Putting Numbers to the "Steelers Championship Roster"

In Ivan Cole's recent article, Projecting a Steelers roster for a championship run in 2013, Ivan makes the case for a very similar roster to this season, with a few tweaks. Here was the first comment to his article:

Wow, how bold

To basically type up the 2012 roster, minus Wallace. I agree that Mendenhall is the most talented back, but I don’t think they’ll keep him… plus they can find someone just as good in the 2nd or 3rd round of the draft. I also wouldn’t be against drafting Cooper to replace Colon.

Apparently the writer forgot to use the sarcasm font for his first sentence, but it comes through just fine anyhow. And somehow or other, it really annoyed me. There were several reasons for this, some better than others. First of all, I just had a colonoscopy. (Sorry, TMI, I know, but it's probably part of why I'm so sensitive. Literally...) That's the bad reason. The second reason is, I got ripped a new one by a commenter in a post a few months ago for saying the 2012 team came close to being a great team, until the collective weight of injuries pulled them down. (I was vindicated when Dick LeBeau said more or less the same thing a month ago. Except, of course, in the eyes of those who think the old coots like LeBeau, and probably me as well, should be swept away in the youth revolution).

And this brings me to the final reason. Although I believe the writer was being sarcastic with the heading "Wow, how bold," in fact Ivan was being pretty bold. A lot of the commentary on this blog in the past several months has been some variation on "get rid of the old farts, draft a bunch of youngsters, and then we'll really have something!"

As unpleasant as the reminder is, let us revisit the excitement after the 2012 draft. We landed David DeCastro! We are going to be kings of the universe! We can finally get the old folks off the field! Except that Ramon Foster, under-appreciated vet, ended up holding down the right guard position for most of the season, and only moved to LG when Colon was unable to return. David DeCastro did not play nearly as well as Foster in replacing him. This isn't to say DeCastro doesn't, hopefully, have a much greater upside than Foster, or that he will not (hopefully soon) be a better player. But for last season, the most solid player on the line other than Maurkice Pouncey was Ramon Foster, at least according to Pro Football Focus. Pro Football Reference gave Max Starks a slightly higher Approximate Value than Foster, but both men were more highly rated than any other lineman, after Pouncey. So, considering Starks was not at all well thought of by PFF, Foster wins : )

I decided that rather than be annoyed by the comment, perhaps I should try to put some data to Ivan's proposal, and see where it leads us. So here goes. At this moment, I have no more idea than anyone else what it is going to show, and I'll attempt not to massage it to prove a point. (Or, as Mark Twain was purported to say, "There are lies, damn lies, and statistics.")

Let's begin with the assumption Ivan makes, without defending it—Ben Roethlisberger will be the starting quarterback. I don't intend to defend it, either. The Steelers just extended his contract. But what about Charlie Batch as his backup? I personally believe Ivan is correct. Batch wants to come back, and even if he loses the battle in training camp, we've been down that road before. Remember 2010, when he was Quarterback No. 4? Remember 2011, when he was No. 3? Remember 2012, when he was No. 3? Funny thing, he ended up as the backup every season.

But let's take another look at Batch, since we're going to question everything that isn't patently obvious (such as the Steelers have every intention of starting Ben at quarterback, Lord willing and the creek don't rise.) Blackadar made this comment:

QB – We need a young rookie and a backup. Batch was ineffective last year (QB rating 64.9) and we need someone who can actually step in and back up Big Ben effectively. A good backup last year would have probably gotten us into the playoffs. I wouldn’t be surprised to see a guy like Matt Cassell rejoin Haley in Pittsburgh.

Let's take a look at how Batch actually played last season, and compare it to earlier years. In 2012 he played in two games, and his average NFL quarterback rating was, as Blackadar correctly points out, 64.9. However, if we break those two games down, we find he played considerably better in the second game. In fact, his QB rating for that week (89.6) puts him ahead of the usual underachievers such as Brandon Weeden and Ryan Fitzgerald. His rating was also better than the following quarterbacks for Week 13:

Steelers Championship Roster

Would David DeCastro be on a Steelers championship roster if we named one today?

