Game Probabilities - Week 17

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week I also discuss home field advantage when a dome team visits a cold climate.

The 2009 All-WPA Team

The day after the Pro Bowl rosters are announced there are the obligatory "snub" articles in local papers around the country. In the DC area, the annual London Fletcher snub article is simply reprinted from last year's Post. So what about those Vincent Jacksons and Cedric Bensons who were unfairly left off the roster in favor of big name stars who may not have had a particularly good year? Who really earned their ticket to Miami?

I'll compare players using two different stats. Win Probability Added (WPA) measures each play's increase or decrease in a team's chances of winning. For every play that a player is mentioned in the play-by-play description, including penalties, turnovers and everything else, the WPA is tallied in his name. WPA is a narrative stat. It tells the story of what happened and is very context-dependent. It measures performance when it matters most. It has limited applications in terms of predicting future player performance, but it my mind it's perfect for comparing Pro Bowl and MVP contenders--even Hall of Fame candidates once there's enough data.

Request: Who Are The Pro Bowl Snubs?

I'm working on a comparison of players who were either selected for or snubbed from the Pro Bowl rosters. I'm tabulating each player's Win Probability Added (WPA) and Expected Points Added (EPA). In the comments here, list the players you think were legitimate all-pros but weren't picked. "Fantasy" positions only please (QB, RB, WR, TE). Except in rare cases, I can only analyze the positions that get run with, catch, or throw the ball.

Thanks!

Andy Reid Is No Longer My Hero

Note: This post was taken from a series of comments on the recent Games of Week post. Thanks to all who contributed to the discussion.

About a month ago Eagles head coach Andy Reid was my hero of the week for his daring, and smart, onside kick to open the game against the Redskins. Although it failed and gave up 7 points, he got them back by going for it on 4th down inside field goal range, succeeding, and getting the touchdown.

This past week, however, Reid is again the goat. With 3 minutes left in a tie game against the Broncos, the Eagles faced a 4th and 1 from their own 49. The Eagles punted. This was a big mistake, and you don't even need fancy math or some win probability model to prove it. Worse, Reid made the exact same call a year ago in the infamous tie against the Bengals.

If you have a 4th and 1 at about the 50 and you go for it, the value of the two possible outcomes are exactly equal. In other words, either I have a 1st down at the 50, or my opponent has a 1st down at the 50. It's perfectly symmetrical. Now consider what punting means.

Team Playoff Probabilities - Week 17

Courtesy of NFL-Forecast.com, here are the latest playoff forecasts.

The NFC teams are locked up. Only the seeds are left to be settled. The AFC wildcard contenders keep losing. Only the Jets and Ravens control their own destinies this Sunday.

Run-Pass Imbalance on 2nd and 3rd Downs

I've recently been looking at the imbalance in the payoffs for running and passing on first downs. The results suggested that most teams should generally pass more often outside the red zone and run more often inside the 10-yard line. What about 2nd and 3rd downs?

Game theory tells us that when the payoffs for two strategy options are unequal, the strategy option with the higher payoff should be selected more often. As the opponent adjusts to counter the new mix of strategies, the payoff of the favored option will decline while the unfavored option becomes more lucrative. Eventually, the payoffs for both options equalize, and at this point the overall payoffs are optimum. In two-player zero-sum games this is known as the minimax, or more generally as the Nash Equilibrium.

I used Expected Points (EP) to value the payoff of each play. Expected Points measures the net point advantage that the play result gives to an offense. It captures the value of yardage gained and lost, first downs, sacks, penalties, turnovers, and everything else in terms of equivalent point value. The change in EP resulting from a play is called Expected Points Added (EPA).

One of the things EP does not measure is the time value of a play. In situations when a team has a significant lead, the true value of a run includes the time burned off the clock. To a team behind late in a game, pass attempts have more value because they are more likely to stop the clock. For this reason I only include plays in the first and third quarters and when the score is within 10 points. This excludes trash-time plays and plays affected by the clock.

Team Efficiency Rankings - Week 17

It looks like the Bengals really are frauds, as Rodney Harrison kindly called them last Sunday night. Well, maybe not frauds, but they are an average team posing as a dominant division winner. Anything can happen in a single elimination tournament like the NFL playoffs, but I wouldn't count on them going very far.

