≡ Menu

Least-Conforming Games of 2015

The Buccaneers were not very good last year. Tampa Bay finished with the worst SRS in the NFC, and the second-worst in the NFL ahead of only Tennessee. But that doesn’t mean the Bucs season was predictable; in fact, Tampa Bay had arguably the two weirdest games of the year.

The Bucs opened the season with the least-conforming game of the first half of the season: Tampa lost, at home, to Tennessee, by 28 points! That’s incredible: the Titans only other two wins were by 3 points against Jacksonville and in overtime against the Saints.

But, amazingly, that wasn’t even the least-conforming game of the Bucs season. In week 11, in Philadelphia, Tampa Bay beat the Eagles 45-17. The same team losing at home by 28 points to Tennessee and winning by 28 points on the road in Philadelphia? That’s pretty weird.

The table below shows all 512 regular season games from 2015, and how it differed from expectations.  Here’s how to read the first line. The biggest outlier game was Tampa Bay against Philadelphia, which came in week 11.  You can click the Boxscore link to go to that game’s boxscore on PFR.  Tampa Bay had an SRS rating of -7.7, while Philadelphia’s rating was -4.7.  As a result, given that the game was in Philadelphia, the Expected Margin of Victory for Tampa Bay was -6.0.  In reality, Tampa Bay scored 45 points and allowed 17, for a 28-point Margin of Victory. That exceeded expectations by 34.0 points. [click to continue…]

{ 7 comments }

Last year, Denver had a pretty tough schedule. In only five games did they face an opponent that an average team would have been favored to win:1 home games against San Diego, Baltimore, and Oakland, and road games against the Colts and Browns. In those games, Denver went just 3-2, with all five games being decided by one score.

The Broncos had six games against top-8 teams by the SRS: two games against the Chiefs, and games against Cincinnati, Minnesota, New England, and Pittsburgh. In those games, Denver went even better at 4-2, with five of those games being decided by one score.

The middle five games of the schedule by SRS standards was where the Broncos really dominated: the Broncos went 5-0 in road games against Oakland, San Diego, Chicago, and Detroit, and a home game against Green Bay, with three of those five wins coming by double digits.

As it turns out, Denver had the third “strangest” season in the NFL last year. How did I define strange? I measured the correlation coefficient between two variables: the actual margin of victory in a game, and the opponent’s SOS (after adjusting for home field advantage). The Broncos had a CC of 0.18, which means (in case you couldn’t figure it out above) that there wasn’t a big relationship between results and expectation. [click to continue…]

  1. I.e., home games against teams with SRS ratings of 3.0 or worse, or road games against teams with SRS ratings of -3.0 or worse. []
{ 6 comments }

Joe Gibbs Inherited a Very Underachieving Team

theismannYesterday, we looked at the teams that overachieved their projected wins total by the largest amount based on the strength of their offensive and defensive passing games. Today, the reverse: the biggest underachievers. And that starts in Washington, D.C., the year before Joe Gibbs arrived.

The head coach was Jack Pardee, who was in Washington for three years, going 8-8 in 1978, then 10-6, and then 6-10 in 1980. Pardee was fired after the season, and you can see why: Washington didn’t just have a good pass defense in 1980, but a great one. It ranked as the 12th best pass defense from 1950 to 2013. Both corners, Lemar Parrish and Joe Lavender, made the Pro Bowl. Both safeties, Mark Murphy and Tony Peters, were in the primes of their careers, and would make a Pro Bowl under Gibbs.

Washington had an absurd 8.4% interception rate and a 9.9% sack rate, which helped the defense allow just 2.4 ANY/A, nearly a full yard better than any other team and 2.51 ANY/A better than league average. And the team went 6-10! The offense had Joe Theismann, Art Monk, and Wilbur Jackson; Theismann ranked 17th out of 30 qualifying passers in ANY/A, but that shouldn’t have been enough to keep the team out of the playoffs, let alone below .500.

Washington’s offense finished with a Relative ANY/A of -0.33, and its defense had a RANY/A of +2.51. The team had a 0.375 winning percentage, but “should” have had a 0.704 winning percentage. That means the team underperformed by 5.3 wins, the most of any team in the Super Bowl era. [click to continue…]

{ 2 comments }

The 1968 Cardinals and Outlier Teams

Hart had a great career, but was still developing in '68

Hart had a great career, but was still developing in ’68

In 1968, the St. Louis Cardinals did not have a very good passing offense. The Cardinals averaged 3.9 ANY/A, good enough for 11th place in a 17-team league where the league average was 4.5. The main issue? St. Louis finished dead last with an anemic 44% completion rate. That was mostly due to the second-year starter, 24-year-old Jim Hart. His 44.3% completion rate remains the lowest by any Cardinals quarterback in history with a minimum of 300 pass attempts, and no quarterback with even 160 pass attempts for the Cardinals has dipped below 45% since Hart in ’68. The defense was also below-average against the pass: the Cardinals allowed 6.2 ANY/A, 5th worst in the NFL.

