Quick Reads
The best and worst players of the week according to Football Outsiders stats.

Divisional Round Quick Reads

by Vince Verhei

The Baltimore Ravens lost four games out of five to close out the regular season, but they've turned things around in the playoffs with two wins in a row to reach the AFC championship game. That two-game win streak coincides with Ray Lewis' return to the Baltimore lineup, and considering that Baltimore started hot this year with Lewis on the field, it seems that the Ravens' fortunes hang simply on the presence (or absence) of their future Hall of Fame linebacker. A closer look at the numbers shows that Lewis makes a real impact on the Baltimore defense, and that he has been one of the best players in the league this postseason. However, there is a hole in his game, and that weakness could cost the Ravens dearly against New England.

Lewis played in each of Baltimore's first six games, five of them wins. However, he tore his triceps late in the Week 6 win over Dallas and missed the rest of the regular season, with fourth-year man Dannell Ellerbe taking over most of his snaps. Including the playoffs, the Ravens have gone 7-1 with Lewis on the field, but just 5-5 in the ten games he missed. That's clearly a huge dropoff, from near-perfect to perfectly average, but you can't attribute that difference entirely to Lewis. In some ways, the defense actually played better without him. With Lewis on the field, the Ravens gave up an average of 20.3 points, 399 total yards and 137 rushing yards per game. (These numbers include the two playoff wins, but they barely change if we look only at Lewis' regular-season games.) In his absence, they actually cut their total yards and rushing yards allowed, to 324 and 115 yards per game respectively, although they let opponents score 22.6 points per game.

Football Outsiders' advanced statistics tell a similar story. With Lewis, the Ravens defense has posted a DVOA (explained here) of -3.4%. Without him, it's only a little higher, 1.2%. Those numbers, though, look at a team's overall defense, which is a little unfair when discussing an inside linebacker. Lewis' primary responsibility is to stop the run, and there's no question that it's tougher to run against Baltimore when Lewis is around. With Lewis, the Ravens fare much better in yards per carry (from 4.30 without Lewis to 3.98 with him), rush defense DVOA (4.6% DVOA to -5.0%) and running back Success Rate (54 percent to 48 percent).

Even there, though, rush defense is a team effort, not a one-man game. What do Lewis' individual numbers look like? At Football Outsiders, we judge defenders in three categories: Plays (any tackle, pass defensed or intercepted, or fumble forced or recovered), Successes (any play that holds an offense to a short gain), and Defeats (any play that causes a loss of yardage, a turnover, or a stop on third or fourth down.) In his six games in the regular season, Lewis averaged 9.7 Plays, 4.3 Successes, and 0.5 Defeats per game. For comparison's sake, Carolina's Luke Kuechly led all players in Plays (10.8 per game) and Successes (6.1) this season. Lewis looks respectable in those categories, but not in Defeats, where Philadelphia's DeMeco Ryans led all inside linebackers with 1.9 per game. Way back in 1999, Lewis himself once produced 45 total Defeats (2.8 per game), the best year of any player in our records until this season, when Houston's J.J. Watt had an stunning 56. Frankly, three Defeats in six games is a pretty impotent total for a starting linebacker. And two of those came in one game, as Lewis limited Andrew Hawkins to a 6-yard gain on third-and-10 and also recorded a sack-fumble in Week 1 against Cincinnati. It seemed that Lewis was finally slowing down when he was injured.

And then the playoffs started, and Lewis was reborn. Lewis leads all defenders this postseason with 24 Plays and five Defeats, and he's second with 10 Successes. Obviously, it helps that he has played two games, but even on a per-game basis he's first in Plays and among the top ten in the other two categories.

It is remarkable that Lewis has more Defeats in two playoff games than he did in six games in the regular season. Here's how those five plays break down:

  • In the first quarter against Indianapolis, he stuffed Vick Ballard for a 1-yard loss.
  • In the same quarter, he tackled a scrambling Andrew Luck for a 3-yard gain on third-and-7.
  • In the second quarter of that game, Luck completed a pass to Coby Fleener, and Lewis tackled Fleener for a 7-yard gain on third-and-14.
  • In the third quarter of Sunday's game, Denver had a third-and-5 near midfield. They handed off to Jacob Hester, and Lewis made the tackle after a gain of 2 yards.
  • Finally, Ronnie Hillman tried a run around left end in that same quarter, and Lewis met him behind the line of scrimmage for a loss of 3 yards.

