as any time outside of the final two minutes of the first half, and any time outside of the final five minutes of the game. Early timeouts are typically used to avoid procedure penalties, saving the team five yards, but research by Brian Burke at Advanced Football Analytics (among others) has shown that the relative value of using a timeout to avoid a procedure penalty is minuscule compared to using a timeout at the end of the game (or half) to regain possession.

Noonan concluded in October that McCarthy was, in fact, “wasting” more timeouts than other NFL head coaches, but he also noticed that successful teams like the Los Angeles Rams called plenty of early timeouts, too.

So now that the Packers season has concluded, Noonan downloaded the play-by-play data for the whole 2018 season. This time, he took a somewhat different approach.

My focus this time was to count early and late timeouts, but also to capture the Expected Points Added (EPA) of each play following a timeout. For those unfamiliar with EPA, this is a useful primer on the subject, but in short, EPA quantifies how much more or less likely you are to score after a given play. My ultimate goal was to show which teams use the most early timeouts, whether they use them on offense or on defense, and how much benefit they derive (or cost they incur) from doing so.

Noonan also made the cutoff for early timeouts in the first half three minutes instead of five, because he’d seen in the previous data that most first-half timeouts are called in the final three minutes. And he also eliminated timeouts called before punts or field goals from his calculations.

So what remains are what teams are able to do — either offensively or defensively — after calling a timeout. I’ve copied Noonan’s data and removed the columns for total EPA so you can focus on the key numbers — the raw numbers of timeouts called on offense and defense, and the average EPA for the successive play — and also included rankings (in parentheses) for each value.

Ready to be a little surprised?

Timeouts and EPA

Team Tot TO Off TO Def TO Tot Avg Off Avg Def Avg
KC 17 (18) 16 (21) 1 (3) 0.71 (1) 0.81 (1) 0.86 (27)
NO 25 (28) 24 (30) 1 (3) 0.55 (2) 0.63 (2) 1.43 (31)
LAC 13 (10) 12 (18) 1 (3) 0.30 (7) 0.50 (3) 2.19 (32)
MIA 12 (7) 11 (14) 1 (3) 0.48 (4) 0.43 (4) -0.95 (4)
HOU 14 (12) 8 (8) 6 (24) 0.13 (12) 0.42 (5) 0.26 (21)
BAL 13 (10) 6 (6) 7 (30) 0.09 (14) 0.38 (6) 0.16 (19)
LA 29 (32) 23 (29) 6 (24) 0.39 (6) 0.35 (7) -0.55 (7)
TEN 19 (21) 17 (24) 2 (8) 0.19 (9) 0.23 (8) 0.14 (18)
IND 4 (1) 3 (2) 1 (3) 0.41 (5) 0.22 (9) -0.98 (3)
CHI 25 (28) 25 (31) 0 (1) 0.18 (10) 0.18 (10) 0.00 (13)
TB 14 (12) 8 (8) 6 (24) -0.10 (17) 0.17 (11) 0.46 (23)
CAR 20 (23) 17 (24) 3 (13) 0.04 (15) 0.02 (12) -0.14 (11)
JAX 12 (7) 8 (8) 4 (17) -0.37 (25) -0.02 (13) 1.07 (30)
SEA 24 (27) 21 (27) 3 (13) -0.02 (16) -0.03 (14) 0.01 (16)
CIN 16 (17) 11 (14) 5 (21) 0.16 (11) -0.07 (15) -0.65 (6)
NE 12 (7) 7 (7) 5 (21) -0.32 (23) -0.09 (16) 0.64 (24)
OAK 27 (31) 25 (31) 2 (8) -0.16 (20) -0.17 (17) 0.00 (15)
PIT 20 (23) 16 (21) 4 (17) -0.36 (24) -0.18 (18) 1.05 (29)
MIN 5 (2) 3 (2) 2 (8) 0.52 (3) -0.20 (19) -1.59 (2)
BUF 14 (12) 9 (12) 5 (21) -0.16 (21) -0.24 (20) 0.02 (17)
NYJ 21 (26) 15 (20) 6 (24) -0.24 (22) -0.24 (21) 0.24 (20)
WAS 25 (28) 22 (28) 3 (13) -0.14 (19) -0.26 (22) -0.68 (5)
SF 20 (23) 18 (26) 2 (8) -0.11 (18) -0.31 (23) -1.69 (1)
ATL 14 (12) 4 (5) 10 (32) 0.20 (8) -0.49 (24) -0.48 (8)
CLE 8 (4) 8 (8) 0 (1) -0.51 (28) -0.51 (25) 0.00 (13)
GB 19 (21) 16 (21) 3 (13) -0.39 (26) -0.55 (26) -0.42 (9)
PHI 11 (6) 9 (12) 2 (8) -0.50 (27) -0.63 (27) -0.09 (12)
NYG 15 (16) 11 (14) 4 (17) -0.58 (29) -0.68 (28) 0.33 (22)
ARI 17 (18) 13 (19) 4 (17) -0.93 (31) -0.95 (29) 0.89 (28)
DET 9 (5) 3 (2) 6 (24) -0.90 (30) -1.09 (30) 0.80 (25)
DAL 18 (20) 11 (14) 7 (30) -1.01 (32) -1.11 (31) 0.85 (26)
DEN 7 (3) 1 (1) 6 (24) 0.10 (13) -1.36 (32) -0.34 (10)

What we see here is that in 2018, Andy Reid called about an average number of early timeouts — almost all of them on offense. But most importantly, he got more benefit from having called those offensive timeouts than any other coach in the league.

What this data does not take into account is the reason a timeout is taken — which can’t really be determined from the raw play-by-play data, and it sometimes isn’t even clear if you’re watching the game. Is the timeout taken because it’s third (or fourth)-and-long in a critical game situation, and you want to be sure that you’re calling the right play? Or is it because the play clock is about to expire... or you have too many men on the field... or the defense lined up in a way that would crush the next offensive play? There’s just no way to know any of this from the raw data.

But here’s what we know: once a timeout is called — for whatever reason — coaches always have more time to figure out what to do on the next play. And in 2018, Andy Reid benefited more from those offensive timeouts than anyone else.

Having said that, it should be noted that there’s a little bit of bias built-in to this data. A very good offensive team will tend to do better on offense whether a timeout is called or not, so it’s not a surprise that teams like the Chiefs and Rams would be near the top. Still, there are other teams near the top that aren’t offensive powerhouses, so it might be interesting to see how these post-timeout numbers would line up compared to the EPA average for all a team’s offensive (and defensive) plays.

In any case, this is a fascinating project that Paul Noonan undertook, and my hat is off to him! I invite you to read his October article — and the most recent one, too — to learn more.