2014: The Year Offensive Line Continuity Died
We've got this little stat we use in Football Outsiders Almanac each year called the offensive line continuity score. A guy named Jason McIntyre came up with the idea for an article in one of our early books, and we still track it every year. We just ran the numbers for 2014, and wow, last year's OL continuity scores are really low.
As in, we didn't realize this before, but there was not a single team in the NFL that started the same five offensive linemen in all 16 regular-season games.
Continuity score takes three variables: number of starters used, number of times the starting line changes between games, and longest stretch starting the same five guys in the same five spots. It works out so that the highest possible score in each variable is 16, and the highest total is 48. We have this tracked going back to 2000. Every season at least one team had a 48, until last season.
The Packers came closest at 43. They started the same line for all 16 games except in Week 2, when Derek Sherrod started at right tackle for an injured Bryan Bulaga. Arizona was also close at 42, starting the same line except in Weeks 14 and 15 when Jonathan Cooper started at left guard and Ted Larsen moved over to replace the injured Paul Fanaika at right guard. Houston and St. Louis also had only six starters on the offensive line last year.
Surprisingly, San Diego was not in last place or even close. The Chargers had 27, ahead of eight other teams. Although they started five different centers, they also got 16 games out of three of their four other starters, and right guard Johnnie Troutman only missed Week 17. (Willie Smith technically had a start because the Chargers ran a six-OL set on their first play of Week 15, but we don't count that for continuity score.)
Instead, the lowest score belonged to Indianapolis at 19. This is the second-lowest continuity score since 2000. Only the 2007 Rams were worse at 17. The Colts last year had 10 different players start on the offensive line and changed the line 11 times. Anthony Castonzo started the whole year at left tackle, but the rest of the line was a total mess. At right guard, for example, the Colts switched back and forth between Lance Louis and Hugh Thornton six times between Week 1 and Week 12, then ended up with Joe Reitz for two weeks, then back to Louis. A.Q. Shipley started at center for four weeks, then Jonotthan Harrison for 10, and then Khaled Holmes for two. Jack Mewhort played most of the year at left guard, but when he was injured for two weeks the Colts started Louis in his place one week and then Reitz the next, and then in Week 17 Mewhort moved over to right tackle because Gosder Cherilus was out.
We're not even getting into the playoffs here, where swing tackle Xavier Nixon was out for the AFC Championship game because he missed the freakin' plane.
This is not a one-year problem for the Colts. Andrew Luck covers up a lot of problems on the offensive line with his awesomeness. The Colts were close to the bottom of the league with a 29 continuity score in 2013, and were tied for dead last in the league with 22 in both 2011 and 2012. The Colts haven't had an offensive continuity score above 30 since 2007. For comparison purposes, in 2014 the average team was at 30.7, and that was the lowest of any year in the data.
Here's the whole thing; thanks to Sterling Xie for helping me put this together.