[ad placeholder 3]

With all that said, there is one area where the Ravens have clearly played better without Lewis, and that is in pass coverage. When Lewis has played this season, opposing running backs have caught 70 percent of the passes thrown their way, and averaged 5.6 yards per target. Those numbers are a little lower than the average for running backs. In his absence, though, that catch rate fell to 62 percent, and the average gain to 3.9 yards.

That might not matter against some teams, but against the Patriots it could be a fatal flaw. To a large degree, New England is a running team (only Seattle had more carries in the regular season), and that means Lewis' performance on early downs will be crucial. However, the Patriots' running backs are also dangerous receivers. New England averaged 8.1 yards per pass when throwing to their running backs (mostly Danny Woodhead, but also Shane Vereen, Stevan Ridley, and Brandon Bolden) this year, highest of any team in the league. Vereen might have been the MVP of New England's playoff win over Houston, catching five passes in six targets for 83 yards and two touchdowns. Obviously, if the Pats split Vereen out wide against Baltimore like they did against the Texans, Lewis won't be vacating the middle to cover him. However, Lewis will be in charge of covering receivers out of the backfield, and that could be a serious problem. If Baltimore wants to win one more for Ray, they'll need to help him out on passing downs.

Quarterbacks
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
1.
Russell Wilson SEA
24/36
385
2
1
265
237
28
In a losing effort, Wilson breaks his own record for best total DYAR for a rookie quarterback in a single game. (With updated opponent adjustments, his Week 13 game against Chicago now stands at 214 total DYAR.) We'll start with what he did as as a rusher. His first four carries each produced first downs and a total of 45 yards. His fifth carry was a 1-yard touchdown. His last two carries were 9- and 5-yard gains on first-and-10. That's a 100 percent Success Rate on seven runs. As a passer, his first half was up-and-down, as he went just 10-of-17 for 144 yards with one very crucial sack. In four third-down dropbacks, he had one conversion (for 4 yards on third-and-1), two incompletions, and the sack. And then came the second half. On his first three drives, each of which came with a deficit of at least 13 points, Wilson went 10-of-10 for 185 yards with one sack. Nine of those completions led to first downs or touchdowns; the tenth was a 7-yard gain on first-and-10. Each of those three drives ended with a Wilson score via pass or run. His next drive was a three-and-out with three incompletions in a row, but on the next drive, needing a touchdown to take the lead in the final minute, he went 4-of-5 for 56 yards and three first downs. His final pass, an interception on a Hail Mary, is treated as an incomplete pass for DYAR/DVOA purposes. On deep balls alone, he went 8-of-13 for 195 yards and a touchdown.
2.
Tom Brady NE
25/40
344
3
0
224
224
0
Brady was a little erratic (for Tom Brady, anyway) towards his own end of the field, but when he neared midfield he was scary. Inside the New England 40-yard line, he went 12-of-22 for 113 yards with only five first downs and a sack. Beyond that point, he went 13-of-18 for 231 yards with 10 first downs, including three touchdowns. Those touchdowns came on gains of 8, 5, and 33 yards, and all came on first down.
3.
Joe Flacco BAL
18/34
331
3
0
203
201
2
Throwing to his right (where the bulk of the Torrey Smith vs. Champ Bailey action took place), Flacco went 11-of-17 for 261 yards with three touchdowns and seven other first downs, plus a 25-yard DPI. If we limit that to deep right passes, he went 4-of-7 for 185 yards with three touchdowns, plus the DPI. That's 146 DYAR on those eight plays alone. Flacco didn't throw a single pass in the red zone, mainly because he went right past it. However, the Broncos pinned him deep at the other end of the field quite often, but he was able to escape pretty easily. Inside his own 20, he went 4-of-6 for 55 yards, with every completion gaining a first down, plus that DPI.
4.
Colin Kaepernick SF
17/31
263
2
1
150
87
63
