Thursday, August 22, 2013

Trickle Down Flacconomics? Flaccometrics?

I wanted to explore an idea here, though it probably won't lead to anything terribly important.  A lot has been made about Joe Flacco's playoff run, and the rather high TD/INT ratio he produced in this time (11 TDs and 0 int, in the playoffs).  Many people equate this with the team's decision to fire Cam Cameron, and promote Jim Caldwell to the offensive coordinator position.  "They took the handcuffs off of Joe!", the people seemed to cry.  Well, maybe.  Or maybe it was something else altogether.

It's perfectly natural to see a relatively major shakeup in the coaching staff, followed by a change in statistical performance, and think the two things are related.  Unfortunately, with this particular theory, there were some other significant changes that took place at the same time as the Caldwell promotion, that may be more plausible explanations for what happened.  Namely, this involves the reshuffling of the offensive line.

Starting with the first playoff game, against the Colts, the Ravens reinstalled Bryant McKinnie as the left tackle, who had been on the bench all season.  Do I think that McKinnie is a great left tackle?  No, not really.  Still, when motivated by paying off his debts to strippers, and not weighing 400#, he appeared to at least be a reasonable improvement over Michael Oher at the LT position (Oher's combine measurables also suggest that he shouldn't play LT).  This also allowed Michael Oher to move back to the right tackle position, where he is a also at least a moderate improvement over Kelechi Osemele.  Am I saying that Osemele is bad?  No.  I'm just saying that he probably isn't suited to playing tackle.  When Osemele is relieved at RT, he promptly moves inside to left guard, a position he is better suited for, replacing Jah Reid.  Again, this provides another incremental improvement.  So, by bringing one player off of the bench, this actually managed to improve 3 positions on the offensive line.  But is there any evidence that this actually is the case?

In Flacco's five years in the league, his stats have generally been fairly stable and predictable.  For now, though, we are only going to look at one specific section of his statistics, to see how often he gets sacked, and how he performs in games when he is getting sacked more frequently.  Starting with his rookie year (2008), Flacco was sacked once in every 13.375 pass attempts.  In 2009, it was once in every 13.861 attempts.  In 2010, it was once in every 12.225 attempts.  In 2011, it was once in every 17.48 attempts.  Finally, in 2012, it was once in every 15.17 attempts.  To provide a little context, the average number of attempts per sack in 2012, among the top 34 QBs, was 16.31.  So, in 2012 Flacco was basically pretty close to the average result (Flacco's result is the sixteenth best, the definition of average).  In fact, he has hovered rather close to the middle of the pack in this area throughout his career, with minor variances.  So, what happens if we dig a little deeper?

Well, let's first look at how Joe did in the regular season, under both of these coaches, before we start to examine what happened after changes were made to the offensive line in the playoffs.

Coach    Comp%         YPA        TD      INT   QB Rating    Att/Sack
Cameron  59.95 7.2 18 9 87.09 14.41
Caldwell 58.09 7.1 4 1 91.22 21
Flacco's Season Total 59.09 7.2 22 10 87.7 15.17

The problem here is that Caldwell only coached 3 games at the end of the season, one of which was a week 17 game where the starters only played for about a quarter.  Still, we can say that there was no significant movement to Joe's completion percentage, which actually went slightly down under Caldwell.  The same can be said for Joe's average yards per attempt.  His overall QB rating, and TD to INT ratio went up slightly in the last three games, though it's hard to say if this means anything.  While his attempts per sack number improved from 14.41 to 21, the limited sample size of these 3 games at the end of the season makes this hard to know how much value to place on this.  So, why not take a closer look at those three games?