Teams that are below-average at passing and stopping the pass are usually not very good. In the Super Bowl era, each additional yard of ANY/A (on either offense or defense) relative to league average increases a team’s winning percentage by about nine percent. The Cardinals, at -0.5 Relative ANY/A on offense and -1.7 RANY/A on defense, would therefore be expected to win about 30% of their games. Instead, the Cards won 68% of games, going 9-4-1.

The table below shows the 100 teams of the Super Bowl era that have most exceeded expectations based on Offensive and Defensive RANY/A. In general, using ANY/A and RANY/A will get you most of the way there with a team’s record, but not for these teams. The Cardinals had an Off RANY/A of -0.52, a Def RANY/A of -1.72, and therefore, an expected winning percentage of 0.296. By having an actual winning percentage of 0.679, the Cardinals exceeded expectations by 6.1 wins per 16 games. [click to continue…]

{ 10 comments }

538: Front- and Back-loaded Schedules

Today at 538, a look at which teams have front-loaded (the Jets) and back-loaded (the Ravens) schedules. The methodology will be familiar to regular readers: I created implied NFL ratings based on Vegas point spreads, and then calculated general and then weighted strength of schedule ratings. The weight, of course, was based on how late in the season a particular game occurred.

You can read the article here.

{ 3 comments }

On average, passing yards is a pretty meaningless measure of quarterback play.  Consider that the winning team and the losing team in a game both generally throw for about the same number of yards. Last year, for example, winning teams averaged 258 gross passing yards per game, while losing teams averaged 259. In 2013, it was 253 for the winners, 251 for the losers. In 2012, it was 246 for the winners, 248 for the losers. Since 2000, winning teams have averaged about 5 more passing yards per game, thanks mostly to 2009 (244 for winning teams, 222 for losing) and 2014 (261/242) as big outliers.

Joe Flacco, for example, has averaged 233 passing yards per game in wins and 231 in losses. But just because the averages are close together doesn’t mean every quarterback follows this same formula. And two of the best examples of that are Nick Foles and Blake Bortles.

Foles has lost 17 games where he was the starting quarterback; in those games, his average stat line was 21/38 for 214 passing yards, 0.7 TDs and 1.1 INTs. He also has started and won 19 games; in those games, his average stat line was 19/30, for 258 passing yards, 2.1 TDs, and 0.4 INTs. That paints the picture of a guy who is much better in wins than losses, which makes a lot of sense.  (Also, 7 of his 17 losses have come during his ugly time with the Rams, compared to just 4 of 19 wins.) [click to continue…]

{ 5 comments }

Guest Post: Bryan Frye on Adjusted Drive Yards

Friend of the program Bryan Frye is back for another guest post. As regular readers know, Bryan operates his own fantastic site, http://www.thegridfe.com. You can view all of Bryan’s guest posts here, and follow him on twitter @LaverneusDingle.


For some time, I have wanted to create a new metric that used elements from Total Adjusted Yards (TAY) in order to quantify a team’s production on each drive. Past work from both Chase and Brian Burke has given us insight into the value of touchdowns, interceptions, fumbles, and first downs, translated into yards. This work has been fundamental in the development of stats like Adjusted Net Yards per Attempt, Adjusted Rushing YardsAdjusted Catch Yards, and TAY.

Those metrics have given us valuable insight regarding statistical measurement of individual player performance. I’ve also used TAY to measure the output of offenses and defenses.

However, I wanted to attach generic values to every way a drive can end.1 This is not a rigorous study, and it is meant to be a starting point for future research rather than a conclusive formula to govern the way anyone interprets on-field action.

With that in mind, I’ll briefly cover the generic yardage values for various drive outcomes. [click to continue…]

  1. With the exception of kneel down drives to end halves or games, as those don’t demonstrate an offense’s (or defense’s) ability to actually play the game. []
{ 44 comments }

Adjusted Completion Percentage

In 1991, Dave Krieg led the NFL in completion percentage. He completed a career-high 65.6% of his passes, and while that mark was very good for that era, it doesn’t mean Krieg was great that season. In fact, he arguably wasn’t even good: Krieg actually finished just 24th in ANY/A that year.

One reason, I think, that Krieg was able to lead the NFL in completion percentage is because Krieg “ate” a lot of his incomplete passes. What do I mean by that? Krieg took a ton of sacks — he was sacked every ten times he dropped back to pass. When under duress, some quarterbacks eat the ball, to avoid an interception; that’s bad (well, it’s better than n interception) but it doesn’t get graded that way when calculating completion percentage. Other quarterbacks will throw the ball away; that’s good (assuming it isn’t intercepted) because no yards are lost, but it does hurt the quarterback’s completion percentage.

