Archive for the ‘Charting Our Progress’ Category

Charting Our Progress – Time of Possession

April 26, 2007

Today we look at another of college football’s great axioms – time of possession, in our charting the progress series. I’d harbor a guess that this was one of the most discussed (or at least mentioned) variables during the Osborne era. Win the time of possession, win the game. Right? Maybe not.

According the work of SMQ, this statistic means very little, only slightly more than yards penalized in college football.

Here is what SMQ’s tables tell us about time of possession:

Really not much to look at there. Time of possession, just doesn’t seem to be all that important in college football. Or at least it wasn’t in 2006.

So it appears we have another reason to turn the volume to mute during televised college games – too much discussion about a variable that just doesn’t seem to matter. But why doesn’t time of possession matter like we think it should? Well there are actually already some pretty good explanations already out there, so let’s start with those.

First we have a fantastic rant from Brian at MGoBlog from a point last season in which Michigan was leading the nation in TOP (they ultimately finished 2nd behind Texas A&M).

Time of possession is a fraud. It is a fraud for these reasons:

You cannot “keep the ball away from your opponent” any more than a basketball team can keep the ball away from their opponent. When you score or turn the ball over, they get the ball back, no exceptions. Unless you attempt an onside kick, your opponent is getting the ball back after you’re done with it. They will have the exact same number of possessions you do, plus or minus one depending on end-of-half and end-of-game hijinks.

It describes the actions of teams after they acquire a big lead and not what they do to get said lead. One of the primary reasons Michigan is #1 in time of possession: they’ve jumped out to massive leads in many games and cruised home. Opponents like Michigan State and Notre Dame have spent entire halves in a spread hurry-up emphasizing quick movement of the ball. Meanwhile, Michigan leisurely pounds the ball into the line until the game is over. Result: in the second half Michigan three-and-outs can take more time than 80-yard touchdown drives by the opponent. This is hugely distorting and tells us nothing more than “Michigan has a big lead.”

It places undue emphasis on the run game. Michigan features a crushing ground game and a crushing run defense. Result: lots of Michigan runs and very few opponent runs. This naturally helps TOP, but the reason Michigan is good isn’t because they possess the ball for relatively large amounts of time but rather those crushing units. Time of possession obscures the real reasons for Michigan’s success.

Well played. I suspect TOP obscuring the real reasons for success can be said for a lot of folks. But we’ll come back to that.

There is also real, academic research that also contradicts the implied importance of TOP. Harold and Daniel Sackrowitz in an issue of Chance, argued against the use of a ball-control, TOP-favoring offense. They claim:

that “a team using ball control may reduce the number of possessions and points scored by its opponent, but it will lose more often than if it did not use ball control.”

In their study, the Sackrowitz team developed a mathematical model for evaluating the efficacy of three styles of play: unconstrained (a team’s normal mix of offensive maneuvers and actions), time-consuming (using additional time), and hurry-up (using less time than normal).

Beware dorky math mumbo-jumbo ahead.

They modeled a typical time-limited game by a multidimensional Markov chain — a sequence of random vectors, each of which can be written as a string of numbers. The first three numbers, or coordinates, may indicate which team has possession, the current point difference between the teams, and the amount of time remaining. The remaining numbers of a given vector could represent the values of other variables deemed important for the particular game being considered, such as the number of time-outs remaining. Such a vector describes the state of the game at the beginning of each possession.

Such a computational model allowed the researchers to check game outcomes over a broad range of possible situations involving weak and strong teams adopting different styles of play at different times. A consistent pattern emerged in all the cases studied: An unconstrained strategy is preferable to either the time-consuming or the hurry-up strategies for both teams, even when one team is demonstrably weaker than the other.

And what the hell does that mean?

“The results also force us to the realization that, despite what one feels emotionally, a proficient ball-control offense reduces the number of possessions for both teams,” the authors note. “Thus, if anything, one might guess that the better team would decrease its probability of winning by using ball control, particularly if it had reduced its probability of scoring.”