Matthew Stafford

Colin Kaepernick

Matt Ryan

Tom Brady

Andrew Luck

Andy Dalton

Drew Brees

and yes, the Most Highly Compensated Quarterback in the NFL, Joe Flacco himself. Flacco's quarterback rating of 61.9 in Week 13 was, if I am calculating correctly, lower than the 64.9 rating Batch averaged for the two games.

Nor did Charlie Batch have the advantage which often accrues to a young back-up, that of playing a team who is unfamiliar with him. He played Cleveland (not well, alas,) and Baltimore.

Let's have a look at Batch's average QB ratings for the seasons since he joined the Steelers as a backup:

I have not included games in which he came in for a few plays, only those in which he was the quarterback (even if he took over from an injured QB.) Notice that, aside from 2006, when he was active for half of the games (he was on the roster for eight games and had significant snaps in three of them) his average rating for the season sits right around 66.0. Now let's see how he got there:

Hopefully this chart format isn't too confusing. The top of the black area represents the quarterback ratings for his first game of the season, the gold for the second. Note that in every single case the rating of the first game is substantially lower than the second, even in 2006, his glory year.

Even in 2006 he only played most or all of three games. Of course, he started two games and played most of a third game in 2010. Curiously, his QB rating for the third game in both cases was in the 50s (50.9 in 2006, 57.8 in 2010). It's hard to say whether this is just an artifact of a small sample size or whether one good game is his limit. But, all in all, I think what this says is if Batch has to come off the bench for a couple of games, he won't play terribly well the first week but the second week will be more than respectable.

Many of you may not find this particularly comforting. However, it is probably a reality for most backup quarterbacks, particularly the sort of development projects the Steelers typically draft in the lower rounds. Because the Steelers aren't looking to actually replace their quarterback anytime soon, it doesn't make a lot of sense to draft someone high. And if you draft someone in the lower rounds, it's pretty hard to imagine them being much better than Charlie Batch, particularly if they aren't getting any game experience. There's generally a reason quarterbacks who fall to the lower rounds in the draft are never heard from again.

Pro Football Reference gave Batch an Approximate Value of 1 for 2012. This is pretty much the AV he gets every year in which he plays any amount, with the exception of 2006, when he played a lot more (and better.) He received a 2 in 2006. A 1 is a respectable rating for a backup quarterback. For comparison purposes I grabbed the first name that floated into my head for a backup QB in 2012, which happened to be John Skelton of the Arizona Cardinals. I looked him up, and PFR gave him a -1 AV last season. I've never even seen a negative AV, although I've seen plenty of zeros, including a couple of Charlie Batch's seasons.

For a further comparison, Kirk Cousins of the Redskins, who played so well in relief of RGIII, received an AV of 2. If the Steelers draft someone equivalent this April, bearing in mind the Redskins drafted Cousins at the top of the fourth round, and if said youngster looks better than Batch in the pre-season, I expect Batch will once again be relegated to third QB.

Would the Steelers bring in one of the FA quarterbacks soon to be available? It was suggested that Matt Cassel might wish to be reunited with Todd Haley. I strongly suspect, though, in a quarterback-hungry league, Cassel would be able to pull down a lot more elsewhere than the Steelers could afford (or would wish) to give.

The offensive line is the one area of the team where Ivan projected a fairly different lineup than the one in 2012. (Although it might be reasonable to ask, "which one of the 2012 offensive lines?") Since I just finished a massive post on the Steelers' offensive line, and since this unit is the one with the most change projected, I will merely refer you to my earlier post if you wish to look and numbers and rankings for last season's O line members. Since there is also historical data going back to 2008 you can look at performance trends for the older players if you wish.

As far as the receivers go, Ivan's projection didn't stir up too much controversy, other than those disputing the likelihood of the Steelers picking up Steve Breaston in free agency. The top of the order, e. g. Antonio Brown and Emmanuel Sanders, is really not in any doubt, assuming they are both healthy.

The tight ends were another story. Since many commenters seemed to divide into the pro-and-anti David Johnson camp, I will next look at the historical data for Johnson, since there is, obviously, none from last season.