Even ignoring the yards lost due to sacks, fundamentally, a sack is no better than an incomplete pass. So why are quarterbacks who take sacks rather than throw the ball out of bounds given an artificial boost when it comes to completion percentage? Well, that’s largely just an artifact of how the NFL always graded things. The NFL was not always good at recording metrics, and somewhere along the way, sacks were either included as running plays, ignored, or included as pass plays. I don’t think a lot of thought went into it, but in my view, it makes the most sense to include sacks in the denominator when calculating completion percentage. Otherwise, we give undue credit to quarterbacks that take a lot of sacks, and penalize quarterbacks who throw the ball away when under pressure. [click to continue…]

{ 34 comments }

The 1978 Patriots, Part II

The 2001 Rams had Kurt Warner, Marshall Faulk, Isaac Bruce, and Torry Holt.

The ’92 and ’93 49ers have prime Steve Young and prime Jerry Rice, along with the first two years of Ricky Watters’ great career.

The ’88 Bengals had MVP Boomer Esiason, Pro Bowler Eddie Brown, HOFer Anthony Munoz and Pro Bowler Max Montoya on the offensive line, and a running back tandem of James Brooks and Ickey Woods. Two years earlier, the ’86 Bengals had those players save Woods, but also had Cris Collinsworth in the prime of his career.

The ’51 Rams had Norm Van Brocklin and Bob Waterfield — two HOFers — at quarterback, along with Elroy Hirsch, Dan Towler, Dick Hoerner, and Tom Fears.

Those are 6 of the 7 teams since 1950 to lead the NFL in both average yards per rush and average yards per pass. Can you guess the 7th team? You have three guesses, but the first two don’t count. [click to continue…]

{ 15 comments }

The 1978 Patriots, Part I

Here’s what I wrote in my first post at Football Perspective:

I’ll be blogging about everything football-related, from Jerry Rice to Bobby Douglass, and from the 1978 Patriots to who is the greatest quarterback of all time.

The New England Patriots rushed for 3,165 yards, an NFL record that still stands. Take a look at the individual players on that team:

Games Rushing
No. Age Pos G GS Att Yds ▾ TD Lng Y/A Y/G A/G Fmb
39 Sam Cunningham* 28 FB 16 14 199 768 8 52 3.9 48.0 12.4 4
23 Horace Ivory 24 rb 15 3 141 693 11 28 4.9 46.2 9.4 5
32 Andy Johnson 26 RB 15 13 147 675 3 52 4.6 45.0 9.8 4
14 Steve Grogan 25 QB 16 16 81 539 5 31 6.7 33.7 5.1 9
44 Don Calhoun 26 rb 14 2 76 391 1 73 5.1 27.9 5.4 1
37 James McAlister 27 16 0 19 77 2 16 4.1 4.8 1.2 3
86 Stanley Morgan 23 PR/WR 16 16 2 11 0 6 5.5 0.7 0.1 6
29 Harold Jackson 32 WR 16 13 1 7 0 7 7.0 0.4 0.1 0
30 Mosi Tatupu 23 16 0 3 6 0 3 2.0 0.4 0.2 0
4 Jerrel Wilson 37 P 14 0 1 0 0 0 0.0 0.0 0.1 1
83 Don Westbrook 25 16 0 1 -2 0 -2 -2.0 -0.1 0.1 0
Team Total 26.2 16 671 3165 30 73 4.7 197.8 41.9 35

[click to continue…]

{ 6 comments }

He gained 120+ yards pretty frequently

He gained 120+ yards pretty frequently

Yesterday, I posted a list of the career leaders in receiving yards after removing “junk” yards gained on an individual game basis. I’ve defined junk games as somewhere between 32 and 40 yards in 2015, and a lower threshold in less pass-friendly eras. You can view the Justin Blackmon example here.

While I presented the career list yesterday, I thought it would make sense to plot the career yards after removing junk yards (using 2.5x as the baseline) against each receiver’s plain career receiving yards (in both cases, since 1960). That’s what I’ve done in the graph below, with actual career receiving yards on the X-Axis and career yards after removing junk yards on the Y-Axis. Jerry Rice is literally off the chart (22,895; 13,786) because including him would require using a much broader (and less helpful) chart. Let’s just ignore Rice and focus on the other 99 receivers: [click to continue…]

{ 10 comments }

[Note: Due to a scheduling blunder, you may have missed yesterday’s post on single-season leaders.]

The GOAT

The GOAT

On Sunday, I explained one methodology to modify receiving yards in a way to give more value to top receivers while devaluing junk games. You can read that explanation here, and see the Justin Blackmon example.

Jerry Rice, of course, will rank as the top receiver by this or any other methodology, especially if that system excludes Don Hutson (today’s data only goes back to 1960). In fact, using a 3X baseline, Rice still gained 15,314 receiving yards after removing junk yards, more than every wide receiver in NFL history has gained including junk yards other than Terrell Owens. Rice was just incredible.

Perhaps the first real surprise on the list is Don Maynard, who ranks 6th among all players since 1960 by this methodology. The Jets Hall of Famer currently ranks 26th in career receiving yards, but 30 years ago, he was the all-time leader in that category. Maynard benefits here for some era adjustments — his 14-game seasons get prorated, the baseline for junk seasons was lower in the ’60s and ’70s — and his dominant play for a long stretch is rewarded.