Kaepernick had four total touchdowns and a quarterback rushing record and still ranks fourth this week. That's partly because it was a REALLY good week for quarterbacks. Joe Flacco was the top-ranked quarterback of the wild card round, and he only had 129 DYAR that week against Indianapolis. Looking back over the regular season, when there are four times as many quarterbacks fighting for spots each week, 150 DYAR would usually get a quarterback in the top three. It would have been first overall in Week 2 and Week 10. Kaepernick also had an interception and a fumble, and while he was explosive, at times he struggled to keep drives going. On third downs as a passer, he went 5-of-7 for 90 yards, but only three of those plays went for first downs. He did much better running on third downs, gaining 75 yards and four first downs (including a 20-yard touchdown) on third downs. His only third-down run that did not pick up a first down gained 9 yards on third-and-10. And speaking of rushing numbers, Kaepernick finished with 14 carries for 183 yards and seven first downs, including two touchdowns. That's 63 rushing DYAR, which is the fourth-highest total for a quarterback since 1991. (We talked about this in Week 6 when Robert Griffin ran for 140 yards.) However, there's a bit of an asterisk with that. On second-and-9 in the middle of the second quarter, 49ers center Jonathan Goodwin snapped the ball early. It bounced off Kaepernick's knee, and he picked it up and scrambled for no gain. Looking back at the play, it's clear that the mistake is on Goodwin. The 49ers' receivers and linemen just stand around watching. Take that play away and Kaepernick had 78 DYAR rushing, and that would be second only to Michael Vick's 173-yard day against Minnesota in 2002 (82 DYAR).
5.
Aaron Rodgers GB
26/39
257
2
1
128
133
-6
Rodgers ranked above Colin Kaepernick in the ESPN version of Quick Reads, but that was due to a spreadsheet error that was giving him positive value (a lot of it, actually) for his interception. Further poking around revealed that the error involves "just as good as a punt" deep interceptions on third down, and about a half-dozen plays each season will need to be fixed in our data for quarterbacks when things slow down after the Super Bowl. (The error doesn't exist in the team numbers, thankfully.) Rodgers also padded his overall numbers with a meaningless fourth-quarter touchdown drive that still left Green Bay down by 14 points with a minute to go. On that drive, he went 8-of-10 for 63 yards, plus a 12-yard DPI, for a touchdown and four other first downs. That's 72 DYAR that did nothing to help Green Bay win the game.
6.
Matt Ryan ATL
24/35
250
3
2
113
110
3
Ryan had significantly more plays in Seattle's half of the field than in his own, and he played better there too. On his side of the 50, he went 10-of-14 for 94 yards with seven first downs and an interception. Across the 50, he went 14-of-21 for 156 yards and nine first downs, including three touchdowns, with another interception.
7.
Peyton Manning DEN
28/43
290
3
2
-9
-9
0
For 17 weeks, Manning was one of the best two or three quarterbacks in the NFL. In the eighteenth week, he wasn't, going 28-of-43 for 290 yards with three touchdowns and two interceptions, including a pick-six. He was also sacked three times, fumbling twice, with Baltimore recovering both balls. He struggled particularly badly on third downs, going 4-of-7 for 56 yards, and while each of those completions produced a first down (including a 15-yard touchdown to Brandon Stokley), he also gave up an interception (the pick-six), two sacks and a fumble.
8.
Matt Schaub HOU
34/51
343
2
1
-15
-11
-3
Schaub spent a lot of time trying and failing to throw short passes over the middle, going 11-of-18 for 98 yards with only four first downs and one interception to that area of the field. Meanwhile, on deep passes, he went 4-of-6 for 87 yards and a touchdown. Schaub failed to convert his first six third-down plays, and didn't pick up his first conversion until Houston was down by 11 points in the third quarter. His next third-down pass afte that was intercepted. Popular belief during Houston's late-season slide was that Schaub's play-action skills made him effective when playing with a lead, but he was so deficient otherwise that he was useless in a comeback situation. So let's check the numbers. During the regular season, Schaub's DVOA when ahead in the second half was 41.2%. Out of 25 quarterbacks with at least 50 passes in that situation, only Aaron Rodgers (49.1%) was better. When behind in the second half, though, his DVOA fell to -15.6%, 24th out of 38 qualifying passers. Sorry Texans fans. The public perception in this case is pretty accurate.
Five most valuable running backs
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
Shane Vereen NE
41
1
83
2
95
34
61
In 62 runs and 13 pass targets in the regular season, Vereen had two plays for 20 yards or more, and one of those came on a defensive pass interference foul. He had three 20-yard plays against Houston, and there were no referees involved. First came a 25-yard catch in the first quarter, then a 22-yard run in the second, followed by a 33-yard catch for a touchdown in the fourth. His final numbers: Seven carries for 41 yards, one touchdown, and two other first downs, with five catches in six targets for 83 yards, two touchdowns, and one other first down.
