- Home Archives for September 2010
The team rankings below are in terms of generic win probability. The GWP is the probability a team would beat the league average team at a neutral site. Each team's opponent's average GWP is also listed, which can be considered to-date strength of schedule, and all ratings include adjustments for opponent strength.
Offensive rank (ORANK) is offensive generic win probability which is based on each team's offensive efficiency stats only. In other words, it's the team's GWP assuming it had a league-average defense. DRANK is is a team's generic win probability rank assuming it had a league-average offense.
Mike McCarthy missed an opportunity Monday night. With the score tied at 17 and 1:44 on the clock, the Bears managed to earn a 1st and goal from the 9. The Packers had only one timeout, so CHI could run nearly out the clock and kick a field goal if they chose. FGs from that range are successful about 94% of the time, and with the good conditions, it's probably even higher.
Had CHI scored a TD on 1st down, that puts GB down by 7 with 1:40 to go, which gives the average team about a 10% chance of winning. A FG attempt after 3 clock-burning runs, gives GB almost no chance to win. GB's only hope would be to prevent a TD on 3 straight downs and hope a FG misses.
I got a few requests to analyze Rex Ryan’s decision following the Jets' touchdown to go up 7 points with 2 minutes left to play Sunday night. Would it be smart to try for the 2-point conversion? If the Jets convert, they go up by 9, and the game is effectively over. If they fail to convert, they’re still up by 7. On the other hand, kicking the XP puts the Jets up by 8, requiring the Dolphins to go for 2 points themselves if they manage to score.
Looking at things from the Dolphins’ perspective, they would have about a 0.11 Win Probability (WP) being down by 7 with just less than 2 minutes left. (I’ve done some extra analysis by including all similar situations where teams need a TD to win or tie with 2 min left. This expands the sample greatly if we assume it’s the final drive of the game, and each team has a 50/50 WP in overtime. Teams get a needed TD 22% of the time, putting the WP at 0.11, which is the same as what my general WP model estimates.)
I’ve added a new page to complement the individual player advanced stats. Until now, there were stats for all defenders, offensive skill players, and player stats by team on both sides of the ball. I’ve recently added a team-level page, which is a summary of each offense and defense in terms of advanced stats. You can always get to them via the 'Stats' menu item in the site header. There’s Win Probability Added (WPA), Expected Points Added (EPA), and Success Rate (SR) for every offense and defense since 2000. Stats for the run and the pass are broken out.
For example, so far in 2010 New England’s offense is back firing on all cylinders, leading the league with 43 EPA. This basically means that the NE offense has generated about a net point advantage of 43 points over three games. That’s about 14 net points per game. Everyone knows about their passing success, but they’re running the ball very well too. They’re 3rd in the league with about a 50% SR in the running game.
The regular flash-based live graphs are still available as always.
Advanced individual player stats will be updated immediately after each game this season.
Game probabilities will start for week 4. They'll be featured at the NY Times again this season.
TJ uses EP to analyze a big play call in the DEN-SEA game from last weekend.
NYT article on the 3-4.
From the Advanced NFL Stats Community reboot:
Dan Schlauch on placekicker salary distribution. Leave some comments and give Dan some feedback.
John Candido is parsing weekly play-by-play for everyone this season. If you use John's data, be sure to leave behind a thanks. I'll be updating 2010 data periodically, but not every week.
Keep the submissions coming! Thanks again to Ed for tending to 'Community.'
This week's edition of The Weekly League features notes about Sunday afternoon's Indianapolis-Denver game and also the Monday night contest between Green Bay and Chicago.
The following games have been chosen as they'll be available to the greatest portion of the network-watching audience, per the NFL maps at the506.com.
The Four Factors you see for each game represent each team's ability in four important categories (pass and rush efficiency, pass and rush efficiency against) relative to league average (where 100 is league average and anything above is good). They're only for this season's first two weeks of games, however. In other words, they're to be taken with the giantest grain of salt ever.
Finally, a glossary of all unfamiliar terms can be found here.
Indianapolis at Denver | Sunday, September 26 | 4:15pm ET
Kyle Orton Hears a "Who?" (And Probably a "What?" and "How?" Also)
If you look at our quarterback leaderboards and sort by either EPA/P (Expected Points Added per Play) or AYPA (Adjusted Yards per Attempt), you'll see a surprising name either near, or even right at, the top of each list: Kyle Orton. Through the first two weeks of the season -- first at Jacksonville, and then home versus Seattle -- Orton has impressed both with his raw totals (he's thrown for a little over 600 yards now, good for sixth in the league) and by means of more advanced metrics (by EPA/P, Orton is the best quarterback in the NFL, with a 0.41 mark; he's also fourth among QBs with a 7.4 AYPA).
