Turnover Margin and Predictions

Joe Fortenbaugh with a cool article.  Usually the kind of thing I like to research and I’m a little surprised I hadn’t thought it of it before.

I disagree with some of his conclusions (of course) because I think he looks too much at single seasons without looking at spreads of seasons for anomalous numbers.

First the link:  

Now the chart:

From this it is easy to look at 3-4 year spreads and ask “what number doesn’t fit”.  For example, Kansas City’s +9 in 2010 was out of character for them and corresponded with a somewhat improbable playoff run.

Looking at 2011 we see that Fortenbaugh highlighted Jacksonville’s jump, but the real anomaly was 2010, not 2011.  2011 was corrective.

So looking at this, the teams who I would expect to be “corrective” in 2012 would be (in order of confidence):

San Francisco     -15.5

Philadelphia      +13.5

Pittsburgh          +12.5

Tampa Bay         +11

I would normally expect Baltimore to improve by about +6 but the loss of Suggs may mitigate this improvement a bit.  I’ll leave it to you guys to decide whether last year’s dip was merely normal variation or a combination of weak offensive line play and aging playmakers.  You’ll recall Reed’s somewhat bizarre INTsplosion in a half season in 2010.  It could be that +2 is closer to the new normal for the Ravens.

I would like to point to Arizona.  A team I already thought was a bit remarkable, going 8-8 with no quarterback and trying to overcome heavy losses from the ’10 offseason.  Additionally we see that their TO margin also dove.  While this could directly correspond to their personnel problems, it could also point to a team that suffered bad luck and might be worth a couple more wins.

Finally, looking at Denver.  This is an interesting team.  Made the playoffs (and won a game!) last year with an absurd run of 4^th quarter comebacks.  Now they’ve replaced their awful QB with an all-time great.  It is impossible to predict exactly what Peyton will be, but it is fairly easy to project that he will be better than Ortonbow.  This is a team that could take dramatic steps forward, given the right Manning, similarly to Minnesota in Favre’s penultimate season.

Not predicting, just predicting.

Read more...

The Manning Who [didn't] Choked Away the Superbowl

Today, the discussion in the media regarding the Manning INT is all about how great a read Tracy Porter made, and how Peyton Manning has [in some cases "yet again] blown it in the clutch in the big game. Throwing the INT on a late 4th quarter, game-tying (saving?) drive, is certainly damning evidence.

The problem is, it's surface perspective only, and doesn't tell you the real story...that Manning isn't at fault for that INT, and didn't choke this Superbowl away with that throw.

Disclaimer - I'm not an NFL coach, so realize that I have no true experience breaking down game tape or telling anyone the precise technical nature of the game. All I can do is tell you what I saw, upon repeated viewings, and what I believe based on what I saw.

Porter is getting a lot of credit for making a great play. But I'm not sure he's getting enough credit for how great it truly was. Manning is getting much of the blame for throwing the INT. I don't think he deserves any of it. If anyone on the Colts deserves any blame for the pick, it's Reggie Wayne.

I saw a great angle on one of ESPN/NFLN (I was flipping between the two for about two hours post-game last night) which showed what looked almost like a coach's tape view of the play...virtually the entire field was shown. This was a terrific angle, and I wish I'd grabbed it on TiVo to keep to watch more than the couple times I flipped back through it last night.

The timing of the three players is what's critical. After the snap, as Manning drops, Wayne begins what looks like a go route, headed straight up-field. Manning appears to be reading the whole left/center side of the field on drop-back. It looks to me as though he begins his throwing motion prior to Wayne breaking off and cutting in. At this point, Porter isn't breaking on the ball. Porter in fact looks to be back-pedalling, probably ready to pivot and run, giving Wayne cushion to run the route with him as there didn't appear to be much help over the top if Wayne broke straight.