2.
Frank Gore SF
119
1
48
0
63
41
23
Gore was stuffed just twice in 23 carries, with three runs for 10 yards or more. He finished with five first downs on the ground, including a 2-yard touchdown. The 49ers threw him two passes, both on third-and-long. One of those was a meaningless short gain, but the other turned into a 45-yarder.
3.
Arian Foster HOU
90
1
63
1
50
31
20
Foster was stuffed four times in 22 carries, with two runs of 10 yards or more. He had a touchdown and four other first downs, three of which came with 1 yard to go. Foster caught seven passes in nine targets for 63 yards, with one touchdown (also with 1 yard to go) and three gains of 10 yards or more. In the 16 regular season games, Foster caught 40 passes for 217 yards. In two playoff games, he caught 15 for 97.
4.
Stevan Ridley NE
82
1
13
0
48
42
6
Ridley was stuffed three times in 15 carries, with one touchdown and four other first downs, including two runs of 10 yards or more. He only caught one pass, but that one pass was a 13-yard gain on second-and-12.
5.
Michael Turner ATL
98
0
0
0
40
40
0
Each of Turner's 14 carries gained at least one yard, five of them gained first downs, and three of them gained at least 10 yards. While we're here, let's discuss Turner's teammate, Jacquizz Rodgers. Rodgers finished with 64 yards on only 10 carries, but 45 of those yards came on one play. Only one other time did he gain successful yardage, and on three second-down carries (with 3, 5, and 8 yards to go), he gained a total of 1 yard.
Least valuable running back
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
Jacob Hester DEN
11
0
7
0
-21
-22
2
None of Hester's eight carries gained more than 2 yards. He did pick up two first downs, but he failed on three other third- or fourth-down runs. He was also thrown two passes. One was caught for 7 yards on first-and-10, the other was incomplete on third-and-8. That's an average of 1.8 yards in ten plays, and the fact that Jacob Hester was involved in six third-down plays is probably a big reason why Denver's season has come to an end.
OTHER BACKS OF LITTLE VALUE: Ronnie Hillman, DEN (22 carries for 83 yards; three catches in four targets for 20 yards); Bernard Pierce, BAL (five carries for 14 yards); Ray Rice, BAL (30 carries for 121 yards, but only four first downs; no catches in two targets).
Five most valuable wide receivers and tight ends
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Michael Crabtree SF
9
11
119
13.2
2
69
The 49ers threw Crabtree five passes on second and third downs. All were complete, for a total of 71 yards. Two went for touchdowns, and two others went for first downs. The fifth was a 14-yard gain on third-and-15.
2.
Zach Miller SEA
8
9
142
17.8
1
64
Miller had at least 685 yards receiving in each of his last three seasons in Oakland, but gained 629 yards, total, in two seasons after signing with Seattle in free agency. He only went over 40 yards four times in those two years, but he gained 48 yards on four catches in six targets in the wild card win over Washington, and then he exploded for eight catches and 142 yards and a touchdown in the loss to Atlanta. He had one touchdown and five other first downs on Sunday, including four plays of 20 yards or more. Is this two-game surge a random numerical fluke, or a sign that Russell Wilson is developing and learning to use all his receivers? Guess we'll have to wait until September to find out.
3.
Golden Tate SEA
6
8
103
17.2
1
59
Each of Tate's receptions produced a first down or touchdown. On deep routes, he caught four out of five balls for 88 yards.
4.
Torrey Smith BAL
3
6
98
32.7
2
49
In addition to the numbers shown above, Smith picked up a 9-yard DPI call on third-and-4. His 59- and 32-yard touchdowns were worth 55 DYAR between them.
5.
James Jones GB
4
6
87
21.8
1
43
Jones did most of his damage when the game was still close. In three first-half passes, he had one incompletion, one 44-yard gain on third-and-5, and a 20-yard touchdown.
Least valuable wide receiver or tight end
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Owen Daniels HOU
9
15
81
9.0
0
-27
Remember our discussions of Matt Schaub's struggles with short middle passes? Eight of those were thrown to Daniels, and he caught five of them for 50 yards, 34 of those on two plays. Daniels only had four first downs on the day.
OTHER RECEIVERS OF LITTLE VALUE: Doug Baldwin, SEA (one catch in three targets for 6 yards); Randall Cobb, GB (five catches in six targets for 24 yards with a fumble; two runs for 23 yards); Kevin Walter, HOU (two catches in two targets for 15 yards, both of them failed third-down plays, one of them fumble).