The table below shows the top players by this methodology since 1960. Here’s how to read the table, using the Owens line. Using a 2.5X baseline, he ranks 2nd all-time. His career began in 1996 and ended in 2010, and he had 9,386 receiving yards above that junk baseline. Using a 3X baseline, he still ranks 2nd, and had 10,493 non-junk receiving yards. [click to continue…]

{ 11 comments }

Adam Steele is back for another guest post. You can view all of Adam’s posts here. As always, we thank him for contributing.


There have been countless attempts at deducing the clutchiness of NFL quarterbacks, most of which involve tallying playoff wins and Super Bowl rings. Today I’m going to take a stab at the clutch conundrum using a different approach: Pythagorean win projection. If a quarterback’s actual win/loss record diverges significantly from his Pythagorean estimated record, perhaps we can learn something from it. I began this study having no idea how it would turn out, so there were definitely some surprises once I saw the end results. This study evaluates the 219 quarterbacks who started at least 32 games since 1950, including playoffs but excluding the 1960-64 AFL (lack of competitive depth).

Here’s how to read the table, from left to right: points per game scored by the QB’s team in games he started, points per game allowed in his starts, total starts, total wins (counting ties as a half win), Pythagorean projected wins based on the points scored and allowed in his starts (using a 2.37 exponent), and the difference between his actual win total and Pythagorean win projection. [click to continue…]

{ 39 comments }

Brad Oremland noted in his last post that Stanley Morgan is the only player in history to average more than 19 yards per catch in a career with at least 500 receptions, and that such distinction will probably stand forever. Brad’s likely right: given today’s environment, Vincent Jackson and Calvin Johnson are the two preeminent deep threats of the last decade with at least 500 catches, and Jackson (16.97) and Johnson (15.89) were far shy of that mark.

That’s a fun bit of trivia, but let’s expand it. You can use reception cut-offs to come up with lots of Yards per Catch Kings. Here’s an exhaustive one:

  • Jerry Rice is the all-time leader in yards per reception (14.78) among players with at least 1,079 receptions.
  • Terrell Owens (14.7811 to Rice’s 14.7805) is the all-time leader in yards per reception among players with at least 1,025 receptions.
  • Isaac Bruce is the career leader in YPR, at 14.85, among players with at least 983 receptions.
  • Randy Moss (15.57) is the only player to average 15 yards per reception and record 820+ receptions.

[click to continue…]

{ 20 comments }

The 2015 season was another spectacular one for wide receivers. Pittsburgh’s Antonio Brown outgained the NFL’s leading rusher by a record 349 yards. On a game-by-game basis, the leading receiver for every team in every NFL game this year, including playoffs, averaged 94.3 receiving yards, a post-merger record.

In fact, the average number of receiving yards gained by the leading receiver of each team has been steadily rising, which isn’t surprising.  The average was below 80 as recently as 1992, and below 70 in 1977, the year before the big passing rules changes went into effect.  But the 1962 NFL season had a slightly higher average, at 95.2, while the average leading receiver in a game in the ’64 AFL even broke 100.

The graph below shows the average number of receiving yards gained by each team’s leading receiver in every game in each season since 1960.  In all graphs today, the NFL line is in blue, while the AFL line is in red. [click to continue…]

{ 4 comments }

The 2000 NFL Draft was supposed to bring an incredible infusion of wide receiver talent. Peter Warrick, Plaxico Burress, and Travis Taylor were top-10 picks, making it one of only four classes since 1970 were three wide receivers drafted in the top ten. In addition, Sylvester Morris, R. Jay Soward, Dennis Northcutt, and Todd Pinkston all went in the top 36 picks, one of only seven classes since the merger with seven wide receivers in the top 36. Avion Black was the 20th wide receiver taken with the 121st pick: add it all up, and the 2000 draft had unmatched levels of quality and quantity. The graph below shows the amount of draft value spent on wide receivers (you can click here for value spent on wide receivers and tight ends) in each draft from 1970 to 2011: [click to continue…]

{ 24 comments }

Over the last couple of days, I’ve been looking at receiving yards by class year. I’ll continue that today, with a look at the best classes in wide receiver history.

The 2014 class looks to be a very special one. It set a rookie record by gaining 18,321 receiving yards in 2014, the most by any set of rookies in NFL history. Then last year, those same players gained 23,727 last year, the most by any class in any single season in history.

Of course, while impressive, we have to remember the pass-friendly environment we are experiencing. The Class of 2014 — which includes all players selected in the 2014 Draft and all undrafted players whose first season began in 2014 — gained 14% of all receiving yards two years ago, and then 18% of all receiving yards in the NFL in 2015. Thought of another way, the class of 2014 has averaged 16% of receiving yards in their first two seasons.

thru 2 years

The 1987 class was a bit inflated by the replacement players who all register as rookies. The only other class since the merger with at least 15% through two years was the 1974 class, which got strong rookie seasons from Charlie Wade, Nat Moore, Paul Seal, Joel Parker, Harrison Davis, and Roger Carr, and then had Lynn Swann, Ken Payne, Moore, Ray Rhodes, Carr, Charlie Smith, and John Stallworth play well in 1975.