It's no surprise the Jaguars have fallen apart down the stretch. They have been a slightly below-average team for most of the year.

The reigning champion Steelers are still a very good team, but are a long-shot for the playoffs. Had Roethlisberger been able to play in the the first game against Baltimore, which was lost in overtime, the wildcard situation might be very different right now. That was a division and conference game, which would come in very handy in a tie-breaker.

The Jets had a relatively small chance to make the playoffs through most of the season despite their 3-0 start. That all changed in the 3rd quarter against the Colts. On the first series without Manning and their other starters, the Colts gave up a sack, a fumble, a touchdown, the lead, and a perfect record. They also handed the Jets a very strong hand going into the final week. Colts fans are upset at losing the shot at perfection, but maybe the fans who should be most upset are those of the Broncos, Steelers, Jaguars, Texans, and Dolphins.

Games of the Week

Tonight's game notwithstanding, the game of the week was the Denver-Philadelphia game. What started to look like a blowout turned into a barnburner. Down big early, the Broncos came back to tie before surrendering a late field goal. DEN 27 PHI 30 (EI 4.9).

The runner up was the Baltimore-Pittsburgh game (4.0 EI). The Ravens were down by 10 at the half but came back to tie in the 3rd quarter. They had multiple opportunities to win the game, but imploded with penalties. BAL 20 PIT 23.

The comeback of the week was also the upset of the week. New Orleans gave up a commanding lead over Tampa Bay, and ultimately lost in overtime. TB 20 NO 17. At their low-point in the game the Buccaneers had a 3% chance of winning (CBF 33).

Game Probabilities - Week 16

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week I also look at just how big a blowout Monday night's Giants-Redskins game was using the WP model.

Troy Polamalu and 'WPA Plus/Minus'

Troy Polamalu is one of my favorite players in the league. As a Ravens fan whose heart despises everything that is black and gold, I'm compelled to say that he is truly spectacular. (The picture to the left is his interception return for a touchdown that saved the AFC Championship game for Pittsburgh last year.) He's also a classy and selfless player who never showboats, a breath of fresh air in today's NFL. Unfortunately for Steelers fans, his injury this season has revealed just how great a player he is.

Polamalu has played in five games this season and has missed nine. Using the Win Probability (WP) model, we can calculate the Win Probability Added for any play (WPA). We can also sum WPA for any team, squad, or player to estimate the context-dependent contribution of that player to his team's chances of winning.

WPA for defensive players is particularly problematic. If a running back evades all ten other defenders but is tracked down 60 yards later at the 5-yard line by the eleventh, it's the tackler that will show up in the play-by-play and be charged with the loss in WP. That doesn't make sense.

But Polamalu's 2009 season provides us with a special opportunity. We can measure the WPA of the entire Steelers defense when Polamalu is playing and compare it to the WPA of the Steelers defense when he's not.

Team Playoff Probabilities - Week 16

Courtesy of NFL-Forecast.com, here are the latest playoff forecasts.

The Ravens have further strengthened their position in the AFC, while the battle in the NFC continues to be a battle between the Giants and Cowboys. The Cowboys currently have the upper hand, but the Packers' loss last week has made them vulnerable. The Ravens' weekend was miraculous. They won, while the Jaguars, Jets, Dolphins and Broncos all lost. Except for the Steelers, who still only have a small chance of making the playoffs, every single other competitor for the AFC wildcard lost ground. The Ravens went from the 'outside looking in' to the #5 seed.

Team Efficiency Rankings - Week 16

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.

GWP is based on a logistic regression model applied to current team stats. The model includes offensive and defensive passing and running efficiency, offensive turnover rates, defensive interception rates, and team penalty rates. If you're scratching your head wondering why a team is ranked where it is, just scroll down to the second table to see the stats of all 32 teams.

Desperation Graphed

[Note: I can't remember if I already posted this or not. I did this back over the summer and saved it for the season. But I don't think I ever clicked 'publish.' ]

How desperate do NFL coaches have to be before they start going for it on 4th down?