But what about seemingly irrelevant Super Bowls involving the Big Tuna?

In a 1993 New York Times article about football coach Bill Parcells, the reporter stated, “His masterpiece was the 1991 Super Bowl, in which his Giants defused the powerful and innovative offense of the Buffalo Bills through the simple expedient of denying Buffalo the use of the football.”

The Sackrowitzes have a different view. “Even in the supposed ultimate endorsement of ball control, the 1991 Super Bowl, the Bills had ten possessions (but punted six times),” they remark. “In that game, a great defense helped to create the illusion that ball control is effective.”

Hopefully you are at least starting to understand that time of possession is not a key to college football success.

Now let’s look at Nebraska’s time of possession under Bill Callahan.

Here we see some pretty amazing progress during Callahan’s tenure. Unfortunately it is in a statistic that we have highlighted to be irrelevant to success. But wait, here is where it gets a bit tricky.

If we look at Nebraska in 2006 we will notice that the Huskers actually went 8-2 in games in which they won the time of possession battle. They were just 1-3 in games their opponents had the ball longer. So TOP does matter then in Nebraska, right? Wrong.

Time of possession is still a shaky predictor of success. It looks like the football equivalant of a “red herring”. We think it matters, but research and a little better understanding of its ability to mask other, more important variables (that I will be focusing on later), allows us to see it a more appropriate fashion.

But for those that are still a bit confused, or who insist on hanging on to the importance of TOP, here is a bit more.

In 2006, Nebraska finished fifth in the country in TOP, but our record was just 9-5. Arkansas State finished one slot ahead of us at fourth in the nation. They finished the season with a 6-6 record.

In 2005 Nebraska won four games when they possessed the ball more than their opponent. The Huskers also won four games when their opponent won the time of possession battle.

It just doesn’t matter. Or at least not to the degree we might think it does.

Charting Our Progress – Yards Penalized Per Game

April 23, 2007

With the rebirth of this series I have decided to look at the statistical categories in order of relevance from least important to most important according to the fine work of SMQ. To accomplish this I have combined both parts of his stats relevance watch to determine which statistic was the least relevant in terms of W-L record and Top 25 ranking in 2006.

The overall loser was yards penalized per game. You can see its lack of relevance in CFB in the following table.

Now, here is the chart of Nebraska’s standing relative to the national average in this statistic.

And here is how they ranked nationally in each of those three season.

2004: 46th; 51.00 yards/game
2005: 63rd; 57.75 yards/game
2006: 41st; 43.86 yards/game

I could spend time analyzing Nebraska’s progress in this statistic, and it might actually be interesting to do so. But I won’t. Why spend time analyzing a statistic that has been shown to be irrelevant. How irrelevant? In 2006, Nebraska went 4-2 in games where they were penalized less yards than their opponent. In games the Huskers racked up more penalty yards they were 5-2 (the Cotton Bowl was a push as Nebraska and Auburn were both penalized 45 yards).

Therefore, I thought it would be interesting to examine why this statistic might not be as important as the casual observer might think.

First, let’s go with what SMQ said. After Part I:

“First counterintuitive result: the most penalized teams were slightly better as a whole than the least penalized teams. Penalty yardage, over the course of an entire season, had no discernible effects on winning and losing. You can probably think of a situation that would specifically argue otherwise, cuz penalties are definitely bad, mmmkay?, but they’re bad more as situational mistakes than an overall, cumulative drain.

Then after Part II:

”Again, penalty yards stand out as utterly meaningless; as in Part One, higher penalty yardage actually correlates slightly more with success, which makes no sense and should not indicate that jumping offsides is desirable or even, in the short term, meaningless (hello, Louisville), but the overall, cumulative consequences of flags were apparently nil.”