The 18% number produced by the 2014 class in year 2 was the highest rate since by a sophomore class since 1958.  That year, second-year players Del Shofner led the NFL in receiving yards, while R.C. Owens and Tommy McDonald finished in the top ten, with Joe Walton, Jon Arnett, and Billy Ray Barnes rounding out the class.

We can also look at the best classes as rookies, and over 2-, 3-, 5-, 7-, and 10-year periods. Finally, the last column simply sums the percentage of receiving yards from each class in every year of their careers.

Year1First 2First 3First 5First 7First 10Total
195018.5%20.7%20.4%19.3%17.3%14.5%147.9%
195118.3%17.9%16.6%16%14.7%11.9%123.2%
195215.8%16.9%15.9%16.8%16.6%14.5%156.2%
195311.8%10.3%10.4%11.4%10.5%8.7%97.5%
195416.1%13.4%14.3%12.8%11.9%9.7%105.1%
195512.9%10.7%9.1%7.3%6%4.4%43.6%
195610.9%13.3%14.5%14.3%13.8%11.6%118.9%
195711.4%16.2%16.7%16.2%16%13.7%144.3%
195810.4%13.2%15.8%15.6%15.3%13.5%137.7%
19596.4%8.2%8.5%9%8.1%7.2%75.5%
19608.6%10%10.1%9.4%8.3%6.7%70.8%
19618.6%12.1%12.7%13.2%12.5%10%102.1%
19625.8%6.9%7.6%8.1%7.6%6.6%67.2%
196310.5%9.8%10.9%11.9%11.6%8.8%90.2%
196413.1%13.3%14.6%14.4%13.3%10.7%112.2%
19658.7%11.8%13.8%15%15.1%13.1%136.7%
19663.9%7%8.4%8.8%8.4%6.6%67%
19678.7%11%12%11.3%10.3%8.2%84.6%
19688.4%9.9%11%10.9%9.8%8.2%89.9%
196911%13.9%14.8%15.1%13.2%10.5%114.6%
197010.7%12.5%13.7%13.7%12.3%9.8%101.3%
197111.5%13.2%14%13.4%11.9%9.7%103.4%
19727.9%10.1%10.7%10.9%9.9%8.1%85%
197310.9%13.9%14.1%13.2%11.2%8.8%90.6%
197413.7%15.2%16.7%17.1%15.7%12.3%129.8%
197511.5%12.3%12.7%12%10.5%8.6%89.4%
197612.7%14.1%15.2%15.6%14.5%12%124.6%
19779.3%11.8%12%12%10.9%9%94.1%
197810.8%12.1%11.8%11.8%11%9.4%100.2%
197910.7%13.8%15%15.9%14.5%11.9%127%
19809.6%9.9%10.8%11.4%9.9%7.7%81.5%
19818.4%9.8%10.8%11%9.8%7.9%80.4%
19828.1%10.4%10.8%10.9%9.8%8%85.6%
198311.7%13.4%14.4%14.2%13.4%11.4%120.3%
19849.8%11.7%11.3%11.6%10.7%8.8%97.4%
19859.6%13%13.3%13.5%13.1%11.6%130%
198612.1%12.7%12.5%12.7%12%10.3%106.4%
198716.5%16.2%15.9%14.4%12.4%9.7%103.2%
198811.1%12.4%12.8%13.9%13.5%11.6%123.4%
198910.9%11%10.7%10.3%9.5%8.3%87.4%
19909%11.1%12.2%12.7%11.6%10.1%110.3%
19917.2%10.4%11.3%12.9%12.8%11.4%123.5%
19925.5%6.6%7.6%7.8%7.6%6.2%66.1%
199310.3%10.5%11.2%10.9%10.3%8.6%90.3%
19948.4%10.4%11.5%11.3%10.8%9.2%98.6%
199510.8%12.3%12.8%12.3%11.5%9.4%100.3%
199610.6%11.4%12.4%13.3%12.9%12%137.1%
19976.5%8.4%9.4%9.8%9.1%7.8%88.8%
19989.5%11.6%11.8%11.2%9.9%8%86.3%
19998%9.5%10.1%10.6%9.6%8.2%87.5%
20009.2%10.3%11%10.5%9.4%7.4%75%
200110.3%12.3%13.7%13.9%12.9%10.9%118.3%
200211%11.6%12.7%12.7%11.5%9.1%91%
200310%11.5%11.9%12%11.5%9.8%107.2%
20049.5%11%12.5%12.4%11.1%8.8%92.2%
20058.6%9.3%9.9%9.6%8.8%7.3%74.9%
200610.2%11.7%12.3%12.1%11.5%9.7%97.4%
20079.3%11.1%12.4%12.2%11%86.6%
200810.2%12.3%12.8%12.6%11.5%85.4%
200911.3%13.7%13.5%12.4%10.9%76%
201012.1%13.7%14.2%13.7%77.4%
201111.7%12.6%13.1%12%60.1%
201211.8%13.4%12.9%48.7%
201313.4%14.3%14%42%
201414.2%16%32.1%
201512.6%12.6%

The 1996 class, with Marvin Harrison, Terrell Owens, Keyshawn Johnson, Muhsin Muhammad, Joe Horn, Eric Moulds, Amani Toomer, et. al., is often considered one of the best classes ever. That’s not quite so clear early on — a number of classes have them beat through 7 years — but the longevity is incredible.  Take a look at this graph, which just shows the total percentages; that’s obviously going to be biased against active classes, but it’s a fun graph to look at anyway:

overall wr perc

As always, please leave your thoughts in the comments.