My weekly post at the Washington Post's Redskins Insider site breaks down the decision making in overtime of the Texans-Redskins game Sunday. Both Kubiak and Shanahan faced the nearly identical situations: 4th down on the opponent's 34. One punted and one attempted the field goal. Which coach made the right call?
In a previous post, I detailed how the 3-4 defenses over the past 10 seasons have, on average, outperformed 4-3 defenses. However, it’s possible that those 3-4 defenses simply had better players and that the 3-4 only appears to be the better scheme.
This time around, to account for each individual defense’s strength, I used a multivariate regression. The dependent variable in each model was a measure of success—EPA, WPA, and Success Rate (SR). The variable of interest was whether or not the defense was a 3-4 (a 1 if it was and a 0 if it wasn’t). The resulting coefficient of this variable represents the advantage of the 3-4 over the 4-3.
Each individual team was also assigned a dummy variable intended to capture the team-specific strength of their defense. Conveniently, this technique also accounts for the general year-by-year improvement in the effectiveness of NFL offenses.
The 3-4 defense appears to be the wave of the future. Over the course of the recent decade, the number of teams that have employed the 3-4 defenses went from 3 to 14. This season, almost half the league will be fielding a 3-4 to one degree or another.
As the pass has become more important, it makes sense that defenses are adapting by replacing larger, slower players with faster, more agile ones. The unpredictability of the 3-4 front makes the offensive line's job that much harder. The success of defenses such as the Steelers, Ravens, Jets, and even now the Saints, suggests there may be a substantial advantage to the 3-4.
Using the concepts of Expected Points Added (EPA) and Win Probability Added (WPA), we can compare the effectiveness of the two defensive schemes over the past 10 seasons. EPA measures every play in terms of the change in the potential gain of net point advantage. WPA measures every play in terms of how much it changed a team's chances of winning the game at hand.
Michael Vick intrigues me, his personal behavior aside. I suppose his playing style just stands out in a league full of clones. Last year I wrote that Vick was a better passer than most people think. His conventional statistics, such as passer rating and completion percentage, obscured his true performance. The difference isn't just the running yards he could gain, but the fact that he tended to throw deep because he was his own check-down option. He's back starting this Sunday, so it's a good time to take another look at Vick's career, this time through the lenses of Expected Points Added (EPA) and Win Probability Added (WPA).
Here's a cool team ranking site. It's based on only score and home field advantage, but I like the algorithm, and the presentation is cool. It reminds of Beat Graphs, which has a new look.
I really enjoy watching the NFL Films series America's Game. Now you can watch them on Hulu. Here's the Ravens 2000 season. Man, am I glad Ray Lewis is on my team.
Florida Danny at Niners Nation does some extra homework on how bad NFL pre-season predictions really are.
A couple interesting posts at Fifth Down. Week 1 winning teams make the playoffs 52% of the time, while week 1 losing teams make the playoffs only 23% of the time. (In case you think that this implies 75% of teams make the playoffs, like I did, that's not the case. 16 teams are 1-0, half of them--8 teams--make the playoffs. 16 teams are 0-1, and a quarter of them--4 teams--make the playoffs. That's a total of 12.) There's two things at work. First, 1-0 teams now have a numerical head start toward earning enough wins to qualify, and second, it's an indication of a good team.
About a year and a half ago, I launched Advanced NFL Stats Community, a site where anyone could contribute a post. It received a few dozen submissions, some more interesting than others, and it was a great start. Unfortunately, the time I devoted to reformatting the articles and their tables of data took away time I wanted to spend on other projects. I decided to gradually let the site fade into the background.
But one of my goals continues to be helping building a collaborative community of football stat heads without “premium” content or “proprietary” black-box stats, and I’d like to offer readers a platform for sharing ideas and analysis. In the recent off-season, I completed a long-standing goal of building an open NFL play-by-play database suitable for research and making available to everyone. Immediately, readers started parsing the data, adding fields for various things, and sharing their insights in the comments section of the post.
This week's edition of The Weekly League features some brief notes about the Miami-Minnesota game and then some less-than-brief notes about the New York Jets-New England match-up.
The following games have been chosen as they'll be available to the greatest portion of the network-watching audience on Sunday, per the NFL maps at the506.com.
The Four Factors you see for each game represent each team's ability in four important categories (pass and rush efficiency, pass and rush efficiency against) relative to league average (where 100 is league average and anything above is good). They're also for Week One games only. In other words, they're to be taken with the giantest grain of salt ever.