As Manning is in throwing motion, Wayne stutters his feet. His strides get noticably shorter, PRIOR to his breaking off and cutting in. This looked to me to be the key to the play, because the moment this stutter begins, Porter plants to stop the back-pedal, and breaks on the ball. By the time Wayne has begun his cut, Porter has begun his break, Manning is letting the ball go...and the game has been decided.

Many of the defensive backs and wide receivers converted to analysts - Deion Sanders, John Lynch, Michael Irvin, etc - seem to be giving Porter the credit he deserves. I heard a few mentions of how much preparation plays into making a play like that. Many others have paid lip-service to how Manning blew it by throwing the critical interception that sealed the game. I've heard at least as many uses of the term "choke" for Manning as I've heard "preparation" for Porter.

To me, it's a shame. While Manning's had his share of "choke" moments in playoff games in the past, he actually played a very good game last night. I don't think he deserves the heat he's taken for that INT. His receivers actually cost him badly last night. Not just in Wayne's telegraphing the throw, but his drop of the 4th down attempt that could have pulled them within 7 to lead to an onsides kick with 44 seconds left, and Garcon's 3rd down drop in the second quarter, after which Manning only got 5 more possessions in over 40 minutes of game-time and were outscored 28-7.

Manning, to me, doesn't deserve the criticism he's receiving from the pick.

Read more...

A Stafford In Peyton's Clothing

As my love hate relationship with Matthew Stafford - or more specifically his diehard fans - continues, I grit my teeth every time I hear him compared with Peyton Manning. And while my gut knows that these two quarterbacks are nothing alike, that even if Stafford has a trajectory similar to Peyton that his style and career will be totally different, I also can't ignore the simple statistical oddity that at this point in their respective careers (first ten games) they are remarkably similar.

Oh no doubt that Peyton put up slightly better numbers across the board, and no doubt that each quarterback got to this point in wildly different ways the juxtaposition of their numbers here is really fairly remarkable:

Comp    Att      Pct    Yards   TD     INT      Rate
209     378      55.3%   2289   15      20      64.6
201     377      53.3%   2267   13      20      61.0

So yeah, if it wasn't already obvious the top number belongs to Peyton but you really have to squint to see the differences and really the touchdown figure is the only thing that stands out even a little, and frankly that is possibly the most random number that a rookie quarterback will generate.

Other similarities?

As I said, they each got to this point in different ways so there really aren't many. Each had gone over a 90 passer rating in a game exactly once, and for each that particular rate was actually between 110 and 120 (118 for Peyton, 113 for Stafford). Stafford had a much greater deviation about his passer rating with two other games in the high 80s and five games under 50. Peyton barely creased 80 one other time but he also only had two games under 50. Stafford has had two games with more than three INT, one with four, one with five. Peyton had none*, however Peyton had 8 games with either two or three while Stafford has had fewer at six.

On other similarity? Both quarterbacks were 2-8 in their first ten starts.

While the similarity here is startling it is difficult to imagine that it is much more than coincidence. At this point of his rookie year Peyton was still riding a magnificent improvement curve while Stafford seems to have stalled a bit (for a variety of reasons) and so it is hard to expect Stafford to match Peyton's rate of improvement from his 11th game on.

But even so, the numbers are strange.

*it wouldn't be until Peyton's fourth season that he would throw as many as four interceptions in a game. To date he has done this twice, one four INT game and one six.

Read more...

What Quarterbacks Learn From The Bench

About once a week I see someone making the argument that quarterbacks can't learn from the bench. This is an unfortunate consequence of spending too much time reading Detroit Lion discussion boards and seeing all of the excuses posters make justifying ... well, virtually everything that Matthew Stafford does. Apparently some time last summer Peyton Manning told some sportswriter that quarterbacks can't learn from the bench and the quote was siezed by a vocal group of fans wanting so see Stafford play.

Never mind that if there is one thing that Manning doesn't know about quarterbacking it is how much can be learned from the bench.