{ 1 comment }

It’s easy to think that as the NFL becomes more of a passing league — a statement that’s undeniably true — that the best teams would be passing most frequently. But that just isn’t the case. The three best teams in Adjusted Net Yards per Attempt last year were Arizona, Cincinnati, and Seattle; those three teams ranked 19th, 26th, and 28th, respectively, in pass attempts. The Saints and Patriots did rank in the top five in both pass attempts and pass efficiency, but that just balances things out; it doesn’t mean the best passing teams are the most pass-happy teams.

There’s a pretty easy way to track this throughout history. The common way to calculate league-average Adjusted Net Yards per Attempt is to measure the league totals of its components: figure out how many league-wide passing yards, touchdowns, interceptions, sacks, and sack yards lost there were in any given season, and run through the calculation.

Another way, though, is to measure each team’s ANY/A average, and take an average of those averages. This approach gives each team the same weight when calculating league-average ANY/A; as a result, if this approach leads to a higher average than the traditional approach, that means the best passing teams are passing less frequently. And if the traditional approach has a higher average, that means the better passing teams are passing more often, because giving those teams extra weight (because of more pass attempts) is leading to a higher average. [click to continue…]

{ 11 comments }

You remember the 1987 Draft, right? It was a terrible draft for pass catchers.  The first TE drafted was Robert Awalt in the third round; only two more, Ron Hall and Jim Riggs, went before the sixth round, and Ron Embree was the final TE selected before the seventh round. At wide receiver, Haywood Jeffires was the first off the board at #20; the only other first rounders were Ricky Nattiel and Mark Ingram. The only other receiver in the top 50 was Lonzel Hill.  Mark Carrier, Kelvin Martin,Curtis Duncan, and Bruce Hill went in the later rounds,  but it was a terrible draft for pass catchers.

Using the Draft Value Chart, there were 177.4 points of draft value used on wide receivers and tight ends in the 1987 Draft.  That was the second year in a row when the league moved away from pass catchers.  Well, in this past draft, less draft capital was spent on wide receivers and tight ends than on any year since 1987. Take a look: [click to continue…]

{ 1 comment }

Consecutive Playoff Losses For a Franchise

From 1993 to 2015, the New York Islanders lost eight consecutive playoff series, beginning with a loss in the conference finals to Montreal in 1993, and culminating in a heartbreaking, 7-game series loss last year to Washington. Last night, the Isles came from behind and defeated Florida, to win the series, four games to two.

So the streak stopped at eight for the Islanders; as it turns out, the longest streaks for consecutive playoff losses in NFL history is also at eight, with two of those streaks being active. [click to continue…]

{ 6 comments }

Yesterday, I wrote how the NBA seemed to undervalue the three-point shot for many years. While the 3-point shot was consistently the better EV play, and the ratio of three-point shots to overall shots was increasing, it didn’t seem to increase quickly enough. As pointed out in the comments, one could make a pretty similar claim about pass/run ratio in the NFL.

It’s a little misleading to start things in 1970, since that’s really the beginning of the dead air era in football history. Pass efficiency was very high in the late ’40s and parts of the ’60s, so a chart beginning in 1970 would inaccurately imply a linear progression of the passing game. That said, because first down data is spotty the farther back we go, and because of the complexity involved in deciding how to treat the AFL, I’m going to limit myself today to the period from 1970 to 2016. [click to continue…]

{ 6 comments }

Yesterday, I looked at the Pythagenpat records for all teams since 2000. Since I crunched all that data, I thought it would be fun to look at the biggest outlier teams.

The 2003 Steelers were not very good. Pittsburgh went 6-10, scoring 200 points and allowing 327 points. Because of regression to the mean, the ’04 Steelers were expected to be a little better, and finish with 7.2 wins. Instead, behind a rookie Ben Roethlisberger and an outstanding running game and defense, the Steelers went 15-1, exceeding expectations by 7.8 wins.

Last year’s Panthers also went 15-1, and have a similar story. Cam Newton, the AP MVP, was more of the driving force, of course, but a great running game and defense powered the team. But based on a mediocre ’14 season, Carolina was expected to win only 7.8 games, so the 2015 Panthers exceeded expectations by 7.2 wins.

The third biggest outlier? That would be the ’07 Patriots, who went 16-0 with a projection of just 9.5 wins. The next year, New England was projected to win 10.99 games, and… went 11-5.