Recently I've shown that, as a rule, teams should be going for it on 4th down far more often than they currently do. But what is 'currently do?' How far do coaches let themselves get backed up against the wall before they start actually doing what's best for their chances of winning?

Games of the Week

Tonight's game notwithstanding, the game of the week was the Oakland upset over Denver (6.2 EI). The runner up was the Green Bay-Pittsburgh barnburner (5.8 EI), which was also the comeback of the week (33 CBF).

Unlike last week, there was a good number of competitive games, including:
CLV 41 KC 34
CIN 24 SD 27
IND 35 JAX 31
MIN 7 CAR 26 (which despite the score was very close until the 4th qtr).

Explanations for Excitement Index (EI) and Comeback Factor (CBF) can be found here.

Was the Steelers' Onside Kick Smart?

Up by 2 points with 3:58 left against the Packers, Steelers head coach Mike Tomlin called for an onside kick. Was this a good decision?

When onside kicks are expected, they are successful only about 20% of the time. But unexpected onside kicks are successful a surprising 60% of the time. I think we can say this particular kick was certainly unexpected. And surprise onside kicks can be the most beneficial when a team is ahead late in the game. Possession is critical.

Roundup 12/19

If you're snowed in like me this Saturday afternoon, you've got plenty of time on your hands. Early season college basketball doesn't really grab my attention. Neither does the apparently sponsor-less "New Mexico Bowl." (Actually, I often root for Fresno State due to being stationed nearby at Naval Air Station Lemoore, CA for many years.) Here are some links to help pass the snowy Saturday.

Chase Stuart from PFR looks at Steven Jackson's huge year for the struggling Rams. That's very unusual for a RB to have such a big year on a losing team. In a contribution to the Fifth Down Blog, Chase looks at the five least likely playoff teams in history based on their early season struggles.

Are Teams Going For It More In 2009?

Is it me, or do I sense that teams are going for it more often on 4th down this year? And maybe I'm imagining things, but I also get the feeling that teams have been even more daring since the Belichick 4th down decision against the Colts.

The three graphs below tell the story. The top two graphs depict the proportion of 4th down decisions by field position. The first graph is for the 2000-2008 seasons, and the second is for the 2009 season. The third graph isolates just the proportion of conversion attempts by field position so we can better compare the previous seasons with the current season.

Game Probabilities - Week 15

Weekly game probabilities are available now at the nytimes.com Fifth Down.

Team Playoff Probabilities - Week 15

Courtesy of NFL-Forecast.com, here are the latest playoff forecasts. The tables below do not include results from this week's Thursday game.

The Ravens have strengthened their position in the AFC, while the battle in the NFC appears to be between the Giants and Cowboys.

Expected Point Values

Lately I've been using the concept of 'Expected Points' (EP) as a measure of success for football plays. It's been the foundation of much of my analysis of fourth down decisions, onside kicks, run-pass balance, and even touchbacks.

Every down-distance-field position combination has an average net point advantage. For example, when an offense has a first and goal at their opponent's 1-yard line, they can expect about a 6-point advantage over their opponent in the long run. A first and 10 at midfield is worth about 2 EP.

Expected Points on first downs are easy to compute because there are so darn many 1st and 10s compared to any other down and distance combination. Here is the EP chart for 1st and 10 (or goal) from my fourth down study earlier this season.

Team Efficiency Rankings - Week 15

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.

GWP is based on a logistic regression model applied to current team stats. The model includes offensive and defensive passing and running efficiency, offensive turnover rates, defensive interception rates, and team penalty rates. If you're scratching your head wondering why a team is ranked where it is, just scroll down to the second table to see the stats of all 32 teams.

Games of the Week

Some weeks can have a lot of really exciting games. Week 14 wasn't one of them. This week's most exciting game was (technically) BUF 16 KC 10 (5.5 EI). True, this wasn't a shootout between two playoff contenders, but if you really cared about who won, this was exciting. Leads of 6 points often produce exciting games because a TD at any time can tip the balance.

The runner up actually had playoff implications.

The comeback of the week wasn't terribly thrilling either. NE had about a 1 in 6 chance of winning (a CBF of 6) late in the second quarter as CAR drove into scoring position with a 7-point lead. NE ultimately won, NE 20 CAR 10.