Miami at Minnesota | Sunday, September 19 | 1:00pm ET
The Miami Defense
I suspect that last week's Buffalo-Miami game was painful to watch in its entirety. We in the Upper Midwest joined the game with a couple minutes left, and even that was probably too much. With the exception of a fourth quarter drive on which he went 7-for-7 with 82 yards (including lost yards from a sack), quarterback Trent Edwards looked absolutely lost, checking down almost instantly despite his team being down by five and deep in their own territory.
CHI RB Matt Forte was named the NFC offensive player of the week for his amazing 7-catch, 151-receiving yard, 2-receving TD game against the Lions. When I looked up the advanced stats for RBs, I expected to see him near the top of the list for WPA. I was surprised to find him more than halfway down, posting a net negative WPA. I thought there must be some bug in the system. There’s no way a guy with 151 receiving yards and 2 TDs could have a negative effect on his team’s chances of winning, especially in a victory.
I’m not suggesting that WPA should be the ultimate stat for handing out awards, but taking a closer look at Forte’s plays can reveal a lot about WPA and what kind of events truly affect a team’s chances of winning.
Advanced statistics for individual players are now available for week 1 of the 2010 season. Win Probability Added, Expected Points Added, Success Rate, Tackle Factor, and a myriad of other numbers will be tracked throughout the season. (Check out the glossary for definitions of each statistic.)
Who was the hero of week 1? Eli Manning with +0.58 WPA in the Giants' victory over the Panthers.
Miles Austin tops the wide receivers with an incredible +0.68 WPA in the Cowboys' loss to the Redskins. But that was mostly due to a +0.60 WPA play at the end of the game. Austin's 30-yard catch over the middle set Dallas up for what was almost a miraculous victory, if not for Alex Barrons' clothesline hold on Brian Orakpo.
Reader 'slushhead' asks:
What do you think about "taking points off the board?" Specifically, should Washington have declined the penalty on the third quarter field goal?
Leading 10-0, WAS kicked a successful FG with 10:29 to go in the 3rd quarter. An offside penalty on DAL gave WAS the choice between keeping the 3 points or accepting the penalty, which would give them a 1st and 10 on the DAL 12.
I'm hoping not to turn this site into 4thDownAnalysis.com, but I did get several requests to look at another controversial 4th down decision. (And I'll always be happy to take requests.) Down 14-13 to DET, CHI faced a 4th and goal from the 1 with about 9 minutes left in the 4th quarter. A FG is nearly automatic, giveing CHI a 2-point lead and a 0.59 WP.
Conversion attempts from the 1 are successful about 68% of the time. A successful TD gives CHI a 6-point lead and a 0.79 WP. A failed attempt leaves DET at its own 1, but with a 1-point lead, worth 0.44 WP for CHI. On net, the conversion attempt is worth 0.68*0.79 + (1-0.68)*0.44 = 0.68 WP. Going for it was the right call by a large margin--0.68 to 0.59.
CHI went for it, but Matt Forte was stopped for no gain. But that's only half the story.
I received a few requests to analyze MIN’s final drive from last night. With a little over 5 min left, down by 5 points, they faced a 4th and 11 from the NO 44. MIN punted, and NO was able to convert 3 first downs and seal the game. Should MIN have gone for it?
The 2010 season starts tonight, and Advanced NFL Stats returns in mid-season form. I'm going into a contract year, and I'm in the best shape of my career. I had a great off-season working out at the team facilities. I've finally put my messy personal problems and last year's four-game suspension behind me, and my coaching staff is expecting big things. So draft Advanced NFL Stats first and put it #1 in your RSS reader.
The Win Probability graphs will be going strong again with a few enhancements. The model has been tweaked, with more refined estimates when the ball is near the goal line. I’ve added a live update of the Expected Points of the current drive. It’s a little abstract, but it gives a feel for the impact of events in the game. (That sack on 3rd and 4 just cost us 3 net points!). I also plan to add the current probability of a TD and probability of a FG at some point.
Also, during the last couple weeks of the playoffs last season, I added a live chat/comment capability under each game’s graph. We had a lot of fun with it, and I’m bringing it back this season. The chat from last year’s Super Bowl was great with all its twists and turns. I don’t expect every game to get flooded with instant commentary and analysis, but for the bigger national games I know readers will add their two cents.
Say there are two inside linebackers on the same team, John Andrew and Dan Farley. They’re considered equally good players, and both are in a contract year. Their team is good, and one of the early favorites to go deep in the playoffs this season. Each one has some choices to make over the course of the upcoming season.
Both Andrew and Farley want a big payday. In order to get as big a contract as possible, they need a lot of gaudy numbers to point to—lots of sacks, QB hits, forced fumbles. Millions of dollars are on the table, and the temptation is strong. However, leading their defense to a successful year and going deep in the playoffs will increase their values too. But that can only happen if both Andrew and Farley play selflessly and put their team first.