Comments

101 comments, Last at 18 Jan 2013, 5:37pm

1 Re: Divisional Round Quick Reads

"Ridley might have been the MVP of New England's playoff win over Houston, catching five passes in six targets for 83 yards and two touchdowns."

It was actually Shane Vereen.

3 Re: Divisional Round Quick Reads

I just posted this in the DVOA thread, but it belongs better here...

"This is why stats in a single game need to be taken with a very large grain of salt. The sample size is simply too small. Over 500-600 attempts (a season) the issues of context get balanced out to some degree, and over the 5000-6000 attempts (a career), the issues of context usually get balanced out to a very, very, large degree, although sometimes this is not true, with regard to the quality of teammates. In the sample size contained in a game? It is a really large mistake to simply look at the stats, even advanced stats, and make strong conclusions regarding the quality of play of an individual. There is no substitute for breaking down film, if you want to really know how well, or how poorly, somebody, even somebody at the position that best lends it self to statistical analysis, played in a particular game."

I haven't done the film work, but my guess is that Flacco's numbers overestimate his performance the most, because they can't take into context the hideous nature of some of the play of the Denver dbs; that last td in particular was little better than a Hail Mary completion. I think Russell Wilson's play may actually be better than his numbers. Everybody else may be about right, I'd guess, although Manning's 1st pick tends to overstate how badly he played; 3 fumbles are 3 fumbles, after all, and I don't think any came on a particularly fast sack when he couldn't prepare for contact.

10 Re: Divisional Round Quick Reads

Flacco takes advantage of bad matchups and defensive mistakes, so his play was overrated. Atlanta and Houston, meanwhile, never had a defensive breakdown or matchup problem. Just making sure I'm clear here.

11 Re: Divisional Round Quick Reads

Yeah, I respect Will and usually agree with him, but I'm kind of scratching my head at that one. A QB really shouldn't get penalized for taking advantage of a mismatch or broken coverage.

13 Re: Divisional Round Quick Reads

I'm not saying anything more than I would if the stats credit Joe Webb for a long td on a busted coverage, which happened the week before. It's not a bad play, but the yardage contained in the play makes up a pretty good percenatge of the day's positive work, and it really isn't due to any thing above average, in terms of qb performance. Over the course of a season, these things even out, but within one game, such a play, positive or negative, tends to overstate things.

15 Re: Divisional Round Quick Reads

Like I said, I haven't done the film work, so I was just giving my impression. If you want to argue the opposite of my point, which was that sample size contained within one game does not lend itself to accurate statistical representation of a qb's quality within one game, you go right ahead.

29 Re: Divisional Round Quick Reads

If you want to argue the opposite of my point, which was that sample size contained within one game does not lend itself to accurate statistical representation of a qb's quality within one game, you go right ahead.

What else besides the sample size contained within one game can represent a qb's quality within that game? These are the plays Flacco made within that game. Including the entire body of work, he did this well. The statistical anomaly of hitting three TDs on eleven deep throws does not belie the fact that Flacco did, in that game, throw those TDs, and they must be counted as part of his quality in that game.

As far as it being nothing out of the average, I'm not sure an average QB can even throw as far a Flacco did on some of those, let alone put the ball where a receiver could get it (or the DPI call), busted coverage or no.

Besides, Flacco only needs two more small-sample-size successes to win a big, expensive ring. For which every pundit will credit Ray Lewis and his intangibles. ;-)

43 Re: Divisional Round Quick Reads

Are you still a paying member of the Trent Dilfer Fan Club?

If I wasn't clear enough, I'll restate. The statistics provided in a sample size of one game do not give insight as to the context which provided the statistics. Only breaking down film can do that. Over a season, and even moreso, over a career, identifying the context which provided the stats is not very important, because with a large sample size, issues of context are evened out. This is not the case within the sample size that one game provides.