Run-Pass Imbalance by Year

I've been chronicling the imbalance in the payoffs of runs and passes on first down in several recent articles. One of the possible explanations for the imbalance is that coaches are too slow to adapt to the NFL rules that seem to become friendlier to the pass year after year. If so, maybe the adaptation to the new realities can be seen in a decrease in the payoff imbalance over recent years.

The graph below charts the difference in payoffs between passes and runs by year. As with my previous posts, data are limited to 'normal' football situations--when the score is close and when time is not yet a factor. As you can see, there may be a slight decreasing trend in the imbalance, suggesting coaches might be catching up.

Team Playoff Probabilities - Week 14

Courtesy of NFL-Forecast.com, here are the latest playoff forecasts. The tables below do not include results from the Thursday games.

The Jaguars and Broncos have strengthened their positions in the AFC, while the Giants, Eagles, and Packers have improved their chances in the NFC.


Game Probabilities - Week 14

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week I also lead-in with a discussion of the luckiest and unluckiest teams so far this season.

Luckiest and Unluckiest Teams - Week 14

For some reason this is my favorite stat.

I estimate team "luck" by using my efficiency regression model to calculate each team's expected wins--how many wins a team can normally expect, on average, given their actual performance in offensive and defensive running, passing, turnovers, and penalties. The difference between the expected wins and actual wins is what I loosely call team luck.


What do I mean by luck?

In my own life, I'm a big believer in hard work, preparation, focus, execution and everything else that isn't luck. Coaches and players can't let themselves think any other way for a single second. But once we account for all those things, what do we call what's left over? Statisticians call it "residual," and a substantial portion of any residual is due to random effect, including sample error and what I call "bunching." In a bounded and meticulously measured system like sports, a vast amount of the residual from any decent model will be due to randomness. A season of 16 games simply isn't long enough for the breaks to even out.

Team Efficiency Rankings - Week 14

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.

GWP is based on a logistic regression model applied to current team stats. The model includes offensive and defensive passing and running efficiency, offensive turnover rates, defensive interception rates, and team penalty rates. If you're scratching your head wondering why a team is ranked where it is, just scroll down to the second table to see the stats of all 32 teams.

Games of the Week

Game of the Week: OAK 27 at PIT 24 (6.7 Excitement Index)
Runner Up: NE 21 at MIA 22 (6.2 EI)

Comeback of the week: NO 33 at WAS 30 (100 Comeback Factor)

Defensive Run-Pass Balance

I've been writing a lot about offensive run-pass balance lately, recommending that most teams should pass more often. So far I've ignored half of the equation--Defensive strategies matter too.

Unlike offensive play choices which are discrete defined options (such as either run or pass), defenses have a continuum of options ranging from a prevent defense to a goal-line stand. Defenses don't have to commit to one or the other. Instead, they can bias their strategy anywhere from 100% pass to 100% run.

With play-by-play data, we know what option an offense chose. Unfortunately, without coach's film of every game and an intimate knowledge of a teams' defensive schemes and intentions, it's nearly impossible to know where on the continuum a defense stood on any given play.

Still, it may be useful to see how often a team was run against or passed against, and see how successful each play type was. The table below lists each defense's proportion of passes faced on first downs. Also listed is the average success of each type of play as measured by Expected Points Added (EPA). The difference in average success between passes-against and runs-against roughly measures the imbalance in play selection.

MVP Candidates Through the Lens of WPA

One of the cool things to do with a Win Probability model is to look at the total Win Probability Added (WPA) of a player. Each play in a game changes a team's chances of winning either positively or negatively, and the WPA for the plays in which a player takes part can be tallied. WPA discounts 'trash-time' performance, and emphasizes 'clutch' performance. The result roughly measures the past value of a player in terms of what matters most--winning games.

WPA is a narrative statistic. It tells the story of how a player performed, but it does not predict how he'll likely play in the future. It includes all the flukes, miracles, and everything else that makes football unpredictable and fun to watch. In essence, WPA is what the MVP award is all about.

I've tallied the WPA for the notable MVP candidates through Week 12 of the 2009 season. The table below lists various players and their WPA expressed a few different ways. There is total WPA, which is simply a raw sum. There is also WPA per game and WPA per play, which tell a slightly different story.