This post isn't about Stafford. At this point there is no particular reason not to play him, assuming he is reasonably healthy. It is more about the question of how much can be learned by watching the game, particularly by rookie quarterbacks.

For one thing this should be a reasonably simple exercise in counting, and I will do that. For another it is simply a matter of generating a composite season of those quarterbacks who got a lot of playing time as rookies versus those who did not. To avoid survivor bias I will try to focus on quarterbacks who are current starters, although to be fair quarterbacks who played extensively as rookies and ultimately failed should at least be considered.

For the first part, we will simply count current starting quarterbacks to see how many played as rookies. My minimum is five starts, which may seem a bit arbitrary, but it is to avoid those frequent situations where rookies gathered a couple of starts at the end of a lost season. This is the eyeball test with the Alex Smith Exception which will be explained later (the is also known as the Vince Young Exception. You can see where it's going).

Team                      Quarterback       Rookie Starts
NFC East
New York Giants            Eli Manning            7
Philadelphia Eagles        Donovan McNabb         6
Washington Redskins        Jason Campbell         0
Dallas Cowboys             Tony Romo              0     

NFC North
Chicago Bears              Jay Cutler             5
Green Bay Packers          Aaron Rodgers          0
Minnesota Vikings          Brett Favre            0

NFC South             
Atlanta Falcons            Matt Ryan             16
Carolina Panthers          Jake Delhomme          0
New Orleans Saints         Drew Brees             0

NFC West      
Arizona Cardinals          Kurt Warner            0
Seattle Seahawks           Matt Hasselbeck        0
St. Louis Rams             Marc Bulger            0

AFC East
Buffalo Bills              Trent Edwards          9
Miami Dolphins             Chad Henne             0
New England Patriots       Tom Brady              0

AFC North
Baltimore Ravens           Joe Flacco            16
Cincinnati Bengals         Carson Palmer          0
Cleveland Browns           Brady Quinn            0
Pittsburgh Steelers        Ben Roethlisberger    13

AFC South
Houston Texans             Matt Schaub            1
Indianapolis Colts         Peyton Manning        16
Jacksonville Jaguars       David Garrard          1

AFC West
Kansas City Chiefs         Matt Cassel            0
Oakland Raiders            Jamarcus Russell       1
San Diego Chargers         Philip Rivers          0

From this you can see a few omissions. Three of them are teams currently starting rookie quarterbacks and for whom we have no real baseline. The other three are San Francisco, Tennessee, and Denver. In each of those case their current quarterback played extensively as a rookie and then ultimately was benched. Evidenced by the more recent play of Orton, Smith and Young it is safe to hypothesize that each benefited from watching the game from the sideline. It is possible that Russell or Kyle Boller may also join this class of quarterbacks.

Now if you were to say 'that's a nice list but it really doesn't tell us anything', I would probably agree. The one thing that I would argue is that it demonstrates that there is no inherent advantage to playing as a rookie.

The second peek may be a little more interesting. As I see it, if there is no benefit to sitting and watching then quarterbacks who play right away as rookies should accrue very similar statistics to those who wait and watch. My gut says that this simply cannot be so, but it is simple enough to test. If we build composite seasons of the two groups we can make a head-to-head comparison. What I've done is to take the first 16 starts of every quarterback and then combined them into average seasons, one for each group.

So here are the first sixteen starts for every current starting quarterback in the NFL (omissions are for QBs who do not have 16 starts yet):

                   Att Comp Pct Yards TD INT
Tony Romo          323 510 63.3% 4348 31 19
Eli Manning        248 494 50.2% 3079 21 18
Donovan McNabb     280 508 55.1% 2753 20 16
Jason Campbell     266 468 56.8% 3032 19 13
    
Kurt Warner        325 499 65.1% 4353 41 13
Alex Smith         228 403 56.6% 2502 10 18
Matt Hasselbeck    262 447 58.6% 2833 10 9
Marc Bulger        347 545 63.7% 4262 28 19
    