Perhaps surprisingly, SMQ wasn’t actually the first to discover the lack of correlation between fewer yards penalized and success. Football Outsiders also examined this finding in the NFL back in 2003. They point out that this phenomenon was actually first noted in a 1988 book, The Hidden Game of Football written by Bob Carroll, Pete Palmer and John Thorn. However, according to Football Outsiders these authors made a key error in describing it, when they said:

“[Penalties] don’t make a whole lot of difference. Over the course of a season, they tend to even out. For every drive-killing holding penalty, there’s an interference call that keeps a drive going.”

Football Outsiders doesn’t buy the “even-ing out” hypothesis, because:

”The truth is, penalties don’t even out. Looking at the whole of last season, it’s clear that some teams were consistently penalized more than other teams. The difference is equal to a few hundred yards, which is also the difference between the best teams and the worst teams in punt return yardage. Would we say punt returns don’t really matter and even out over the course of the season? Certainly not.”

Football Outsiders goes on to discuss that:

What is perhaps even more surprising is the discovery that the majority of Super Bowl champions have actually been more prone to penalties than their opponents in the regular season. In fact, of the 37 Super Bowl champions, 23 actually had more penalty yards than their opponents.

From here the article goes on to do the work for me. The talented Michael David Smith of Football Outsiders provides his theories for the finding that many of the most successful teams are actually penalized more in the NFL. I would guess that we could also apply these theories to college football as well. I will present each of Smith’s theories in italics and then provide my response after these.

1. Good teams have the lead late in the game, which means they’re on defense against the pass more often. This makes them more likely to be called for defensive pass interference, which is the only penalty that can cost more than 15 yards.

I never would have thought of this, but it seems like a possible explanation in the NFL. It is less helpful for college football given that pass interference is still just a 15 yard penalty.

2. Good teams are more likely to decline their opponents’ penalties and have their own penalties accepted. All the NFL’s statistics are for accepted penalties only; declined penalties are treated as if they never occurred. It would make sense that a good team is more likely to have a successful play and therefore decline an opponent’s penalty, whereas a bad team is more likely to have an unsuccessful play and take the penalty yards.

Another thought provoking hypothesis, that would be difficult to prove or dispel.

3. Good teams are more aggressive, and while aggressiveness is usually a positive trait in football, it can lead players to be penalized.

I actually like this hunch. It was the first one that came to mind. Think Florida State in the 90s.

4. Winning teams could be smarter about taking penalties at the right times. For instance, it’s often advantageous to take a delay of game penalty rather than waste a timeout. (This only happens a few times a season and probably isn’t statistically significant.)

I certainly agree with the lack of statistical significance portion of this hypothesis.

5. When discussing penalties, it’s important to keep in mind that, contrary to what coaches and commentators tend to say, penalties shouldn’t really be called “mistakes.” When an offensive lineman holds Michael Strahan, he didn’t do it on accident. He did it on purpose because he knew Michael Strahan would beat him otherwise. He just hoped he wouldn’t get caught. Ditto a defensive back interfering with Randy Moss. Yes, there are some penalties that are mistakes — offsides, false starts, delays of game — but even those would seem to happen more often against better opponents. I’d expect a tackle to be called for illegal procedure much more often against Jason Taylor than against some practice squad scrub. So when you see that the Giants’ opponents were flagged for more penalties than any other team’s opponents last year, don’t assume the Giants just got lucky. The Giants certainly played a role in it. Also keep in mind that NFL officiating crews are not all created equal. Some crews call more penalties than others. But even if one team was stuck with a flag-happy crew more times than another team, it would make no difference in the net penalties shown here.

Now we’re talking! This is one that definitely needs to be considered and should probably be analyzed more carefully. It certainly doesn’t clarify the entire picture, but no one ever seems to bring this one up.

Ok, let’s conclude this piece by keeping our wits about us. Teams should continue to attempt to avoid penalties whenever possible. And we should not expect to hear coaches come out and endorse a high number of penalties, but at the same time we now know a little bit more than most announcers about their relative importance. As Football Outsiders concludes:

So does this data say that penalties don’t matter? It most certainly does not. We’ve all seen penalties that had game-altering implications. But penalties are probably less important than coaches and commentators would have us believe. And this probably deserves further study.