2009 Team-Specific Run-Pass Balance

Recently I've been looking at run-pass balance on first downs based on a principle of game theory. When strategy mixes are optimized, the two strategies will ideally produce equal payoffs. If they aren't equal, then the better strategy should be selected until the opponent responds with his own counter-strategy. Results suggested that, in the NFL overall, the gains by passing on first down exceed those by running. In turn, this suggests that offenses should pass more often than they currently do.

However, every team has its own relative ability between passing and running. You can't just tell the 2009 Raiders to start passing more often. Their running game may actually be superior to their passing game in terms of expected payoffs, so while most teams should be passing more frequently, it's possible a minority of teams should be running more often.

Team Playoff Probabilities - Week 13

Courtesy of NFL-Forecast.com, here are the latest playoff forecasts. The tables below do not include results from the Thursday games.

It looks like the AFC wildcard picture has changed with the Jaguars' and Texans' losses, and the Ravens' win over the Steelers. Although the Ravens currently hold the 6th seed (I think), they're projected to have a 40% chance of holding on to it.

The NFC wildcard picture has changed too. The Packers' and 49ers' chances have improved while the Giants' and Falcons' chances have taken a hit.

Game Probabilities - Week 13

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week I also discuss the playoff prospects for the Jets and Giants, and why they may seem pessimistic to most fans.

NFC Playoff Picture: Wildcards

Chris Cox from NFL-Forecast.com, who computes team playoff chances using his Monte Carlo simulation, provides the current playoff breakdown in the NFC. His AFC breakdown is here.

Yesterday, I posted a summary of the AFC playoff race. Today I turn my attention to the NFC. As a reminder, my approach is to tentatively declare Cowboys, Vikings, Saints, and Cardinals as division champs so that I can focus on the wild card races. The NFC East, and to a lesser extent the NFC West, are still up for grabs, but this assumption greatly simplifies the analysis.

Team Efficiency Rankings - Week 13

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.

GWP is based on a logistic regression model applied to current team stats. The model includes offensive and defensive passing and running efficiency, offensive turnover rates, defensive interception rates, and team penalty rates. If you're scratching your head wondering why a team is ranked where it is, just scroll down to the second table to see the stats of all 32 teams.

JaMarcus Russell: Concorde of the NFL

With Jamarcus Russel’s recent benching, there’s been a lot of talk about when it’s time for a team to cut its losses on a failed quarterback. I don’t have hard numbers at my fingertips, but I’d be fairly certain that if a QB isn’t playing above average football or there hasn’t been steady improvement, by the end of his second year, it’s time to move on. [Edit: Here's a good look at that very question at PFR.] There’s no question teams tend to stick with struggling QBs well beyond their expiration date, even when better alternatives exist. The real question is, why?

Let’s say you’re an out-of-town Bills fan, and before the season began you were understandably optimistic about the team’s prospects. You bought prime tickets to the January 3rd game hosting the Colts, including parking and a hotel room. Altogether the bill comes to $400. In August, this feels like a great deal.

As the season wears on, it becomes clear the Bills aren’t contenders. The coach is fired, and the upcoming Colts game is not looking promising, as the Colts appear likely be playing for home field advantage in the playoffs. Everything points toward a humiliating blowout. What’s worse, as the game approaches the weather isn’t looking good. Bills fans are always the hardy type, but the foercast is beyond bad—snow, wind, freezing rain, and bitter cold. You’re not exactly excited about the prospect of going to the game.

AFC Playoff Picture: Wildcards

Chris from NFL-Forecast.com has done an excellent break down of the AFC playoff picture. Enjoy.

With 5 weeks remaining in the season, there are clear leaders in seven of the eight divisions. The NFC East division champ is very much up for grabs but in each of the other divisions there is one team that has an 80% chance or better of winning the division. The Cowboys have the inside track to the NFC East title, and even if the Eagles or Giants ultimately win that division, the Cowboys are likely to claim one of the wildcard spots. For the moment, if we designate the Patriots, Bengals, Colts, Chargers, Cowboys, Vikings, Saints, and Cardinals as division champs, we can focus on the wild card races. In this article, I will focus on the AFC wild card race.