Mike Smith Was Right

by Kevin Meers

Statisticians usually complain that NFL coaches do not go for it enough on fourth down. They usually take the “safe” way out of 4th and short and punt the ball, going for the most publicly appealing choice instead of their best chance of winning.

When my roommate and I were watching the game, we told each other how dumb it was, that we knew coaches should go for it more, but not here. People are going to want Mike Smith’s head tomorrow morning because of his decision to go for it on 4th and 1 from his own 29-yard line.

We were wrong. Mike Smith was right.

His decision to go for it on 4th and 1 maximized his team’s chance of winning – exactly what the Falcons pay him to do. At that down and distance, the Atlanta Falcons’ win probability was 0.42. Given his situation, the Falcons were probably going to lose. Mike Smith had two options: punt, and hope his defense could hold the Saints offense, or go for it.

If they went for it, they could have either succeeded or failed. Historically, teams succeed on 4th and 1 from their own 29-yard line 74% of the time. If they failed, they would give the Saints the ball – leaving the Falcons with a win percentage of 18%. If they had gained that one-yard and gotten a 1st down and 10 from their own 30-yard line, their win percentage would have risen to 57%. In the game, Falcons’ running backs had rushed 39 times for a total of 126 yards, an average of 3.23 yards per carry. They had rushed for 0 or fewer yards just three times in the game, giving them (perhaps) a better-than-historical-average chance of converting.

But let’s stick with historical averages. The Falcons expected win percentage given that they went for it was , which equals 0.47.

Alternatively, Mike Smith could have chosen to punt the ball. Matt Bosher, a rookie, has punted the ball 39 times in his career so far, averaging a net of 35.6 yards per punt. Given an average punt and return, the Saints would have taken over at their own 35-yard line with a win probability of 0.61 – meaning the Falcons would have had a win probability of 0.39.

So Mike Smith made the right decision: his team was 8% more likely to win if he went for it on 4th and 1 instead of punting it. Since his offense did not pick up that yard, he looks like an idiot for giving Drew Brees the ball on the Falcons’ 29. However, this feeling is nothing but hindsight bias: Mike Smith made the right call.

Huge thanks to www.advancednflstats.com for providing all of the win probability data in this article.

About the author

harvardsports

View all posts

3 Comments

  • When you say that the Saints had a win prob. of 0.18 after the conversion failure:
    1) Is this based only upon past overtime games? And if so, how can the sample size be sufficient?
    2) In the particular situation (where hitting a long FG is critical to the analysis), any historical analysis will underestimate the chances of the team in long-FG field position, as FG accuracy has risen significantly in the past 20 years. Not to mention the good dome kicking conditions.

  • Some notes: Due to changing NFL conditions the success rate for a run on 4th and short is lower than the historical average. Furthermore kickers have become more effective at 40+ yard field goals. However, while those 2 effects make the decision relatively worse, it is unclear how whether or not they make it 8% worse in terms of win%.

    http://www.advancednflstats.com/2011/11/running-into-trouble-on-4th-down.html

    http://sports.espn.go.com/nfl/columns/story?columnist=pasquarelli_len&id=3112032

    Seth Burn

  • Kickers are hitting on FG’s from 30-39 yds at an 88% success rate, and at 77% from 40-49. Assuming that the Saints only gain 6 yards and kick a 40 yd FG, the success rate would be around 82% on its own. If they get a first down – hardly an unrealistic probability – that field goal kick gets shorter and more likely to succeed. Or, they could end up scoring a touchdown to win the game as well. And, even if they missed the kick, it’s not like they automatically lose the game. That 18% figure just stinks to high heaven, and I’m disappointed that no one thought it was a good idea to question the accuracy of that number, because it seems absurdly high. I also think the 74% success rate seems high, but I have nothing more than my intuition to back that up, though I suspect the sample size of situations like this is very small.

Leave a Reply

Your email address will not be published. Required fields are marked *