Pro Football Focus doesn't have Johnson on their radar as a tight end in 2009, the year the Steelers drafted him in the seventh round. But in 2010 he shows up as both a tight end and a fullback. They liked him much better as a tight end. In fact, in a limited number of snaps he graded out better than both Matt Spaeth (no surprise) and Heath Miller (enormous surprise). Which I expect will just go to prove the illegitimacy of any conclusions based on PFF data for many of you. But let us persevere. (If it makes you feel any better, Heath ranked quite highly in 2009 and 2011.)

Here is the trend of his PFR AV for three years (they consider Johnson purely as a TE) and his PFF ratings for two years:

Both lines are ratings rather than rankings, hence higher is better. The thing to note is the steady upward progression. But what was this for? Well, Johnson's drop rate was always right around the league average, once he had enough targets to have a drop rate. (Note, however, in 2011 his drop rate of 7.93% was lower than Heath Miller's.) But where his gifts mainly seem to lie is in run blocking, and in 2011 he had one of the highest ratings for run blocking in the league. As to where he ranked overall, in 2010 he was ranked No. 26 (out of 104 ranked TEs) and in 2010 he was ranked No. 15 (out of 108 TEs). Like William Gay, David Johnson somehow or other became a player Steeler Nation loved to hate, but this view is not shared by the coaching staff. It is interesting to note it isn't shared by PFF either.

As a fullback, PFF wasn't particularly enamored of him, other than, once again, in blocking. But that was in 2009 and 2010. In 2011 no one had a single snap at fullback in Pittsburgh except Trai Essex, who had exactly one.

For one more view of Johnson we turn to Football Outsiders. Johnson doesn't appear on their charts until 2011, and even then it is in the 10-24 targets list, which isn't ranked. However, they give him a DYAR of 16 and a DVOA of 8.9%. (For an full explanation of these metrics follow the link.) To put these in perspective, the DYAR of 16 would have put him at around No. 26 out of 45 among the highly-targeted TEs. (It was No. 5 out of 23 among the players with 10-24 targets.) His DVOA (value per play) was considerably better than that of Aaron Hernandez.

Of course, Johnson is returning from a significant injury, and hasn't been playing for a year. But it is reasonable to look at the continual progress he made and understand why Ivan would suggest putting him into the TE roster. As with most every other position, all bets are off if the Steelers draft someone in the upper rounds, and whoever else is on the roster for training camp, Johnson will have to compete with them for a roster spot. But if we make the assumption the Steelers won't draft a stud TE, I think Ivan's supposition is quite a reasonable one.

As far as the running backs go, I've also recently put up an extensive post on the backs, as well as information on them as receivers. Both posts have a considerable amount of historical data. Based upon the information I dug up, my suggestion was to keep Isaac Redman. But I'm not opposed to keeping Dwyer as well, if it makes financial sense to do so. For that matter, I'm not opposed to keeping Rashard Mendenhall—I just don't think it is going to happen. I think he is likely to be happier, and play better, elsewhere. But I would love to be proven wrong. Baron Batch is already signed, and I think he will be kept as well.

So this takes us to the defense, where the bulk of the dissension lies. So let's have a look at the players at issue and see whether Ivan's projections make sense.

Here was his projected D-Line: Casey Hampton, Brett Keisel, Ziggy Hood.

Probably the most controversial member of this trio is Hampton. Most people on BTSC, at least, seem to think Keisel has got another good year in him, and Hood is signed through 2013. Cutting Hampton has the most "upside" in terms of cap relief, if I understand correctly what I've read. So is there any good reason to keep him?

Last season he no doubt began poorly as he attempted to return to game conditioning after rehabbing an injury. (In the case of someone as large as Hampton, "conditioning" takes on whole new layers of meaning.) The questions I wish to explore are these: did he improve over the course of the season, and was that improvement substantial enough to make him worth bringing back?

We all know how critical the nose tackle position is in the Steelers defense (or any 3-4 defense for that matter.) Although the NT isn't expected to make many "splash plays," they are expected to take on double teams and be able to get movement into the backfield, hopefully collapsing the pocket in the meantime on passing downs. How effective was Hampton?