Jay Cutler         275 437 62.9% 3385 22 15
Aaron Rodgers      341 536 63.6% 4038 28 13
Brett Favre        323 502 64.3% 3390 20 17
    
Matt Ryan          265 434 61.1% 3440 16 11
Jake Delhomme      312 530 58.9% 3688 19 22
Drew Brees         320 526 60.8% 3284 17 16
    
Trent Edwards      289 468 61.8% 3240 13 12
Tom Brady          313 481 65.1% 3360 23 13
    
Kyle Orton         212 406 52.2% 2053 9 14
Matt Cassel        338 537 62.9% 3782 21 13
Jamarcus Russell   210 398 52.8% 2631 14 10
Philip Rivers      284 460 61.7% 3388 22 9
    
Joe Flacco         257 428 60.0% 2971 14 12
Carson Palmer      332 529 62.8% 3683 26 20
Ben Roethilsberger 219 335 65.4% 3133 21 9
    
Matt Schaub        298 474 62.9% 3424 16 16
Peyton Manning     326 575 56.7% 3739 26 28
David Garrard      284 483 58.8% 3258 17 12
Vince Young        216 400 54.0% 2492 14 14

And finally, here are the two composite seasons:

Average NFL Starter With Limited Experience As A Rookie, First 16 Starts:

COMP ATT PCT YARDS TD INT     RATE
305 495 61.6% 3547 22 15      85.5

And here is the composite first 16 starts of all current starting quarterbacks who accrued at least five starts their rookie season:

COMP ATT PCT YARDS TD INT RATE
256 444 57.6% 2981 17 15 76.8

So while it clearly isn't a disaster to start a player as a rookie, there is significant evidence that quarterbacks in fact do grow quite a bit from observing the game, participating in practices, film sessions and off-season activities, to the extent that with this additional experience their first "full" season is significantly better than that of the player who is simply thrown into the fire.

Read more...

Manning and Colts converting 3rd down

My dad and I were talking at the Colts/Ravens game Sunday and I was saying on one of the third down plays that I actually would rather see the Colts in 3rd and 1 or 2 than 3rd and 4 or 5 cause it just seems like Manning always picks up the latter, while the former might not be as easy since there’d be a greater chance they run the ball and the Ravens had a better chance of stopping the run.

So I decided to test that theory, looking at Manning’s conversion rate (first downs or touchdowns divided by total third down attempts, not including kneels or spikes), the Colts overall conversion rate (adding rushes into the mix), and then for the heck of it looked at the league’s overall conversion rates.

On third downs, with the following yards to go, this is the conversion rate (first down and touchdowns divided by attempts, removing penalties, spikes, kneels, and other non-plays) for Manning throwing, Colts overall, and the league overall. Note that the Colts overall attempts and conversion rates include Manning's passing.

3rd and 1
Manning: 60% (5 att)
Colts: 67% (18 att)
League: 66% (430 att)

3rd and 2
Manning: 50% (8 att)
Colts: 55% (11 att)
League: 54% (300 att)

3rd and 3
Manning: 60% (10 att)
Colts: 50% (16 att)
League: 45% (332 att)

3rd and 4
Manning: 56% (9 att)
Colts: 50% (10 att)
League: 43% (329 att)

3rd and 5
Manning: 58% (12 att)
Colts: 58% (12 att)
League: 39% (363 att)

3rd and 6-9
Manning: 42% (33 att)
Colts: 40% (35 att)
League: 33% (1,145 att)

3rd and 10+
Manning: 26% (19 att)
Colts: 25% (30 att)
League: 20% (1,149 att)

Not surprisingly, I was wrong about rate of success from those distances. Also not surprisingly, the Colts are better at converting on third down from pretty much any distance than the rest of the league. In some cases, a great deal better. In particular, note how much better they are at converting 3rd downs between 5 and 9 yards. Speaks a great deal to their offensive success.

Read more...