Week    Comp%         YPA        TD      INT   QB Rating    Att/Sack
Week 15 50 6.4 2 1 76.5 13.33
Week 16 69.4 8.6 2 0 114.2  INFINITE
Week 17 50 4.3 0 0 61.5 8

So, when we look closer at these 3 regular season games under Caldwell, we see that one game in particular is boosting Caldwell's stats, and that is the week 16 game against the Giants.  In fact, this is the only game of the 3 where you could say that Flacco's statistics were good at all, as his numbers for weeks 15 and 17 were well below his season averages in every category.  Not surprisingly, the week 16 game is the only game of the 3 that the team won, and also the only game where he was never sacked.  In the other 2 games his actual attempts per sack actually got significantly worse than his average 2012 result (15.17, in case you forgot), though the week 16 game managed to inflate Caldwell's overall results in this time frame.  Even though this is a small sample size, it seems safe to say that Caldwell hadn't started sprinkling Flacco with pixie dust and unicorn farts just yet.  Now, what happens when the offensive line is reshuffled for the playoffs?

Week    Comp%         YPA       TD      INT   QB Rating    Att/Sack
Week 18 52.2 12.3 2 0 125.6 23
Week 19 52.9 9.7 3 0 116.2 34
Week 20 58.3 6.7 3 0 106.2 18
Week 22 66.7 8.7 3 0 124.2 16.5
Total 57.93 9.04 11 0 117.16 21

Hmm, so once the O-line was reshuffled, Flacco's stats magically improved in almost every area.  Overall he was being sacked 38% less frequently than was typical during the regular season, with even his worst Att/Sack result being above average (at least by the standards of what he was used to dealing with).  While his completion percentage bounced around a fair bit, it only really dipped significantly in games where he was throwing more deep passes.  More aggressive passes down the field tend to be lower percentage passes, but obviously bring greater potential rewards (it's the yin and yang of playing aggressively, I guess).  His QB rating similarly leapt by 29.46 points, and remained consistently high in each game.  Still, how do we know that it just didn't take some time for Caldwell to make his little tweaks to the offense?  Maybe he still does deserve a lot of the credit for this turnaround.  Or, perhaps, Joe just discovered the eye of the tiger?  It's possible, right?

To attempt to answer that question, let's look at Flacco's 2011 season.  2011 was probably Flacco's worst statistical season, short of his rookie year, though his offensive line was largely the same as it was in the 2012 post-season.  Since we know that he was sacked once every for 17.48 pass attempts in 2011, I will divide his 16 regular season games from that year into to groups.  In the first group, for weeks 1,3,10,11,12,13,16 and 17, we will have the eight games in which he was sacked less frequently than the average of 17.48.  In the second group, for weeks 2,4,6,7,8,9,14 and 15, we will have the eight games in which he was sacked more frequently than the average of 17.48.

2011 Season    Comp%         YPA        TD      INT   QB Rating    Att/Sack
Group 1 57.55 7.01 13 3 91.86 30.62
Group 2 57.57 6.36 7 9 71.82 12.91
Flacco's Season Total 57.6 6.7 20 10 87.7 15.17

Once again, when Flacco is getting sacked less frequently, his numbers radically improve.  When he is sacked more frequently, he is a mere mortal, or borderline Sanchezian (I swear, I like Flacco, so don't leave dog poop on my porch).  I don't mean this as an insult, since I suspect this is the case for most QBs (something to explore later!).  While the results aren't the exact same as what we saw in his playoff run they are surprisingly close.  In these games, when he is being sacked less frequently his average QB Rating leaps 20.04 points over his rating in games where he is being harassed.  Also, his 13 TDs, to just 3 INTs, isn't really that far off from the 11 TDs to 0 INTs which garnered so much attention in the playoffs(or 15 TDs to 1 INT, in Caldwell's 7 game tenure, to make the time span more comparable).  Similar to his performance in the playoffs, his yards per attempt also increases as he is given better protection, which likely is directly connected to making more aggressive passes when he has the additional time in the pocket.  One somewhat odd peculiarity, is that his completion percentages in both of these situations doesn't seem to vary much.