According to Pro Football Focus, not very. They gave him a rating of -9.4 for the 2012 season, which put him at No. 77 of 85 rated tackles. However, I have a bit of a problem with how PFF rates tackles, as they don't differentiate between 3-4 NTs and 4-3 DTs. But let's take their rating as a given, and have a look at how he accrued this rating. I wondered if it was a matter of Hampton playing himself back into game conditioning. Here's what his week-by-week rating looked like:

Hmm. While this does show a general upward trajectory if you connect the highest points, and all of the lowest ones except Week 17, there are an awful lot of dip-downs. This made me wonder about Hampton's performance in a typical season, and I discovered something pretty interesting as a consequence:

The only string of consistently "above the line" games Hampton ever had during this five-year period, according to PFF, were the last six games of the 2011 season. (The final game shown, which dips way down, was the Wild Card game in Denver.) And notice that they came after an okay game and two bad ones, after he returned from an injury. Also notice the best game during this five-year period was the next-to last game of 2012, vs. the Bengals. This is consistent with something else I noticed—Hampton really turns up the heat for the Conference Championships. Whatever that means.

There isn't nearly as much direct assessment of nose tackles as there is for, say, running backs, and I couldn't find a lot of detailed information for seasons earlier than 2008, when PFF's ratings begin. But Pro Football Reference goes way back, as do the NFL rankings. Let's begin with the latter, and see how Hampton compares, in raw stats, with the other active nose tackles, and whether there is a trend:

I've set the upper parameter of the chart at 26, because this is the maximum number of players designated "nose tackle" I was able to count in any one season. (That's out of around 275+ defensive linemen ranked each season.) It was interesting to note the sharp uptick in the number of nose tackles. During the 2001 season I counted nine players listed as nose tackles. In 2005 there were 18. 2007 was the year there were 26, and the number has remained between 22 and 26 every since. I don't know whether this is a consequence of the increasing number of teams running a 3-4 defense, or the NFL being a bit more precise in how they label players, or both. In 2004 Chris Hoke played essentially the entire season, so there isn't a number for Hampton. I don't know that his rating is worth a good deal, but it does demonstrate that, in terms of production, Hampton is playing in the past two years at the average level for his career.

Although it is a rather broad brush with which they paint, it's also worth taking a look at the PFR Approximate Values, since, like the NFL stats, they go back to well before Hampton's career began. Here's what he was given each season:

Higher is better in AV, unlike the NFL rating chart above in which lower is better, and No. 1 is best...

I thought it would be interesting to compare the AV for a number of the well-regarded nose tackles around the league. A few of them even match Hampton's tenure:

I noted with interest that Isaac Sopoaga, who was mentioned as a possible cheap free-agent pickup for the Steelers this past week, only once in his career played better than Hampton (according to PFR, anyhow,) and played considerably less well in 2012. But I guess that's how the market works. Also note that, according to PFR, Hampton played better than any of the listed tackles except Vince Wilfork, including a youngster like B.J. Raji.

But before we decide whether Hampton ought to be brought back or not, let's take a look at his putative replacement, Steve McLendon. (McLendon was the middle-of-the-line guy in Ivan's alternate scenario.) And let me also make it clear I have no idea whether any of these vets we are discussing make sense in financial terms. I suspect Ivan doesn't, either. The front office probably doesn't know in some cases, until they find out how many accommodations the players and/or their agents are willing to make.

Steve McLendon played less than Casey Hampton. A lot less. The final ratio of snaps was 503 for Hampton, 139 for McLendon. Many people wondered why, when McLendon appeared to be so much more productive on the field. Certainly the PFF guys thought so. His rating for the season was 7.0, high enough to put him at No. 19 among all DTs if he had played a sufficient amount to make it through the 25% filter. (If you recall, Hampton was 77 out of the 85 ranked tackles.) This represents a steady improvement since he first got playing time in 2010. So why didn't the coaching staff play him more?

My rather cynical assessment was, the front office wanted to wring every drop of goodness possible out of Hampton in 2012, and shield McLendon from the notice of other teams, who would then drive up the value of the contract the Steelers have to negotiate with McLendon sometime soon. If the Steelers sign McLendon to a contract in this off-season, release Hampton, and McLendon tears up the joint next season, it would lend credence to this theory. But the rub is, I can't imagine the coaching staff would sit quietly by while a player who could greatly improve the team sat on the bench most of the time. So let's see if PFF found any notable deficiencies in McLendon's play.