| OL Continuity Score, 2014 | ||||||
| Team | Cont. Score | No. Starters | Changes | Longest Stretch |
Cont. Score 2013 |
Change |
|---|---|---|---|---|---|---|
| GB | 43 | 6 | 2 | 14 | 38 | 5 |
| ARI | 42 | 6 | 2 | 13 | 42 | 0 |
| KC | 40 | 7 | 2 | 12 | 28 | 12 |
| HOU | 39 | 6 | 2 | 10 | 38 | 1 |
| STL | 38 | 6 | 2 | 9 | 29 | 9 |
| BUF | 37 | 7 | 2 | 9 | 41 | -4 |
| NYG | 37 | 7 | 3 | 10 | 23 | 14 |
| NYJ | 37 | 7 | 2 | 9 | 42 | -5 |
| DEN | 36 | 7 | 2 | 8 | 36 | 0 |
| BAL | 34 | 7 | 4 | 8 | 37 | -3 |
| NO | 33 | 7 | 3 | 6 | 26 | 7 |
| TB | 33 | 7 | 4 | 7 | 33 | 0 |
| DAL | 32 | 7 | 4 | 6 | 37 | -5 |
| OAK | 32 | 6 | 3 | 4 | 22 | 10 |
| CLE | 31 | 8 | 4 | 6 | 35 | -4 |
| Team | Cont. Score | No. Starters | Changes | Longest Stretch |
Cont. Score 2013 |
Change |
| JAC | 29 | 8 | 5 | 5 | 30 | -1 |
| TEN | 29 | 10 | 4 | 6 | 29 | 0 |
| WAS | 29 | 8 | 5 | 5 | 48 | -19 |
| MIN | 28 | 8 | 5 | 4 | 34 | -6 |
| NE | 28 | 9 | 7 | 7 | 32 | -4 |
| ATL | 27 | 10 | 7 | 7 | 23 | 4 |
| PHI | 27 | 10 | 6 | 6 | 48 | -21 |
| PIT | 27 | 7 | 6 | 3 | 27 | 0 |
| SD | 27 | 10 | 5 | 5 | 28 | -1 |
| CAR | 26 | 9 | 7 | 5 | 25 | 1 |
| MIA | 26 | 9 | 6 | 4 | 29 | -3 |
| SEA | 25 | 9 | 7 | 4 | 29 | -4 |
| CIN | 24 | 8 | 8 | 3 | 32 | -8 |
| CHI | 23 | 8 | 9 | 3 | 48 | -25 |
| DET | 23 | 8 | 9 | 3 | 33 | -10 |
| SF | 22 | 9 | 9 | 3 | 39 | -17 |
| IND | 19 | 10 | 11 | 3 | 29 | -10 |
Comments
24 comments, Last at 26 Apr 2015, 2:49am
#1 by JoeyHarringtonsPiano // Apr 16, 2015 - 8:02pm
Interesting that 6 of the bottom 10 teams made the playoffs, but only 2 of the top 10 teams did. I guess it's the kind of statistical anomaly you get with small sample sizes. When you look at former group, the teams either had really good defenses or really good quarterbacks.
#4 by JoeyHarringtonsPiano // Apr 17, 2015 - 5:58am
You're missing Pittsburgh in the bottom 10. But you're right about the top 10. (My recency bias unfairly has the Lindley-led Cardinals burned into my mind and my brain rejected them as a playoff team). This is why I should triple check my counting before I post. I guess my overall point is this stat is interesting, but in 2014 had little corellation to team success.
#10 by hausta01 // Apr 17, 2015 - 2:19pm
Forgot Pittsburgh made it last year. If you look at the 2013 numbers for continuity 8 of the 12 playoff teams were in the bottom half of the league. Maybe this is hinting at playoff teams generally being overall better teams that are affected less by rotating offensive lineman.
#5 by carljm // Apr 17, 2015 - 11:59am
"Andrew Luck covers up a lot of problems on the offensive line with his awesomeness."
I feel like I've been reading this same sentence about the Colts (modulo QB name) for going on fifteen years now.
I guess it's better than having a great O-line and a crappy quarterback.
#11 by Bright Blue Shorts // Apr 17, 2015 - 3:12pm
Thought New England would have been lower given how they struggled early on having just traded Mankins. I guess once they settled in, things pretty much got back on track. Helps to have a QB who gets rid of the ball in under 2-secs.
#13 by justanothersteve // Apr 19, 2015 - 3:27pm
While the Packers may have had the best regular season continuity, it should be mentioned they lost their starting center during the third preseason game (after he had played all the snaps each game). Backup center Corey Linsley turned out to be a find and goes into next season as the entrenched starter, but this could have been a disaster for GB.
#15 by chemical burn // Apr 19, 2015 - 5:57pm
yeah, this seems like a blind-spot in these listings - are the Eagles getting docked for Lane Johnson missing 4 games? Because he's supposed to be the starter and there's a lack of continuity built into his absence. That is, there's missed continuity in those first 4 games because he wasn't there, regardless of who started in the opener. I think what you're saying about the Packers also qualifies as far as there being a "continuity" issue there when a back-up starts the whole season.
