How Andy Reid wins football games with interceptions

Andy Reid’s teams throw a low rate of interceptions. The visual shows how his teams in Philadelphia and Kansas City have had a lower than NFL average interception rate (interceptions divided by attempts) in all but 3 of 18 seasons.

Reid’s Eagles had a particularly good stretch of pick suppression from 2000 through 2004. Led by quarterback Donovan McNabb, Philadelphia never won fewer than 11 games in any of those 5 seasons.

Despite a decreasing interception rate across the NFL, Reid has continued to beat the NFL average over the past four years in Kansas City. Led by quarterback Alex Smith, the Chiefs have won 43 regular season games and never dipped below 9 wins in any one season.

The randomness of turnovers

The visual goes against the typical quant narrative that turnovers are random.

For example, I’ve shown this visual that shows the relation between interception rate the first 6 games of the college football season versus the remainder of the season.

The lack of correlation between these quantities shows that you can’t predict a team’s interception rate later in the season based on the same quantity during first 6 games.

This suggests interceptions are random, and a team has a 50% chance to have a better or worse than average interception rate. However, if you assume this for Andy Reid’s teams, there’s only a 0.37% chance his teams would have had 3 or fewer seasons with a below average interception rate.

Randomness certainly plays a role in interceptions. No one who has ever seen a tipped pass fall into the hands of a defender should doubt that.

However, the Reid visual suggests that some coaches can suppress interceptions over a very large sample of games.

Steelers at Chiefs

This has implications in predicting the outcome of the Steelers at the Chiefs playoff game this weekend.

Kansas City doesn’t seem like much of a Super Bowl threat with the 16th and 11th ranked pass offense and defense, respectively, by my adjusted yards per attempt. I use these pass efficiency numbers to evaluate teams for two reasons:

  • My research shows the importance of passing over rushing in the NFL.
  • Turnovers have little impact on yards per pass attempt.

However, if Kansas City is truly skilled at not throwing interceptions, then these pass efficiency numbers will underestimate their team strength.

Team rankings based on adjusted margin of victory might be a better way to evaluate Kansas City. Their low interception rate will impact margin of victory, as I’ve found that an interception is worth 5 points in the NFL.

My member numbers combine both pass efficiency and margin of victory to make Kansas City a 1.3 point favorite against Pittsburgh. However, my team rankings based on only points would make Kansas City a 3 point favorite.

I interviewed Ben Alamar on the Football Analytics Show this week, and his FPI (Football Power Index) makes the Chiefs nearly a 5 point favorite. They use an expected points added, a metric which accounts for interceptions but their own twist. To listen to that part of the discussion, go to 14:50 of my interview with Ben Alamar.

Grantland, Betting Dork and 3 football predictions

I got into some NFL football this week.

First, I agreed to appear on Betting Dork, the podcast of Gill Alexander. During the NFL season, he invites a guest to appear with his regular round table that talks NFL games. Gill is a friend and all around great guy; dork might be the last word I would use to describe him. You can listen to the podcast each week here.

Second, an opportunity at Grantland came up. They made an excellent video on Kevin Kelley, the high school football coach in Arkansas that always goes for it on 4th down and always onside kicks. They wanted a blog post to accompany the video, so they asked me this: if not punting is one revolution in football analytics, what’s the next big revolution?

My first thought was that NFL teams should stop running the ball.

While this might seem crazy, numbers back up the argument. Including negative yards from sacks, NFL teams throw for 6.10 yards per pass attempt. On the ground, they only gain 4.17 yards per rush.

Moreover, over the last 10 NFL seasons, there is no correlation between rush efficiency, measured by yards per rush on offense, and winning. I found this lack of correlation shocking. The NFL is truly a quarterback’s league. Winning teams can throw the ball downfield while preventing their opposition from doing the same.

The article left a lot of room for further analysis, as people noted in the comments. Pass efficiency might decline with a higher percentage of passes. (Note that I do not think this is a given, especially with good play calling.) There’s also a higher risk for turnovers on pass plays. Hopefully, Grantland will let me follow up on these thoughts later.

You can read the article here. Be sure to watch the awesome video at the bottom on Kevin Kelley’s Pulaski Bruins.

I do think passing matters most in the NFL, especially if you want to predict the future. Yards per pass attempt correlates with winning even more than yards per play, the key stat I look at in college football.

This analysis is based on my NFL yards per pass attempt adjusted for strength of schedule. I’ll make all these numbers available soon.

Of course, I couldn’t resist talking about a college game at the end.