After a short perusal, the answer appears to be no. He graded out better in pass rush (3.7) than run stopping (2.9), but even that run number would put him 25th out of 85. The rush number would put him 21st. And in their Signature Stat of Run Stop %, his 8.6 would tie him for 10th out of 82; his Pass Rush Productivity rating of 9.0 would tie him with Ndamukong Suh at No. 3. I don't need to tell you that Hampton's numbers were much lower. He was tied for last place in Pass Rush Productivity, and was No. 73 out of 82 in Run Stop %.

Here's one more thing to consider:

Like Hampton's chart above, we see good games punctuated by not as good ones. The difference to note is the much higher floor in the case of McLendon (but also a lower ceiling.) I wondered if how well or poorly he played correlated with the number of snaps he had in a game:

To get the actual snap count multiply by ten—I reduced them to put the figures in the same range. So, for instance, he had 17 snaps in the Denver game, etc. The answer to the question seems to be that there isn't really a correlation. Furthermore, there is no correlation whatsoever between his performance and whether the team won or lost that game.

So to return to the question at issue, should Steve McLendon be the Steelers' nose tackle for 2013, or should Casey Hampton be brought back, assuming the money is right? I think to answer that we would have to know for sure why the Steelers gave McLendon so few snaps in 2012, especially after his line coach sang his praises during the off-season. We would also have to know whether the defensive scheme is going to remain at the status quo or whether LeBeau has a new wrinkle requiring faster and more agile D-linemen. But I think the information about Hampton I was able to glean indicates that you pretty much know what he will be capable of next season, and it isn't going to be radically different than it has been for the past several seasons. It's much harder to project how McLendon would perform if he was taking the majority of the snaps. But if he does, we'll find out soon enough.

I'm not going to discuss the linebackers. Harrison will be back if he and the Steelers can come to terms. That's fairly obvious, and therefore makes any of our speculations moot. And yes, I realize he was just released. There is always the (faint) hope that he tests free agency, sees his value isn't particularly high, and re-signs with the Steelers. But I'm not optimistic. So Ivan's roster gets amended to move Jason Worilds into Harrison's spot.

In the case of Larry Foote, I'm expecting an announcement of a contract for him any time now. He clearly wants to return, and the Steelers want him back, although they may not pay much for the privilege. Foote already went down the free agency route and found it wanting, so I don't expect him to make much fuss. Whether he ends up as the second ILB by the end of the season, I expect him to start there, unless the Steelers land a stud in the draft who is an unusually fast learner. LaMarr Woodley isn't going anywhere. Any questions?.

As far as the defensive backs go, the recent signing of William Gay makes Ivan's alternate scenario (Cortez Allen in place of Keenan Lewis) way more likely. Especially as they are paying Gay real money. And a signing bonus. Once again, though, I don't think this lineup is controversial.

Special Teams is actually the place where I would have the most doubt about Ivan's scenario coming to pass. I think Shaun Suisham is an outstanding human being, and he had a great season last year, but it wouldn't completely surprise me to see him lose the battle in camp to some young whippersnapper with a much bigger leg and a reasonable degree of accuracy. Drew Butler certainly didn't set the world on fire, and I expect the Steelers to look for some serious camp competition for him as well. And as for Greg Warren, I liked malaki's suggestion—teach Doug Legursky to long snap, and then he can be a back-up center as well. Mind you, long snapping is harder than it looks, and as we saw last season, a poor snap can turn a chip shot field goal into a missed one.

And finally, can this group go to the Super Bowl? Heck, yes. Last season's team, which looked a good bit like this proposal, was, in my opinion, one injury to Ben, one bad long snap, and eight fumbles shy of competing in the playoffs, despite the injuries. And statistically the Steelers are due to regress towards the mean and have a year in which the normal number of injuries occur across the pack, rather than concentrating in the Pro Bowlers.*

Or at least it is a happy daydream for a sunny March afternoon in beautiful Pittsburgh.

*But wait, are the Pro Bowlers, who are in most cases the very veteran players, statistically more injury-prone? That's a subject for another post. Because right now, I don't know.