A good chunk of Flacco's positive yardage came on one play in particular, where the safety made an extraordinarily unathletic play; the analogy in baseball would be a centerfielder who, standing under a long, high, fly ball for two seconds, let the ball hit him in the forehead, after which the ball bounced into the stands, allowing the hitter to circle the bases. Now, you could, taking a superficial stats approach to such an event, simply say, "The hitter circled the bases; he had a great at-bat, and that at-bat contributed significantly to a great game", or, in contrast, could say, "The centerfielder made a bonehead play, allowed a long fly to bounce off his head into the stands, so we really shouldn't credit the hitter with circling the bases after that at bat, when evaluatng the quality of the game the hitter had". Baseball, being a less complex game, makes such statistical delineations easier, by calling the event a 4 base error, and not a homerun. Football has no such mechanism in collecting stats, even the advanced stats, and thus the stats we collect in fotball are more suspect in terms of measuring true quality of performance, especially when the sample size is small.

Hope this clears things up.

46 Re: Divisional Round Quick Reads

A good chunk of Flacco's positive yardage came on ne play in particular, where the safety made an extraordinarily unathletic play; the analogy in baseball woud be a centerfielder who, standing under a long, high, fly ball for two seconds, let the ball hit him in the forehead, after which the ball bounced into the stands, allowing the hitter to circle the bases.

No.

The safety (Raheem Moore) made an extraordinarily bad decision, creeping up to cover a shorter route and letting the receiver running deepest get behind him. The analogy in baseball is the centerfielder who breaks toward the infield on a fly ball, then can't change direction and reach the fence in time to catch the ball, which was barely over the fence. Which, in baseball, is still scored as a home run, not an error.

The rest of your point (one play, low percentage, not likely repeatable) is reasonable. But Flacco still had to throw the ball over 50 yards in the air to a moving target, while moving himself. That's a pretty valuable thing for an NFL QB to do. If nothing else, film of that play should be extremely valuable to Ray Rice's rushing average.

49 Re: Divisional Round Quick Reads

No.

The safety also mistimed his jump by a gigantic amount. It was extremely unathletic. If you want to make the point that since he actually never touched the ball, the baseball equivalent would have credited a homerun, fine, but then we just back to the original point, which was that taking a purely statistical apprach to evaluating the quality of play within one game will frequently lead to to large missjudgements as to what the qualty of play actually was.

82 Re: Divisional Round Quick Reads

From the replays I've seen, it looks an awful lot like the safety misread how long the pass would hang/how far it would travel, took a bad angle, then jumped at the correct time (from the wrong place). If he had angled back and towards the sideline, rather than directly to the sideline (and thus ahead of Jones), his jump would've actually been a good play. As it was, it ended up being a desperate attempt to make up ground he'd lost due to a poor decision.

85 Re: Divisional Round Quick Reads

He arrived in time. He jumped too soon, and thus had descended substantially from his highest point, when the ball arrived, which resulted in him just barely missing it. If he had timed his jump in just an average fashion, he easily gets his hands on the ball. I have no doubt that the errors compounded on each other, as the poor decisions cascaded throughout an anxiety-filled sequence. In other words, it is hard to play NFL safety with both hands around your throat.

47 Re: Divisional Round Quick Reads

I think there are two different points here.

You are saying first that single-game stats shouldn't be taken too seriously. I think most of us would agree with that.

Your second point is that a good deal of Flacco's success comes down to a play when the key defender made a remarkably boneheaded mistake. And then you relate this issue to baseball, where official scorers note errors.

But errors are not really efficiently assigned. I think I recall seeing a ball bounce of Jose Canseco's head for a home run - no error was assigned. But yeah, on the whole, the error statistics gives us a bit more accuracy to baseball stats.

How would this relate to football? I think it would be very hard to try to tweak official scoring to allow for boneheaded plays like Moore's. But the argument we could make is that on this play, at the very least, it's not really appropriate for opponent adjustments to give the Ravens a boost. They should get a boost when Torrey Smith is able to burn Champ Bailey, an elite CB who hadn't been exposed like that all season long. But the mental mistake Moore made was rare and bizarre.