#17 by justanothersteve // Apr 20, 2015 - 12:13pm
A blind spot in that personnel changes due to injury may happen in preseason. It usually takes at least a few games for a new lineup to gel, especially when the reason for change is a rookie. While an offensive line change may be a coach's decision, I'd guess it's more often due to injury (either the injury itself or a player returning after missing games).
This is the second consecutive year the Packers have lost a key OL in preseason who was replaced by a rookie (C Linsley last year, LT Bakhtiari in 2013). Both times, it took the line several games to find their groove.
#19 by chemical burn // Apr 20, 2015 - 10:32pm
Yeah - that's what I mean. Why would it be less disruptive to "continuity" to lose a player late in the preseason (for several games) than it would be to lose a player for a game during the regular season?
And with a suspension, there's a lack of continuity built into: the line the Eagles started with was going to stop being their line in week 5, no matter what. There's already a disruption of continuity built in there at the start of the season regardless of what happens next.
I know that this is not really supposed to measure those sorts of things, but that's exactly why I feel like these numbers are sorta not useful. It starts from the premise that continuity is inherently good, whereas in the two scenarios just mentioned (getting a player back from a preseason injury or suspension) the disruption to continuity is a very good thing.
Except in the case of getting Lane Johnson back. He stinks.
#20 by Jerry // Apr 21, 2015 - 1:49am
I'd suggest that anybody who wants to can extend these numbers into the preseason, but how do you want to handle the last exhibition game, when many teams don't play starters at all? And, of course, the threshold for sitting guys in August is much different than it is in October.
This doesn't claim to measure anything more than continuity. It doesn't pretend to know if a given change is an improvement or not. And, in fact, reliably quantifying offensive line play is still an open question, so I wouldn't hold my breath waiting for a high-quality offensive line metric. In the meantime, this data is just fine for what it is.
#21 by chemical burn // Apr 21, 2015 - 11:19am
Well, but FO started compiling this info because they had found that o-line continuity has the greatest correlation to running game success. (Or maybe it was a more general offensive success?) Thinking about ways to refine this number and make it more useful isn't a bad thing.
But right now, it doesn't show much of anything, it doesn't even account for injuries very well (that's AGL's job.) This number is SUPPOSED to be indicative of something, that's why it is gathered at all. Accounting for some things like preseason injury seems like an easy-ish thing to do, as would a player taking into a player getting benched versus a player getting hurt - does that improve the correlation between o-line continuity and success or hurt it?
FO doesn't need any more poorly conceived o-line stats that everyone ignores, there's no need for another ALY yards in their repertoire...
#14 by lightsout85 // Apr 19, 2015 - 5:47pm
I would think San Diego's performance "ranks" lower that their score would suggest (although, obviously continuity is not the only thing that matters), BECAUSE the injury issues were primarily at C. Each team's system is different, but on the Chargers the center makes most/all of the calls(/audible/adjustments) that relate to the OL. If the issues had been elsewhere, a healthy Nick Hardwick could have smoothed them out. Once he went down early, it was all down-hill. (Even if *his* back-up hadn't gone down).
#22 by Bobman // Apr 26, 2015 - 2:45am
We're number One! We're number One! We're numb--- wait, that sucks!
Nate Dunlevy has been saying for years that OL play is overrated, and I guess the relative success of the "failures" in this table support that. But damn, I'd like a little more continuity, as would Andrew Luck I am sure.
Worth noting that the Colts' longest stretch of continuity of three games took place in the playoffs, where they saw some success. Better there than weeks 1-3.
I suspect that teams' schemes can either rely a lot on continuity or adapt away from it, so there's no single best answer. Manning, Luck, Brady, Rodgers teams can work around it, but my guess is there are teams for whom continuity is more critical.