Kansas City at Denver

Kansas City has been one of the luckiest NFL teams this season. They have played a soft schedule and have benefitted from turnovers. The Chiefs needed 2 defensive touchdowns to beat Buffalo 23-13 in their last game.

So I was shocked when my numbers came down on the side of the Chiefs. The line has held steady at Denver at 8, while yards per pass attempt predicts Denver by 5. What gives?

I think people understand the problems with Kansas City. ESPN ran a piece on how the Chiefs were the most troubled 9-0 team in the history of the NFL. And I think that’s right.

However, people might be missing how bad Denver’s defense is. They are 28th in my pass defense rankings, which is just terrible for a Super Bowl contender. They have been a bit better the last 3 games since Von Miller has returned.

Overall, Denver gets its edge in this game from Peyton Manning and it’s top ranked pass offense against Kansas City’s 6th ranked pass defense. Denver has better than even odds to win.

However, don’t be surprised to see the Chiefs go to 10-0, especially if they can generate a pass rush against Manning and get some more turnover luck.

Minnesota at Seattle

Seattle is a legit Super Bowl contender. Minnesota is a poor team that features Christian Ponder at QB. However, a line that favors Seattle by 12 seems like too much. Yards per pass attempt predicts a 8.6 point win for Seattle. Remember, this prediction includes the throwing performance of both Christian Ponder and Josh Freeman.

Moreover, the run game could play a role in this game. Minnesota has RB Adrian Peterson, one of the most explosive players in the game. Their rush attack, ranked 5th by raw yards per rush, faces a Seattle defense ranked 21st in rush defense. While I don’t recommend building a team around a RB like Peterson, his presence can certainly affect this game in favor of Minnesota.

Georgia at Auburn

This game plays a surprising role in the national championship race. A Georgia win (with an Alabama win over Mississippi State) locks up the SEC West for Alabama. Auburn would have 2 conferences losses, and it wouldn’t matter if they beat Alabama in 2 weeks.

However, if Auburn wins, then their game with Alabama decides the SEC West. Then an Auburn win puts Alabama out of the title picture… like I predicted in Grantland a month ago.

Can Auburn win? My team rankings predict a 6 point win for Auburn. However, these rankings can be heavily impacted by turnovers, and Georgia has 7 more give aways than take aways this season. Had they performed better in this department, Georgia probably beats Missouri in their key SEC East battle.

Yards per plays predicts a Georgia win by 2 based on the strength of their offense. Despite a rash of injuries to key skill players, QB Aaron Murray has led the Bulldogs to 6th in my offensive rankings by yards per play. The line favors Auburn by 3.5, so expect a tight game that could come down to a last second field goal.

Thanks for reading.

NFL Rankings, Week 12

Rodgers, Manning twice
Quarterback, changes destiny
Happy Turkey Day

Happy Thanksgiving Power Rank readers!

Here are the rankings for tomorrow’s menu:

1.  Turkey legs  10.43

2.  Stuffing  8.74

3.  Turkey breast  5.21

4.  Mashed potatoes and gravy  4.98

5.  Pumpkin Pie  3.16

6.  Cranberry sauce  -25.47

Just kidding!

On a more serious note, here is something I’m very thankful for this year:  19 years of the Packers (#1, 12.14) having a future Hall of Fame quarterback under center.  Not every fan base is as lucky as the one I am proud to be a part of, and it is good to reflect on that unbelievable fortune from time to time.

The Colts (#32, -12.18) are at the opposite end of that spectrum, losing their future Hall of Famer Peyton Manning to neck problems this year, sending them from a top-tier AFC powerhouse to the worst team in the league by almost every metric available.  For fans who have seen Peyton taking snaps every week since 1997, this season is quite a shock to the system.

Other teams haven’t been so stable at the quarterback position over the years, and their ups and downs have shown it.  The Bears, for instance (#5, 5.39) have always seemed to have trouble finding a reliable trigger in spite of years of talented defenses and running backs.  Chicago fans largely applauded the coming of Jay Cutler, the best quarterback talent the Windy City has seen since Jim McMahon.  Now Cutler is likely out for the rest of the regular season, and possibly a portion of the postseason, with an injured thumb.  The Power Rank has been measuring the Bears success with a competent starter leading the offense. With backup Caleb Hanie, their true power is likely diminished, which will show in the coming weeks both in their record and in their ranking.  For this week, do not count on the raw numbers as Chicago is favored over Oakland (#17, -1.15) by 4.5 on the road.