I think, ultimately, these kinds of errors serve as inevitable noise to the system. Maybe I'm too lazy, or maybe I just don't know enough about advanced stats (actually, there's no 'maybe' to either of those two propositions), but I think we're better off just making the note than to try to tweak the system to account for aberrational play.

50 Re: Divisional Round Quick Reads

Oh, I wouldn't support trying to use a scoring system in football which modeled baseball's. Baseball's system is rife with problems as it is, as you note, and football is far more complex. It just wouldn't work. I think advanced stats which takes into account charting data, like dropped ints, have value, but regualrly assigning errors to defenders, which are then used to adjust offensive measurement, on a play by play basis? No, that just isn't feasable. Overall opponent adjustements are about the best that is possible.

My first point about the limited value of single game stats was my major one, and actually I started thinking about it a lot last week, when one play greatly masked, in the statistical sense, the true nature of Joe Webb's game against the Packers. The example of Flacco's last long completion is obviously not nearly as extreme. Unfortunately, it seems to have been interpreted by some as me being the Grand Wizard of the Anti-Flaccites.

77 Re: Divisional Round Quick Reads

I agree completely with Will's point and am somewhat (though not entirely) surprised by the pushback on it.

Of course it is better to win any particular game than to lose it. Of course it is better to have more passing yards and passing TDs in any game than to have less.

But if one wants to draw larger conclusions about value and make more accurate predictions about the future, one should probably note the small sample size of any particular game and particularly note those plays that are outliers and how they affect the stats that are gathered and computed. Yes, certainly, over the course of a season, a 16-game starting QB probably gets one or two long TDs from broken coverages. It evens out to some extent. But in any particular game, the odds of such a play (or multiple instances of such a play) are extremely low, and the results on overall performance are outsized, to say the least.

It hardly seems shocking to evaluate that outlier 70-yard TD and come to the subjective conclusion that the play was much more about poor DB play than good QB play. To me, it certainly seems the more reasonable subjective interpretation of reality, and it diminishes what Flacco achieved in that game with respect to a subjective determination of value.

What this has to do with being anti-Flacco, I have no idea. Flacco played well. He's a decent starter in the NFL. The outlier that was mostly caused by poor defensive play tends to over-value his performance in this one game. Why Will’s description of this is offensive, again, I have no idea.

83 Re: Divisional Round Quick Reads

I also don't read Will as saying anything controversial. If a stat sheet reads that a receiver was thrown three passes and caught all three of them for 50-yard touchdowns, then obviously it looks like he had a great game, and by any reasonable standard he did have a great game. But, if you want to know if that great performance predicts future great performances, you need to look at the film. If you look at the film and you see that on every one of those catches the defender slipped on the turf and fell down leaving the receiver open by 15 yards, and you also see that whenever the defender stayed on his feet the receiver was blanketed, that obviously has some impact on whether you should expect the receiver to have more big games in the future. Isn't that all that Will's saying?

84 Re: Divisional Round Quick Reads

Pretty much. I certainly wasn't saying that Flacco made a bad throw. It was desperation time, and he pretty much had to heave it. However, if future success is based upon the safety making a play like that, then there isn't going to be as much future success, as when the future is based upon a near-perfectly thrown ball that beats excellent coverage. Over a large enough sample size, these things tend to wash out, but one game doesn't provide a large enough sample size. It seems as if discussing qb performance gets a little sensitive sometimes.

4 Re: Divisional Round Quick Reads

Quick note: you guys meant to say Shane Vereen when disussing the Patriots running backs in the intro, not Stevan Ridley. He is a player most outside of the Patriots fan base knew much about prior to last weekends game, but accuracy demands he get his due.

5 Re: Divisional Round Quick Reads

While I thought Wilson played brilliantly, but I'm a little surprised to see him score *that* highly. But why shouldn't DVOA and DYAR love him? Everyone else does. There'll probably be a ballot initiative to rename Seattle Russleville, and secure a bond to build a 98 foot bronze statue of him in Elliot Bay, come November.

59 Re: Divisional Round Quick Reads

You'll notice it's 98 feet and not 100. Besides, we may need to leave room for an asterisk shaped crown of glory (it'll hold a laser light show!) later.

I don't know if I've ever seen this town fall in love with any public figure so quickly.

75 Re: Divisional Round Quick Reads

I remember Ichiro as taking longer perhaps because of his private nature. But I never really followed the Mariners after they traded everyone I liked within the span of two years as a kid. I was thinking either Griffy or Largent, I was a little young during the start of Largent's reign, but it just feels like it took more than a season for them.