Now, the natural response to all of this is to say "duh".  Of course QBs do better when they aren't getting sacked.  The only reason I am mentioning this, is that in all of the hoopla, people seem to have largely ignored this little aspect of what happened with the Ravens in the playoffs, and give an awfully large amount of the credit to Caldwell and Flacco.  I'm not saying they don't deserve some credit (although I might be thinking it, at least in Caldwell's case).  I'm just saying that the offensive line seems to deserve some praise too.  They appear to have performed well above their normal standard in these 4 games.  Unfortunately this makes me wonder if such a fortuitous run is actually sustainable.  For the most part, I don't think it is.  In the last 5 years, the Ravens' O-line seems to have produced average to slightly below average results.  Except for the addition of A.Q.Shipley (who is just a wild card I am rooting for), I don't see any significant improvements to the offensive line as far as talent is soncerned, and as the history has shown, the line's results can vary quite a bit from game to game.  This 4 game stretch is most likely an anomaly.   Still, I think it does point towards the benefits of protecting the QB (again, "duh").

While there are numerous factors that can influence how often a QB is getting sacked, and some would argue that a great QB can make an offensive line look better, this is beyond the scope of what I can get into here.  In general, I am a big fan of the "average QB" (average meaning "cheap" in this case) behind a monstrous offensive line, though I know many people are opposed to this idea.  However you look at it, is entirely up to you.  Still, not that this illustrates things one way or another, but if you are interested in which QBs had the best and worst results in 2012, as far as attempts/sack, I'll list them below.

Player      Att/Sack
Eli Manning 28.21
Peyton Manning 27.76
Drew Brees 25.76
Matthew Stafford 25.06
Tom Brady 23.59
Matt Ryan 21.96

Player      Att/Sack
Andy Dalton 11.47
Jay Cutler 11.42
Chad Henne 11.00
Aaron Rodgers 10.82
Philip Rivers 10.75
Alex Smith 9.08

Except for the peculiar Aaron Rodgers, who can seemingly survive in the worst of circumstances (the cockroach of quarterbacks, and I say that with the greatest respect), I would say that those are two glaringly different lists.


  1. Andy dalton is on the list...? very suprising I always thought of the bengals line as very good. LT Whitworth is a very good LT. LG. is solid C is medicore RG is a run blocker who is OK in pass protection RT (fat and lazy as he mightve been) is pretty solid.... Anyway when looking at flacco... I agree that his protection will undoubtley go down this year.... the pre-season doesn't mean anything but still oher has looked bad.... Yanda hasn't played so maybe he'll open up more holes for rice (and pierce)and keep the right side in check.... As for the left side... Mckinnie has looked solid so far.... Osemele got raped my loutlelei.... Gradkowski and shipley look average to below average.... Flacco will probaly have anoither solid year.... (26 TDs 9 INts is my prediciton) But I don't get why flacco shined under caldwell... the playbook was basically the same (he tweaked it I suppose by adding more slot routes and over the middle routes) but there's nothing he really could've done to really improve the offense.... I think it was all Joe.... he knew what he had to do... send ray off in style.... He took 6 sacks in the playoffs.... which figures to 24 over 16 games... Not bad. Joe I think was really motivated to win.... and it just clicked.... Plus if he failed to win the SB he probaly would've been allowed to walk because he would've been viewed as a guy who couldn't win the big game...

    1. In Andy Dalton's case, his seemed to get better protection in 2011, when he averaged 21.5 attempts per sack. It's also kind of tricky with young QBs, since I think their general awareness improves some after the first couple years, which probably reduces the number of hits they take.

      As for Flacco, I suspect your prediction of 26 TDs to 9 INTs might be a bit optimistic. Considering that he has averaged 22.3 TDs and 10.6 INTs in the last three years, while having better receiving options than he will probably be playing with this year, I expect his results to go down a bit. Still, I can't blame you for hoping for the best.

  2. I'm not hoping for the best the ravens are going to be throwing a bit more now that flacco is being paid like an elite QB (he's not) to be an elite QB you must do it all.... Put up elite numbers win playoff games and SBs... his support cast isn't that great.... (save for torrey smith) but still the ravens will be throwing more.... rice will get more receptions than ever (he needs more touches) Pierce is a good back... but he shouldn't steal touches from the best playmaker in the AFC.... (RB wise) Everyone says arian foster is better but that's b.s he's injured rice has never missed a start since he's been the starter.