A look at Joe Flacco at mid-season

I’ve gotten my hands on a wealth of data through the entire NFL season, thanks to the Football Outsiders game charting project. Thanks to that data, I was able to cut down into some very specific data regarding Joe Flacco’s statistics. I’m not talking about his overall stats. I’m talking about being able to look at things like how he does on 3rd down with 3-9 yards to go when the opposition is winning. Thanks to this, I wrote an article for profootball24x7 (which I'll link here once it gets published).

I cut Flacco’s passing data using the following criteria:
- “No plays” (basically penalties accepted) were not considered
- Who’s winning – Baltimore, the opponent, or a tie game
- What down it is
- Result of the pass – Complete, incomplete, intercepted, sacked, intentional grounding or aborted snap
----- There was only one intentional grounding (in the Pats game) and one aborted snap (the Denver game, where the snap hit Heap)
- Distance to go for a first down
----- Long = 10+ yards
----- Med = 4-9 yards
----- Short = 3 or fewer yards
----- This was a somewhat arbitrary cut. C’est la vie.


So for instance, I can tell you that on third downs with 10 or more yards to go, when the Ravens are losing, Flacco is 5/8 for 56 yards, converting a first down two of the eight times and not throwing a TD nor an INT, and was sacked twice.

In addition, I’ve recorded the number of plays we’ve run in certain situations. Not surprisingly, when losing and facing 3rd and long, the Ravens have attempted a pass ten out of ten times.

Some general observations about some various situations…

Flacco’s 3rd down passing mirrors his complete passing stats.
On third downs, Flacco is 53/83 (64%) for 540 yards (6.5 YPA) with 5 TDs and 3 INTs for an 87.4 QB rating. This is comparable to his 90.2 overall rating, where the only measurable difference is in his YPA. He’s taken 5 of his 17 sacks (29%) on 3rd downs. But the Ravens have had only 88 third downs in 309 passing plays (28%).

Generally, it’s a good sign that Flacco is as good on third downs as he is on first, second and fourth downs. This is typically a pressure down, and defenses are typically playing the pass more frequently, especially in medium and long situations. He has converted 38 of all 88 attempts into first downs or touchdowns (43%). In third and medium/long situations, he’s converted 33 of 82 (40%).

When the Ravens are losing, Flacco is solid, but can improve
When playing from behind, it shouldn’t be too surprising to find that quarterbacks’ don’t perform as well. Typically a team has to pass more, so the defense is playing against the pass. The Ravens are no exception. They ran a passing play on 68% of all plays when losing, vs. 53% when tied or ahead.

Flacco’s passer rating is a paltry 76.4, but primarily due to the fact that his TD:INT ratio is fairly poor compared to the rest of the time. Flacco is 85/139 (61%) for 940 yards (6.8 YPA) with 3 TDs and 4 INTs. He’s taken 10 sacks.

Of note, much of the worst of his numbers was during the Bengals game. Flacco never ran a play when tied or ahead. The Bengals game was by far his worst game, completing 56% with 2 INTs and 4 sacks. Cut that game out, and Flacco’s numbers look like this:
67/107 (63%), 745 yards (7.0 YPA), 3 TDs and 2 INTs with 6 sacks
Once again, his stats look very similar to his full stats. Now, you can’t simply take that game out of the mix…it happened, and it was a poor performance. But in general, Flacco doesn’t perform too badly when losing.

However, there’s room for improvement. Take Peyton Manning as what could be considered the gold standard. Manning’s numbers this year when behind look like this:
96/133 (72%), 1,120 yards (8.4 YPA), 7 TDs, 1 INT and 5 sacks
His 111.7 rating is better than his 105.2 rating through the whole season. His 7:1 TD:INT ratio is gaudy. Ideally, we’d like Flacco when losing to be as good as he is when winning or tied. Honestly, we’d like him to be better, but it’s telling that even Peyton Manning isn’t much better when behind as when tied or ahead.

