Monday Night Fallacies
I didn't see last night's game (the bad thing about Colts-Dolphins games is that one of them has to win). I did just hear it reported that the Colts had the lowest time of possession (TOP) of any winning team since some time in the seventies. Thesis statement: TOP is probably the most overrated stat in football.
My blood boils when I hear some commentator talk about how Pitt's offense kept the ball for 41 minutes against West Virginia, thus limiting the amount of time Pat White and Steve Slaton had to pile up points. This does not translate into any useful advice to a future opponent of West Virginia's other than don't turn the ball over.
TOP can be illustrative of how a game went. For example, if TOP is lopsided but the score close, perhaps one team's yards came mostly from rushing (which plays tend not to stop the clock, so that team racks up the TOP) while the other team passed more (the clock stops on incompletions, so their yardage piles up more on quick plays, without draining much clock time).
On the other hand, if TOP is lopsided and the score is, too, it is usually the case that the winning team also won TOP. This correlation is, I think, often misinterpreted as causation. Whether the winning team built up an early lead and handed the ball off to drain the clock in the second half, or whether the losing team threw four interceptions and kept giving the ball away, letting the winning team accrue TOP throughout, is not explicit by the final TOP figure. And either way, it's not as though any pre-game strategy of maximizing your team's TOP is one that will also help you win. If Pat White and Steve Slaton are your quick-scoring opponents, then holding onto the ball longer will indeed limit their ability to score, but no more than it will your own.
Put another way, Pitt should always run the play that gives Pitt the best chance to score (until, of course, the time remaining in the game is low compared to the score differential, when it makes sense to run out the clock if Pitt is ahead). Otherwise, why not have a receiver take a knee in the open field, taking a fresh set of downs instead of a score?
But so last night the Colts did a rare thing by winning a game while holding onto the ball for less than 25% of the time. More than half of their points came with an 80 yard touchdown on the first play from scrimmage and a quick drive in the final two minutes to go ahead at the end of the game. Add to it that Miami runs the ball a lot and the result is still unusual but now sensible. And that's the (only) real use of TOP: it's something that makes one say How queer! and want to discover how it happened. It's not something you should build a strategy around. At least, not a winning one.
My blood boils when I hear some commentator talk about how Pitt's offense kept the ball for 41 minutes against West Virginia, thus limiting the amount of time Pat White and Steve Slaton had to pile up points. This does not translate into any useful advice to a future opponent of West Virginia's other than don't turn the ball over.
TOP can be illustrative of how a game went. For example, if TOP is lopsided but the score close, perhaps one team's yards came mostly from rushing (which plays tend not to stop the clock, so that team racks up the TOP) while the other team passed more (the clock stops on incompletions, so their yardage piles up more on quick plays, without draining much clock time).
On the other hand, if TOP is lopsided and the score is, too, it is usually the case that the winning team also won TOP. This correlation is, I think, often misinterpreted as causation. Whether the winning team built up an early lead and handed the ball off to drain the clock in the second half, or whether the losing team threw four interceptions and kept giving the ball away, letting the winning team accrue TOP throughout, is not explicit by the final TOP figure. And either way, it's not as though any pre-game strategy of maximizing your team's TOP is one that will also help you win. If Pat White and Steve Slaton are your quick-scoring opponents, then holding onto the ball longer will indeed limit their ability to score, but no more than it will your own.
Put another way, Pitt should always run the play that gives Pitt the best chance to score (until, of course, the time remaining in the game is low compared to the score differential, when it makes sense to run out the clock if Pitt is ahead). Otherwise, why not have a receiver take a knee in the open field, taking a fresh set of downs instead of a score?
But so last night the Colts did a rare thing by winning a game while holding onto the ball for less than 25% of the time. More than half of their points came with an 80 yard touchdown on the first play from scrimmage and a quick drive in the final two minutes to go ahead at the end of the game. Add to it that Miami runs the ball a lot and the result is still unusual but now sensible. And that's the (only) real use of TOP: it's something that makes one say How queer! and want to discover how it happened. It's not something you should build a strategy around. At least, not a winning one.
posted by john at
3:28 PM
1 Comments:
As you know I agree with you 95% of the way on this issue. I certainly think "offensive plays/snaps" would be better. And the only reason I think that might be a predictor at all is b/c of less tangible factors such as player fatigue, momentum, and when a team largely admits that they are the lesser team and therefore want to "shorten" the game, possibly limiting the game to, say, 7-8 possessions each instead of the more common 10-13.
Though the principle behind your thesis is correct, that it is much more useful as a correlation than a causation or predictor.
Post a Comment
Subscribe to Post Comments [Atom]
<< Home