Wilson pretty much had the city eating out of his palm by week 6. Now, he couldn't be more adored if he was the cutest kitten on the internet.

94 Re: Divisional Round Quick Reads

Thanks. My question mark was intentional, as I don't follow the Mariners much at all, except when they're playing the Red Sox. By the time Ichiro-mania reached the Northeast, it was already in full blossom in the PNW, so I had no idea how long the germination process had taken.

101 Re: Divisional Round Quick Reads

The "whoa, Ichiro!" moments began in the first week of the 2001 season (his ROY and MVP year, and the year the Mariners won 116 games). (April 11 in the bottom of the 8th, to be exact, in Oakland.)

There's multiple copies of the video around the net, but may as well go to MLB for it: http://mlb.mlb.com/video/play.jsp?content_id=16865121

Couple that with an HR-saving catch or two, beating out a couple routine grounders to short, and the incredibly fast start the team enjoyed (20-4, then 47-12), and it was pretty much a love affair from that moment on.

27 Re: Divisional Round Quick Reads

I follow Wilson because I follow Wisconsin sports, and Ken Pomeroy's stats, college basketball's equivalent of DVOA in terms of how much respect is paid to it, simply loves Wisconsin, mainly because they can blow out bad teams by a much bigger margin. For that matter, FO's own FEI Ratings love Wisconsin as well for the last two years, and again, I think it's because they can blow out bad teams by a much bigger margin.

30 Re: Divisional Round Quick Reads

Relative to RG3, both Wilson and Luck have been totally dissed by the media.
Maybe that will change in Wilson's case now, but I kind of hope not. There's such
a thing as getting too much attention.

7 Re: Divisional Round Quick Reads

"Those touchdowns came on gains of 8, 5, and 3 yards, and all came on first down."

Uh...I seem to recall a long TD bomb to Shane Vereen.

Did you guys just block out Shane Vereen from your memory?

8 Re: Divisional Round Quick Reads

Ridley might have been the MVP of New England's playoff win over Houston, catching five passes in six targets for 83 yards and two touchdowns.

I believe you mean Vereen here?

12 Re: Divisional Round Quick Reads

BITTER BRONCO FAN.....

FLACCO DYAR 100... DENVER DEFENSE -100
The Pick Six against Manning was an uncalled PI.
The two PI calls Flacco got were questionable. It was good defense by Carter against an uncatchable ball; 2nd against Bailey was ticky-tack, especially when considering the no-call on the Pick Six.
Flacco was also the benefactor the of a Bill Buckner level error by Rahim Moore.

CAN'T PUT BAILEY ON AN ISLAND ANYMORE
Smith and Flacco made Bailey look old. Bailey might not need to go to Safety just yet, but his days of being on an island against a teams best receiver are over.

FOX PLAYED THE END OF THE GAME LIKE HIS QB WAS NAMED DELHOMME OR TEBOW
Taking a knee with 31 seconds and two timeouts was awful.... Could they have at least tried a safe TE pass or screen?

68 Re: Divisional Round Quick Reads

I felt like neither Moore nor Mike Adams were playing well on Saturday, especially compared to some of the regular season games I saw. Maybe Baltimore managed to scheme them out of the picture, but whatever coverage the Denver secondary was in kept them from helping effectively over the top on a number of big pass plays, and Adams didn't do anything against the run, either.

17 Re: Divisional Round Quick Reads

In baseball, runs and hits that result from defensive errors are accounted for differently statistically. Denver's safety play at the end would be a the ultimate definition of an error.

20 Re: Divisional Round Quick Reads

Yeah, baseball just has a much more sophisticated statistical approach, in terms of measuring games, seasons, and careers, in large part due the game being less complex, and easier to measure. Even so, baseball stats usage semtimes will fall into the same trap; hitters will be said to be in a slump, even though they hit the ball on the nose 10 times in their last 30 plate appearances, and had two base on balls on top of it, because they were very unlucky, and had those 10 balls go right at someone.

72 Re: Divisional Round Quick Reads

Good point. They could use chess notation to denote when someone made a particularly good defensive play. Something like 6!-4-3 for a great snag by the shortstop leading to a double play.