Flacco has been very impressive with medium yardage (4-9 yds)
This isn’t just third down. It’s any down, and that can include first downs when they’ve accepted a 5 yard penalty, or first and goal from the 4-9 yard line.

Flacco boasts an impressive 104.2 QB rating, albeit thanks to a stellar TD:INT ratio. His stats look like this:
63/101 (63%), 657 yards (6.5 YPA), 7 TD, 0 INTs, 3 sacks
Flacco converted 48 of 104 (46%). Compare this to his converting 108 of 309 (35%) total passing plays he’s run, and 53 of 185 attempts from long (10+ yards) situations (29%).

His performance with moderate yardage to go is his money situation; he’s a top flight QB when facing medium yardage to go. And interestingly, the Ravens run passing plays on 68% of these plays. His performance in these situations may have something to do with that.

Bottom line
In general, Flacco performs fairly well in tough situations, but has room for improvement. He’s young, and it’s good to see he’s playing as well as he is in these situations.

Better yet, I think we should expect that as Flacco matures – and hopefully gets better receivers eventually – we’ll see continued improvement in these numbers as well. Part of his education is hopefully going to involve playing better in the pressure situations, and we’ve already seen him playing fairly well in them.

It’s an encouraging sign that we have a legitimately good quarterback under center, not just one that’s good any time except when it counts.

Read more...

Expectations for the Now Sophomore QBs

Matt Ryan and Joe Flacco bucked a popular media trend of rookie QBs that do not play well. I say “a popular media trend” because while QBs that start as rookies don’t tend to play as well as they do for the rest of their careers, there are certainly examples of those that have played well. I’ve read some call Matt Ryan’s season possibly the best rookie season since the “passing era” began in 1978, which is difficult to argue in the shadow of Dan Marino and Ben Roethlisberger. I’ve also read an even more ludicrous suggestion that Flacco had the better season of the two, simply because his team made it to their Conference Championship Game…apparently some people see only one player on the field at a time, while most others see twenty two.

But besides that point, the reality is that both Flacco and Ryan had good rookie seasons. Because of this, fans of both franchises are breathing a sigh of relief in belief that they’ve finally solved their quarterback position. In Atlanta, erasing the stain that was Michael Vick was a true achievement for Ryan. Flacco’s version was similar, though Kyle Boller left a different kind of taste in fans’ mouths.

The excitement in Atlanta and Baltimore comes not simply from the two quarterbacks having a solid season, but also from the expectation that they will improve from their rookie campaign. The question is, just how much should we expect them to improve? Many seem to believe both will get better this season. A few suggest the possibility of a sophomore slump (particularly in Baltimore, and especially when Derrick Mason was seemingly retired). How likely is all that to happen?

I went to the books to look at historic trends for QBs that had mediocre to exceptional rookie seasons, and their trends in future years. But first I had to define which QBs would fit the criteria.

I decided to stick to the “modern era” of passing. In 1978, the NFL passed the rule that defensive backs could not come in contact with receivers five yards beyond the line of scrimmage. This ushered in a new era of passers, and quarterback (and receiver) play has for the most part never been the same since. I also decided in a somewhat arbitrary cut to say a QB rating of 70 or better their rookie season would be defined as “mediocre or better.” This isn’t a perfect methodology; it for instance includes Tim Couch, but Rick Mirer is left on the outside looking in. But for the most part it serves as a reasonable criterion for a cut. Finally, I didn’t want to include someone who made a nice splash in a few blow-outs or for two weeks in injury duty, so I set another somewhat arbitrary cut that the rookie QB needed at least 200 attempts to qualify.

This gave me a nice list of thirteen quarterbacks in my sample. Chronologically they are Jim McMahon, Dan Marino, Warren Moon, Jim Kelly, Jeff George, Tony Banks, Jake Plummer, Peyton Manning, Charlie Batch, Tim Couch, Jeff Garcia, Byron Leftwich and Ben Roethlisberger. Taking the combination of their per-game stats, extending them over a full 16 game season, and then looking at the median and average performances of the group, we see the median give a QB rating of 80.7, and a 78.5 average rating. Both are fairly close to Flacco’s 80.2 rating on the year, though Matt Ryan had a far more impressive 87.7 rating.