The table below shows each team since 2000, and their number of projected and actual wins. The table is sorted by the difference column: [click to continue…]

{ 2 comments }

The Browns have been running in place

The Browns have been running in place

A good read from ESPN yesterday about new Clevleand Chief Strategy Officer Paul DePodesta, who is being labeled as the man who will (attempt to) bring Moneyball philosophies to the NFL. Putting aside the inaccuracy of that statement — Moneyball philosophies mean different things to just about everyone, and such philosophies are already a staple in many organizations — there will be a certain spotlight cast on DePodesta in Cleveland. And, according to some statistical analysts, that’s a bad thing.

At MIT Sloan Sports Analytics Conference in March, unilateral fear existed inside analytics community that systemic ineptitude of Browns franchise will be too substantial for even DePodesta to repair. Failure would damage legacy of beloved industry pioneer and set field of sports data science back decades. “If you love analytics and want it to grow and succeed in the NFL, then you know Cleveland is a nightmare scenario,” states NFL executive with 20 years of experience in analytics. “Cleveland is a crazy, terrible place for this to be tested in football.”

The idea that Cleveland is too toxic to be resurrected is…. well, it’s more supported by the data than you might think. Certainly DePodesta could turn things around, but if he doesn’t, he’ll just be the next man in a long line of failed Browns executives. You won’t be surprised to learn that Cleveland has the worst winning percentage in the NFL since re-entering the league in 1999. But even accounting for the fact that the Browns have been bad, Cleveland has still underperformed to the tune of about 26 wins over the last 16 years, most in the NFL.

How did I arrive at that number?

  • First, I calculated each team’s Pythagnpat winning percentage in each season beginning in the year 1999, which is based solely on the number of points scored and allowed by each team. For example, in 2014, the Browns scored 299 points and allowed 337, which translates to a 0.429 Pythagenpat winning percentage (the Browns actually beat that slightly, by going 7-9).
  • Next, I ran a regression on the years 1999 to 2014, using Year N Pythagenpart winning percentage to predict Year N+1 wins. This would, in theory, help out the Browns, because Cleveland would be expected to win fewer games than the average team in Year N+1 because the Browns typically have a poor Year N performance. The best-fit formula was 0.311 + 0.376 * Yr_N_Pyth_Win%. This shows that regression to the mean is a large factor, because past performance only accounts for 38% of what goes into a team’s projection for Year N+1; the remainder is a constant for all teams.

Using Cleveland’s 2014 line as an example, the 2015 Browns would have been expected to win 7.6 games, because the 2014 team had 6.9 Pythagenpat wins, and regression to the mean drives that number towards 8 wins. But Cleveland won just 3 games last year, falling 4.6 wins shy of expectation. And that’s only the second-most disappointing season of the new Browns era: in ’08, Cleveland fell 4.7 wins shy of its Pythagenpat prediction. Take a look at every Cleveland season from 2000 to 2015 (obviously there was no prediction for ’99, since there was no ’98 team): [click to continue…]

{ 5 comments }

Yesterday, I measured the age of each team’s passing attack by calculating the yards-weighted age of each player who gained either a passing or receiving yard. Today, the historical results.

I’ve written a bit about Terry Bradshaw and his terrible rookie season of 1970, mostly in the context of number one picks taking a long time to break out. But here’s something that often gets lost in the mix: Bradshaw was just one of many inexperienced players on the ‘70 Steelers.

Bradshaw played as a rookie that year at age 22 (Terry Hanratty also started 6 games, and was also 22). The top 6 players in receiving yards on the ’70 Steelers were wide receiver Ron Shanklin (age 22), wide receiver Dave Smith (23), tight end Dennis Hughes (22), fullback John Fuqua (24), wide receiver Hubie Bryant (22), and wide receiver Jon Staggers (22). Incredibly, five of those six players were rookies, with Frenchy Fuqua being the sole exception — and he was drafted in 1969! In the ’70 draft, Pittsburgh took Bradshaw with the 1st overall pick, drafted Shanklin at 28, Staggers in the 5th round, and Smith in the 8th round, while both Hughes and Bryant were undrafted free agents that year. That’s unbelievable, and makes the ’70s Steelers passing attack akin to an expansion team — or rather, an expansion team with almost no access to the veteran market. As a result, Pittsburgh’s 1970 passing attack ranks as the youngest in history: [click to continue…]

{ 4 comments }

Bortles led the 2nd youngest passing offense in the NFL

Bortles led the 2nd youngest passing offense in the NFL

After a 1-4 start to the season, it might have felt like an odd time to write about how the Jacksonville Jaguars could have the next great offense. But in many ways, Jacksonville’s passing attack only got better as the season went along. Some (the majority?) of the big numbers were more of a function of quantity than quality, but the numbers really were big. Consider:

  • Blake Bortles finished tied for 2nd in passing touchdowns and 7th in passing yards
  • Allen Robinson finished tied for 1st in receiving touchdowns and 6th in receiving yards. He also had the highest yards per reception average of any player with 1,000 receiving yards
  • Allen Hurns also hit the 1,000-yard mark, and had the 6th highest yards per reception average of any player with 1,000 receiving yards. Hurns and Robinson were one of just four duos (Jets, Broncos, Cardinals) to have two players gain 1,000 receiving yards.

