Thursday, October 8, 2009

What's in a typical drive?

This post is about drives in the NFL. So, naturally, I'll be closing it out by talking about baseball.

A few months ago, I wrote an article about the merits of "grinding it out" versus "airing it out." In that article, I made the assertion that "I read somewhere that the typical NFL game has about 10 drives per team. I'm too lazy to do any real research on that, but it seems about right."

Well, I decided to get over my laziness and see what I could find out, not only about how many drives a team tends to get, but what the results of those drives usually are. I also read somewhere that about 1/3 of drives result in scores for the offense (probably in some article espousing the merits of the current overtime system), and I wanted to see if that matched up, too.

So, with a lot of help from Pro-Football-Reference's 2008 season stats (and a little help from's 4th down stats), I simply added up every "drive-ending occurrence" I could find. I counted a "drive-ender" as any instance of a:

Rush/Receive Touchdown
Field Goal
Missed Field Goal
Lost Fumble
Turnover on Downs
Blocked Punt

What I didn't count:

* Touchdown returns, since those aren't "drives" for the offense

* End-of-half/game drives that didn't result in one of the other options (like a FG attempt). The main reason was because I didn't have stats for them. But I don't think this is a huge problem. These drives usually fall into one of two categories: non-attempts to score, like kneeldowns or "protect-the-ball" runs, which I have no problem omitting; and actual attempts to score by teams in desperation at the end of the game. These should be counted, but, each team probably only experiences a few drives like this per year; often, they turn it over on downs or have a turnover before the clock runs out. Only plays that fail to score a TD on the last play of the half on 1st-3rd down should be counted, and those are really relatively rare.

So, with that exception, I should have compiled the results of every drive in 2008. And the results are:

Missed FG1552.8%
Lost fumble3286.0%
TO on downs2314.2%
Blocked punt130.2%

Drives per team171.5
Drives per team/game10.7

Scoring drives196735.8%
"TO" drives121322.1%

My early estimate -- that 1/3 of drives result in scores -- isn't too far off, as 35.8% of drives in 2008 resulted in either a touchdown or a field goal. But look at the last row. I count a "TO Drive" as a drive that ends in a very bad result for the offense: a turnover (fumble, interception, or downs), blocked punt, missed FG, or safety. In fact, your team has a better chance of bungling an offensive possession than it does of scoring a touchdown! 22.1% seems strangely high for me, but then again, I don't watch many Cleveland Browns games (ha!).

Bonus stat: The average drive scores its offense 1.89 points, if you count a safety as -2 points for the offense (and if you don't, it only raises the average by less than 1/100 of a point).

I'm also pretty close on my "10 drives per game" metric, although that 10.7 statistic would probably be pushed over 11 if it included those game-ending and half-ending drives I'm omitting. I also thought about baseball while putting this together. With 16 games in the NFL season and 162 in the MLB season, people often equate each game in the NFL to 10 MLB games, such as by saying that a three-game losing streak in the NFL is like a 30-game streak in MLB. When I was putting this together, I thought of the 10 drives/game concept and wondered if you could possibly equate each drive to an individual MLB game.

The answer is "yeah, if you're into that." With about 171 drives per season (maybe closer to 180 if we include the "invisible" drives), it's a fair comparison. And it makes one-game playoffs (which the Twins should be experts in by now) even more statistically dubious for their sample size. Imagine that two NFL teams finish with the same record on top of their division. Forget tiebreakers or even a tiebreaker game. We'll determine the division champion by giving each team one drive! It'll be just like a college football overtime game! I love what the Twins did on Tuesday night, but, in the grand scheme of things (especially when you consider how they won -- in extra innings with the lead flip-flopping back and forth), those two teams were identical in ability and the Twins got lucky, thanks to the results of an extremely small sample size.

I warned you there'd be baseball.

1 comment:

Peter said...

I really like the breakdown of potential finishes to drives. Makes me want to measure the Vikings against the baseline for the season so far... maybe look at several seasons to see if there's a correlation between wins and drive success. (and/or opponents' drive success)

Good stuff.