From there I looked at the median, average and range of per game (NOT per start) performance for each QB’s rookie and sophomore seasons. I then calculated the percentage improvement of each group’s rookie to sophomore statistics, and applied those percentages to Ryan’s and Flacco’s 2008 stats.

Matt Ryan
	Comp	Att	Comp %	YD	YPA	TD	INT
By Median	322	495	0.65	3845	7.8	15	12
By Range
Max	382	569	0.67	5816	10.2	43	17
Min	211	340	0.62	2855	8.4	9	8
By Average	302	473	0.64	3854	8.1	18	10
Joe Flacco
	Comp	Att	Comp %	YD	YPA	TD	INT
By Median	312	489	0.64	3311	6.8	13	13
By Range
Max	370	561	0.66	5008	8.9	37	19
Min	205	335	0.61	2459	7.3	8	9
By Average	293	466	0.63	3318	7.1	16	11

By median – due to small sample size – is likely the better measure here. The ranges are calculated based on the rookie median, not average. Also, removing the high and/or low performers from the sample yields little difference in the results.

Some interesting trends in this data:
Completion Percentage – Nine of the thirteen players on the list saw an increase in completion percentage from Y1 to Y2.
Touchdowns and Interceptions – Only seven QBs had their TDs per game and TD:INT ratio improve their sophomore season. This partially explains the difference between the average and median TD:INT projections. We could call it the “Marino/Manning/Garcia” effect. Marino saw a massive increase in his TD production. Manning and Garcia a massive increase in their TD:INT ratio.
Yardage – Ten of the thirteen saw their per-game yardage totals increase. This could easily be attributed to the fact that nine saw their per game attempts go up.
Overall – One last thing worth noting. Only Warren Moon saw a decline in all three categories (yards per game, completion percentage and TD:INT ratio), and no others saw a decline in both completion percent and TD:INT ratio.

In this analysis, it’s probably less important to look at the actual numbers, and more important to look at the trending of the numbers. Their TD:INT numbers are likely to be similar to where they were before, but their yardage will probably improve as they get more reps. Generally it should be expected that they play better, but not necessarily a dramatic improvement.

My own personal belief is that I don’t think either is more likely than the other to either improve or decline. There are strong arguments for both going either way. Matt Ryan had a better season, so there’s less room to improve than decline. Additionally, Atlanta’s offensive line played significantly better than in ’07, and better than expected at the on-set of the ’08 season. They are young and could continue to improve, or they could regress to the mean and set Ryan back with them. Should Michael Turner suffer the “Curse of 370” it could negatively impact Ryan’s ability to build on his rookie year. And of course, Ryan could simply improve with a year of experience under his belt, and the addition of ageless, perennial Pro Bowler Tony Gonzalez as an additional receiving option.

The arguments for Flacco are similar. His offensive line is [mostly] young, and will have two or possibly three new starters this season. It could improve or decline and neither would be surprising. Cam Cameron will almost certainly open the offensive playbook up. But the Ravens also still lack the basic weapons to truly give Flacco solid support. Derrick Mason is in his mid-thirties and suffered a serious shoulder injury last season. If he goes down to injury, the Ravens receiving options are horrendous (see my statistical break-down of their odds in light of Mason’s retirement here). But Flacco has plenty of room to improve upon last season, and his stats in his first five weeks vs. the last eleven weeks are night and day. With one TD to seven INTs his first five games, through his final eleven he threw thirteen TDs to only five INTs, and he posted a gaudy 90.2 QB rating his final eleven games.

Either way, both are more likely to improve than they are to decline. And to fans of two franchises that have suffered through frustrations at the QB position for years, these two are a breath of fresh air.

Read more...