On any given play, a LB can either guess at the play call, trying to get a sack or stuff a run. Or, he can play his responsibility within his scheme by reading and reacting, doing exactly what the coordinator expects him to do. In other words, he can do what his team is counting on him to do, or he can gamble trying to make a play himself, exposing his team to giving up big gains.
Today I'm beginning a weekly contribution at the Washington Post's Redskins Insider site. Well, it will be weekly as long as the commenters don't have their way. (There's a little "Report Abuse" link next to each comment, and I think I might have to report almost every one. Maybe readers were expecting another riveting Albert-Haynesworth-is-lazy post.) My contributions will be primarily Redskins-specific, but will draw upon the general concepts in the Advanced NFL Stats toolbox.
The inaugural post is a back-of-the-envelope calculation about how many wins McNabb can bring with his expected passing efficiency and turnover rates. Of course, there's more to a passing game than just a QB, but this gives Redskins fans an idea of the range of potential improvement.
My thanks to Lindsay and Mitch at the Post for the opportunity. Looking forward to a fun season.
Five notes before we begin:
1. Thank you times a thousand to Brian Burke. His site is crazy good. It's my ambition not to destroy it.
2. Below are previews of two upcoming games: Thursday night's between Minnesota and New Orleans and the Monday game between San Diego and Kansas City. These previews aren't intended to be exhaustive, by any means. (In fact, the attentive reader will notice that there are only, like, 10 words dedicated to the Chiefs.) Rather, the intent is to construct a lens through which to watch these games -- hopefully something that your local paper isn't gonna provide. I discuss this more fully in the introduction thing near the end.
3. Oh yeah, there's an introduction thing ("Introduction-as-Epilogue") near the end of this document. It explores more fully the theoretical underpinnings to this document. If it makes you throw up, I apologize.
On other sites and message boards I often see Advanced NFL Stats referred to as “the Fangraphs of football.” That’s a pretty high compliment, and although I wouldn’t go nearly that far, I’ll certainly take it. There are a lot of similarities, and they were mostly accidental, at least at first. Both sites feature sound statistical analysis, and both feature live win probability graphs for their respective sports. After my own WP graph feature gained a following, I looked at Fangraphs for inspiration for how to further build a respectable sports analysis site. Gradually, Advanced NFL Stats became Fangraphs’ creepy stalker Jennifer Jason Leigh character from the movie Single White Female, and now the transformation is complete.
Carson Cistulli, a regular contributor at Fangraphs, has agreed to come on board to provide us with his brand of analysis, a slightly more entertaining look at football stats than what I can normally provide. He’s planning to start off doing some game previews, but may do some other things down the road.
Aside from his regular gig at Fangraphs, Carson is a writing instructor (maybe he can help me), a contributor at RotoWire, and a real-life poet. Yes poet. (Warning. Click at your own risk.)
So please welcome Carson, and you can look forward to his first post shortly.
In basketball, don't call a timeout at the end of close games. You'll be more likely to draw a foul and win.
A salary cap also needs a salary floor.
The historical distribution of talent in the NFL.
A neuroscientist's struggle to understand regression to the mean.
The Sports Reference sites, including PFR, was named among Time's 50 best websites. Congrats, guys.
Tony Romo is not a mirage.
Advanced Public School Stats. What's next, Fantasy Teachers leagues? ...With his first round pick, Brian Burke selects Jaime Escalante, math teacher, Garfield High School... Stand and Deliver-great flick by the way. Escalante passed away earlier this year.
Navy plays a triple option offense, a somewhat rare scheme even in college these days. The triple option allows Navy to compete against much bigger and faster opponents using discipline and execution as their weapons. They tend to be at or near the top of all Division I (FBS) schools in running every year, partly because they rarely pass and partly because they run the option so well.
Navy has enjoyed tremendous success in recent years, especially for a school that offers its recruits a rigorous and austere lifestyle, an extremely challenging curriculum, and virtually no chance of an NFL career. Last year they won 10 games, including their second win over Notre Dame in three years and a 35-13 drubbing of Missouri in the Texas Bowl. It was their seventh consecutive winning season and seventh consecutive bowl appearance. QB Ricky Dobbs broke Tim Tebow's record for rushing TDs by a QB last season. As my favorite Navy blog put it, archrival Army looked at their two-touchdown loss to the Midshipmen last year as a cause for optimism.
One thing about option football is that there is always a very good chance of a gain of at least a few yards. There aren't sacks or incomplete pass attempts. Navy always seems to be able to gain at least 2 or 3 yards when they need to, even when the defense knows exactly what is coming.