Another team made a big change under center this week.  The Kansas City Chiefs (#31, -8.22) have acquired the benched Kyle Orton from their division rival Broncos (#22, -2.35) to fill in for an injured Matt Cassel.  Tyler Palko was given a shot on Monday night to take the reigns, but a performance that yielded no touchdowns and three interceptions sent the Chiefs looking elsewhere.  Personally, I believe that not only is Orton an upgrade over Palko, he’s an upgrade over Cassel.  This is sure to be controversial, but I think Orton can deliver a ball to gifted receivers, which the Chiefs have in Dwayne Bowe and Steve Breaston.  It may not be enough for them to beat the Steelers (#8, 3.71) this week, but it may be enough for the Chiefs to start climbing out of the cellar and back into playoff contention.

Finally the Eagles (#13, 1.59) have been struggling all year, but managed to beat the Giants (#14, 0.99) with backup quarterback Vince Young last week.  This improbable loss helped bring the Giants’ division leading record closer to what the Power Rank has expected of the G-Men all year (mediocrity) and they are now tied for the lead with Dallas (#10, 2.85).  The Power Rank has considered the Cowboys a strong team all season in spite of a slow start with a tough schedule, but as things shake out they are rising to the top and The Power Rank expects this trend to continue.

That is, of course, assuming that Cowboys fans give their thanks tomorrow to ensure that Tony Romo can stay healthy for the remainder of the season!

Happy Thanksgiving everyone!

NFL Rankings, Week 5

Okay New England, we get it.  You’re good at football.  And St. Louis, it is likely that you’re already thinking more about the race to acquire Andrew Luck than the race to make the playoffs.  But today I’m not interested in the highs and lows of The Power Rank, I’m taking a shot at it’s sweet, juicy center.

To get a look at what truly constitutes the center of the Power Rank grouping, I used a mathematical formula for Standard Deviation, something that defines the variation from the mean (or average) in a data set.  If you’re enough of a nerd to not stick your tongue out at that definition and want to know more, you can look at Wikipedia’s explanation.

If you’re like me and most math classes made you go crossed-eyed and start drooling on yourself, all you really have to understand is that the bulk of a group (about 68%) falls within 1 standard deviation of the mean on either side, and that the bulk of the remainder (about 27%, for a total of 95% of the whole) falls within 2 standard deviations of the mean.

In other words, teams whose rating falls within 1 standard deviation of the mean (always 0.0 for the Power Rank) are all horribly mediocre.  Ok, that’s my inner pessimist coming out.  A more optimistic view for Eagles and Falcons fans might be to say that they are “on the bubble” when it comes to elite NFL teams (or horrible NFL teams, but we won’t dwell on that).  On the other hand, teams that exceed 2 standard deviations of distance from the mean are truly in a class of their own, either high class or low class depending on which side of the curve they are on.

That’s about as much explaining as I can do, although further questions about the mechanics of this process can be emailed to Ed, who will no doubt be able to give you a thorough explanation of the math that goes into this process.  For my part, I just plug numbers into a free online calculation program and analyze the output.  Ah… sweet, sweet technology.

On to football.

The standard deviation in this week’s power rank is 5.49.  That means that the bulk of teams will fall between 5.49 and -5.49, almost all teams will fall between 10.98 and -10.98, and teams beyond those ratings are truly special.

Congratulations to the Patriots (#1, 15.79) and the Packers (#2, 11.31) for pushing the limits and existing beyond the norm.  Perhaps even more congratulations are deserved by St. Louis (#32, -9.89) for not exceeding the norm…

Very few teams fall between the first and second standard deviations.  On the high side only Baltimore, Detroit, and New Orleans (by a hair) make the grade as especially good teams, whereas on the low end Denver, Arizona, Cleveland, Kansas City, Seattle, and St. Louis all currently qualify as truly not very good teams.

That leaves the other 21 teams in the true statistical middle of the road.  Being in the middle isn’t all that bad, as you are supposedly as close to the top as you are to the bottom.  This is great news for 2010’s weekly bottom dweller Carolina, who finds themselves just within the boundaries of that first standard deviation, but not great news for teams hoping to return strong and make another playoff run like Pittsburgh, Atlanta, or Philadelphia.  Most importantly for these middling teams, their current ratings are not a death warrant for the season, they have no cause for alarm and no need whatsoever to join in the chase for the Andrew Luck Sweepstakes.

A few things to consider for these middle teams:

1.  The current standard deviation is almost 1 point bigger than it was at the end of last year when it ended up at 4.59.

