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).
It's true Vick is below average as a passer, but he more than makes up for it with his running. These EPA and WPA stats account for everything--sacks, fumbles, touchdowns, interceptions, etc. His career EPA shows he generated a net advantage of 152 points over his six years and change. He has generated just over 3.3 WPA over the same period. Over half of his career success came during the 2002 campaign, his best season by far. He's started 72 games over his career, so those numbers equate to about 2.1 net EPA per game and 0.05 WPA per game.
Year | Pass WPA | Run WPA | Other WPA | Total WPA |
2001 | -0.74 | 0.31 | -0.09 | -0.52 |
2002 | -0.09 | 1.79 | -0.64 | 1.06 |
2003 | -0.11 | 0.18 | -0.03 | 0.04 |
2004 | -0.85 | 2.18 | -0.11 | 1.22 |
2005 | -0.71 | 1.28 | -0.13 | 0.44 |
2006 | -0.65 | 1.84 | -0.12 | 1.07 |
2009 | 0.20 | -0.06 | -0.10 | 0.04 |
Career | -2.95 | 7.52 | -1.22 | 3.35 |
Year | Pass EPA | Run EPA | Other EPA | Total EPA |
2001 | -42.1 | 28.3 | -5.4 | -19.2 |
2002 | 43.6 | 50.8 | -12.2 | 82.2 |
2003 | -4.9 | 15.7 | -1.7 | 9.2 |
2004 | -51 | 73 | -4.1 | 17.9 |
2005 | -28 | 46.5 | -3.3 | 15.2 |
2006 | -22.6 | 63.5 | -3.6 | 37.3 |
2009 | 9.4 | 4.1 | -4.1 | 9.5 |
Career | -95.5 | 281.9 | -34.3 | 152.1 |
Both stats show his passing as well below average, but his running ability more than compensates. I'm not sure what Andy Reid's advice for Vick is, but
John Von Neumann would certainly tell him to run far more often.
Arguably though, Vick's great rushing numbers could come from a minimax equilibrium established by the defensive priority of shutting him down in the passing game. Whatever he leaves on the table in the passing game is probably worth more than whatever the defense is letting up in the run game, which is why he is not a winning quarterback. If he were to run more, I'm sure that his rushing numbers would go down. There is probably a reason why defenses are willing to let him run. Most likely having to do with an article you recently wrote, "Passing=Winning". I guess the real question is whether the pass sets up the run or whether the run sets up the pass? Based on the statistics, it should be the latter.
Why are Vick's "Other EPA" so high?
Just in general regarding Vick, I think his stats are outstanding considering his coaches, so afraid of the media, have continually tried to turn him into a traditional pocket-passer. Just imagine if he had a coach that actually planned the offense around Vick, rather than trying to force a style on him.
Ian-The columns got swapped somehow. Fixed. By the way, 'Other' includes penalties, such as false starts, delay of game, intentional grounding, etc., plus aborted plays like fumbled snaps.
You have to wonder what would happen if a franchise actually committed to building an offensive scheme around what Vick does best. Given his age and the record, I have difficulty foreseeing that ever happening, at least south of the 49th parallel.
John Candido, I think your rationale is flawed in general, but I'd like to highlight one line in particular in your analysis: "There is probably a reason why defenses are willing to let him run. Most likely having to do with an article you recently wrote, 'Passing=Winning'."
The whole point of that article is not (in a vacuum) that teams strictly pass too little. Basically, a variety of statistics - WPA, EPA, net yards - suggest that passing, in general, is more effective than running and therefore that passing is underutilized. The argument Brian makes, in a more generalized form, is actually "teams overutilize a statistically less effective strategy."
It is that same logic that justifies saying Vick should run more as, for Vick, the same statistics suggest that his runs are generally more effective than his passes. If he were to rush more, his rushing numbers would almost certainly go down - but his overall effectiveness would likely go up, as these statistics suggest Vick is running more infrequently than whatever the optimal balance of Vick runs and Vick passes is. Similarly, if non-Vick teams were to pass more, their passing numbers would likely go down, but (Brian hypothesizes, and I agree) their offensive effectiveness would increase as they got closer to the optimal amount of passing (Brian's analysis suggests, generally, that most teams are below this optimal level).
I'd like to echo this analysis with some subjective observation - in the Phi-GB game, there were several times where it appeared Vick could have easily successfully run and chose to throw a pass that ended up incomplete, most notably on the third down immediately before the Eagles kicked the field goal that took them within 7. Honestly, Andy Reid seems like the worst coach for Michael Vick as a starter, as the most likely to try to shoehorn him into the mold of a traditional quarterback - who happens to run around a lot - but is still expected to throw the ball at least 3-4 times as much as he runs on designed passing plays.
I agree with ajn; I think the idea here is that the running plays that Vick and the OC are "leaving on the table" would add more value than the passing plays that would be replaced.