That’s an impressive trio by any standard, but what’s incredible is that Hurns was born in November of 1991, and he is the oldest of the three! So how young is the Jaguars passing attack compared to other teams? I have decided to create a passing yards-weighted age grade for each passing attack. And in doing so, I chose to count passing yards and receiving yards equally, which of course has the effect of making the quarterback(s) equal to half of the team’s passing game. I’m OK with that. [click to continue…]

{ 3 comments }

AV-Adjusted Team Age (Offense) from 2012-2015

Background:

In 2012, the Jaguars went 2-14 with an offense centered around Blaine Gabbert/Chad Henne, Maurice Jones-Drew, Cecil Shorts, and Justin Blackmon. Since then, the team has been rebuilt, and gotten better and younger. Among offensive players, only Marcedes Lewis was on the team during each of the last four years. I’ll have more on the Jaguars tomorrow, but given the way the Jets have moved from young and bad to old and good, I think that’s the more interesting team to analyze today.

Here’s how to read the table below. In 2012, the Jets offense had an age-adjusted AV of 26.9; that dropped to 26.4 in 2013, then rose to 27.5 in 2014 and up to a league-high 29.2 last season. That’s an average of 27.5, but more interesting (to me) is the variance of 1.1 years. [click to continue…]

{ 0 comments }

2015 AV-Adjusted Team Age

In each of the last four years, I’ve presented the AV-adjusted age of each roster in the NFL. Measuring team age in the NFL is tricky. You don’t want to calculate the average age of a 53-man roster and call that the “team age” because the age of a team’s starters is much more relevant than the age of a team’s reserves. The average age of a team’s starting lineup isn’t perfect, either. The age of the quarterback and key offensive and defensive players should count for more than the age of a less relevant starter. Ideally, you want to calculate a team’s average age by placing greater weight on the team’s most relevant players.

My solution has been to use the Approximate Value numbers from Pro-Football-Reference.com.  The table below shows the average age of each team, along with its average AV-adjusted age of the offense and defense. For the second year in a row, the Jaguars and Rams were the two youngest teams in the NFL; this year, though, the team formerly known as St. Louis took the top spot.

The average AV-adjusted team age last season was 27.1 years; the Rams (25.6) and Jaguars (25.8) were the only teams below 26, while the Jets (28.2) and Colts (28.6) were the only teams above 28 years. Here’s how to read the table below, using the St. Louis line: the Rams were the youngest team in the NFL in 2015, with an average age of 25.6 years as of September 1, 2015. The team’s offense had an AV-adjusted average age of 25.0, the youngest in the NFL, while the defense was at 26.0, the second-youngest. [click to continue…]

{ 2 comments }

Bob Ford, a longtime fan of Pro-Football-Reference and Football Perspective, has contributed a 2-part guest post on Yards Per Carry Leaders. Bob is the owner and founder of GOATbacks.com, which looks at the greatest running backs of all time. Thanks to Bob for yesterday’s and today’s articles!


Yesterday, I looked at the YPC leaders for the 46 seasons since the merger was completed, 1970-2015 at the 100/120/180-carry cutoffs. Today, a look at the YPC leaders since 1970 at three higher thresholds. [click to continue…]

{ 17 comments }

Farewell to one of the greats

Farewell to one of the greats

Detroit Lions superstar wide receiver Calvin Johnson has likely retired. He had a pretty incredible six-year peak: Megatron gained 8,548 receiving yards in his last six years, the most by any player during their age 25-30 seasons. I don’t think there’s much of a debate that Johnson is a Hall of Famer, although I do think he’s not quite an inner circle member of the Hall.

The big reason for that is Johnson’s numbers have always been inflated by playing on a pass-happy team.  I’ve looked at this before, but (a) those numbers are now two years stale and (b) I want to use a different methodology today. So here’s what I did:

1) Calculate the number of pass attempts per game for each team in every season.

2) For the top 200 players, I then calculated the number of career games for that player.

3) Then, in each season, weight the number of team pass attempts per game by the percentage of games that player played relative to his entire career. For example, Johnson played 11.9% of his career games in 2012, and that year, the Lions threw 46.3 pass attempts per game. Therefore, for Johnson’s career, 46.3 pass attempts per game will be given a weight of 11.9%. Do this for every season of each player’s career, and you will derive the average pass attempts per game for that player. [click to continue…]

{ 21 comments }

In 2015, the average points differential was just 11.06 points per game.  That may not mean much in the abstract, but it’s the lowest in 20 years.  Take a look:

pt diff 1950

What was driving the close games this year?  It’s mostly because the “losing teams” wound up scoring more points, but the average points scored by the winning team did dip slightly in 2015, too: [click to continue…]

{ 5 comments }
Previous Posts