2.  At the end of last season The Power Rank was a little more balanced with one team above 2 standard deviations (New England) and one team below (Carolina).  Currently the two teams exceeding 2 standard deviations from the mean are both on the high side. Most likely, either New England or Green Bay will fall back into the sweet center during the season.

3.  When one (or both) of the juggernauts fall they will bring that standard deviation down with them.  This will cut some teams out of of the running for average status (look out Indianapolis, Jacksonville, and Carolina!) but on the upside a few teams may be thrust into greatness without needing to earn it (it may finally be the year for Houston or San Diego to go all the way).

4.  When the standard deviation shrinks a couple struggling teams may also become hopeless.  But come on, we are only one quarter of the way through the season!  Now is the time for Vikings fans to Ponder over whether or not they can finish out 12-4, Miami fans to Marshall their courage, and Colts fans to…  oh, who am I kidding?  Without Manning they have lost their identity…  they should focus on battling St. Louis and Kansas City in the race for the #1 draft pick in 2012.

NFL Rankings, Week 4

Upper middle class
The Eagles, Falcons, Steelers
Room for improvement

This week, we see the rankings compress a little bit more, a trend that should continue over the next few weeks as this year’s games stabilize.  The Patriots remain #1 with a 14.73 rating after losing in a shootout with red-hot Buffalo (#13, 1.20).  The Chiefs helped their fans get off the ledge by keeping the game with San Diego (#12, 2.02) to within a field goal, so despite their loss they climb out of the cellar to #30 with a rating of -8.26.  The Rams now occupy the not-so-coveted #32 spot with a rating of -10.96  after a humiliating loss to #3 Baltimore (7.75).

What this means is that last week the range of ratings between #1 New England and then #32 Kansas City was 35.63, whereas this week the difference between #1 New England and new #32 St. Louis is 25.69.  That’s not encouraging news for Rams fans, or for the entire state of Missouri.  What is does indicate is what the scoreboards have shown this year:  no team is utterly dominant (sorry, Mr. Brady) and no team is completely pathetic.  Bad, maybe…  but not insanely horrible as Kansas City’s -19.52 rating last week indicated.

The news is best for fans of the “upper-middle class” of the football teams.  Not the dominant upper crust teams like New England (#1, 14.73, Green Bay (#2, 11.37), and Baltimore (#3, 7.75) with giant stretches of blue glory next to their names.  No, this news is for the fans of teams between rank 4 and 16.  The fans of teams that were supposed to come out with guns blazing, but have somehow tripped over the starting line.  What happened to the free-agent force of the Eagles (#15, 0.82) that was so highly touted by the offseason hype machine?  Wasn’t the return of Matty-Ice with Julio “the missing piece” Jones supposed to take the Falcons (#16, 0.62) from a playoff team to a Super Bowl favorite?  Did the mighty Super Bowl runner-up Steelers really fall to #6 with a good but uninspiring rating of 3.97?

It is these teams who should take note of this week’s movement.  Success in climbing The Power Rank ladder isn’t going to be as much about raising their rating from the 0-4 range up to the heights of Patriots in the teens.  That just won’t happen.  Because the outliers like the Patriots and Packers have little to gain (mathematically speaking) by crushing teams that The Power Rank expects them to crush.  But they can lose a lot by falling to teams who are beneath them.  Even close wins to inferior teams will lower their prestige.  So Big Ben and the Steelers don’t have to worry about trying to hang 70 on the the Brown and the Bengals to catch up.  The Eagles don’t need DeSean Jackson to return 2 punts for touchdowns every game to become elite.  If they can just continue to win for a few more weeks, the over-inflated Patriots and Packers will fall to more reasonable and reachable positions.

*Interesting note:  Running numbers from this year alone, the Bills would be The Power Rank’s chosen team with a rating of 14.28, with the Raiders not far behind with a 13.97 rating.  St. Louis would still be in the cellar, but at an astounding -23.50 instead of their current rating of -10.96.  While the Bills are definitely the hottest team in the NFL and a quickly growing fan-favorite, their official position of #13 suits them quite a bit better.  The Bills have proven that they have the talent, but they still have yet to prove to the league and to the algorithm that they have the ability to be consistent.  Because after all, that is the mark of a true champion.  By considering last year’s data as well as this year’s, The Power Rank cuts through the knee-jerk reactions and keeps the Bills somewhat humble with objective mathematical analysis, while still allowing them room to continue climbing up the ladder if they continue their run of impressive wins.