Of course that's the idea in a general sense, and some opponents could require specific planning. For example, Detroit's defensive line seems to be somewhat improved from last year, and the linebackers might be decent, but the secondary is awful. Even though Vick seems to be generally worse at passing, against a team that is awful defending the pass, he'll probably add more value by throwing more, as he did Sunday.
AJN, in response, all I can think about is the concept of "opportunity cost". Obviously, it would make more sense for Vick, himself, to run a greater amount of the time. But that doesn't necessarily translate to more of a contribution for his team winning the game. His true value would be illustrated by the ease in which a below average passing quarterback could probably produce the same results as Vick can running all over the place. The only numbers that matter are the weights that go into predicting game win/loss. Passing dominates there, which is why I believe that Brian had the conclusion he did in the article. If passing is a significant variable, and rushing is insignificant, or random, then it follows that the run sets up the pass. If Vick uses his rushing threat to maximize his passing numbers, then that is where he could be a unique asset. But thinking he has any greater ability to influence the game than a running back does because he also throws passes on other downs is a misnomer when assessing his abilities in my opinion.
"...thinking he has any greater ability to influence the game than a running back does because he also throws passes on other downs is a misnomer when assessing his abilities in my opinion."
But Vick doesn't throw passes on "other" downs. His runs and passes both happen on called passing plays. If we are to compare his effectiveness to other quarterbacks, we have to look at his EPA on all passing plays, regardless of whether those plays are completions, incompletions, interceptions, sacks, or quarterback scrambles.
Put another way: Brian's conclusion that Vick should run more often is like saying a particular quarterback should throw to his tight end more often (because of consistently better results). It's an underutilized option within the sphere of passing plays.
Message received! I'm starting Vick. Thanks, Brian.
My problem is in equating WPA or EPA gained in the running game with that gained in the passing game. The only way to truly understand this would be to weight the variables in a binary logistic regression model to compare how much each contributes to winning games. Just like yards per carry in the running game is not a significant predictor of winning games, I would assume that it is the same for rushing WPA and EPA. Perhaps I'm wrong, but if Vick's running threat were to present an advantage anywhere, it would be in increasing his passing numbers (ie. Steve Young). The fact that he has negative passing numbers shows that he is incapable of capitalizing on his running ability to improve the pass. Therefore, in my mind, his contribution on running plays is just as good as handing the ball off to an equivalent running back. If his combined contribution as one individual is greater than an equivalent QB/RB combo then he is valuable. However, given his lower than average passing numbers, it doesn't seem that this is the case. If you think I'm still misunderstanding something here, due to the many opposing viewpoints, please spell it out cause I'm not seeing it. Thanks.
John-WPA would be exactly what you are looking for. WPA (and EPA) is agnostic to whether or not a play is a run or pass. It only measures the start state and end state of each play, and takes the difference in WP (or EP). So WPA already accounts for weighting how much a certain play-type tends to affect winning, and EPA does the same thing, but for scoring.
He may have negative passing numbers (meaning below average), but that doesn't mean his running threat doesn't improve them. They could be even worse had it not been for the threat of the scramble.
Yardage is also agnostic as to whether a play is a run or a pass. Yet, as a coach, what would you rather have during a game, 5 yards per pass or 5 yards per carry? The right answer is 5 yards per pass? Why? After all, they're both five yards. But one is more predictive of score/winning than the other. The same exists for WPA/EPA. Wouldn't it? Hypothetically assume the New Orleans Saints play the Philadelphia Eagles. Without knowing the result of the game, you only know that Drew Brees averaged 1% WPA per pass while Michael Vick averaged 1% WPA per run. Who is more likely to have won the game, Saints or Eagles?
That's not quite right. 5 yds per pass would be below average, and 5 yds per run would be above average.
Passing well (above average) is indeed more predictive of winning. That only means that given the choice, a team should prefer to be above average in passing rather than in running.
However, given that once a roster is set and the team is what it is, the team should do what it does best more often and do what it does worse less often.
Agreed, which is the point I was trying to originally make about "opportunity cost". Your observation about the minimax equilibrium of play selection involving Vick is correct. However, the Eagles would be a lot better off with someone who could productively pass the ball. Or if Vick would be able to utilize his running threat to improve his passing numbers. But that doesn't seem likely given his previous performances. Thanks for the clarification.
I think many, especially mainstream media and analyst continue to overlook the pink elephant in the room regarding Vick's success as a passer now.
He has WR's that can catch the ball and make plays. His coaching is far superior to what he had after Dan Reeves was fired in Atlanta.
In Atlanta Vick ran for his life. The offensive line was pourous in Atlanta. Look at the film. No one wants to acknowledge this for some reason.
Yes, he wasn't the most studious player in Atlanta. However Ben Rothlisberger wasn't the most studious according to some of his teammates and coaching staff. He won 2 championships.