How important is time of possession for Michigan's success?
So, I went back and looked at the time of possession for Michigan and opponent over their first six games:
Time of Possession
|
Game
|
Michigan |
Opponent |
Ratio |
|
Utah |
30:41 |
29:19 |
1.05 - 1 |
|
Oregon State |
38:01 |
21:59 |
1.73 - 1 |
|
UNLV |
30:02 |
29:58 |
1 - 1 |
|
BYU |
38.38 |
21.22 |
1.81 - 1 |
|
Maryland |
34:19 |
25:41 |
1.34 - 1 |
|
Northwestern |
37:05 |
22:55 |
1.6 - 1 |
I vaguely remember Brian arguing that TOP was a useless statistic as far as wins were concerned. Obviously, TOP shows the offense is sustaining long drives, and the defense is forcing lots of three and outs.
But I think it is more important than that. Doesn't the TOP correlate strongly to their defense getting gassed (like Michigan did last year). And our defense being fresh and ready to go after the QB the whole game long?
I'm not a coach, so I'm hard pressed to analyze this intelligently.
(Note: ratios are slightly off. I just rounded the seconds to get the approximate percentage of a minute which is different than the seconds. As an example, a minute and 30 seconds, 1:30, is actually 1.5 when you calculate the ratio).
October 11th, 2015 at 2:26 PM ^
I did not expect that.
October 11th, 2015 at 2:29 PM ^
Because we didn't have the ball as much as we normally do. It was quick scores or short drives/punt or turnover.
Look at the last 3 domination games. They were domination games because we always had the damn ball.
October 11th, 2015 at 8:20 PM ^
Yup. And that shows you why it's a dumb stat to look at.
October 12th, 2015 at 8:28 AM ^
^^^ This ^^^
Related: I love how it's always assumed that "pounding the rock" always results in long, clock-killing drives (instead of run-run-pass-punt) and "basketball on grass" always means three-and-out.
October 11th, 2015 at 2:27 PM ^
Not in FOOTBALL (see Baylor)
But for us, it's critical. We HAVE to give our defense a rest and play field position if nothing else and Rudock is very solid in this aspect. He's the master of a 8 play drive that results in a punt.
I've learned to appreciate it because he usually gets the back back in better position than he left it because of the defense.
We have to win TOP and continue to allow our defense to rest, regroup, gameplan, adjust, etc. Do that and we win. There isn't an offense on our schedule that's good enough to dominate this defense unless they're given short fields, quick turnovers (turnovers), etc.
Anyone that consistently goes 80+ yards on this defense deserves a gameball signed by Durkin and Harbaugh...because that team is just flat out better than us.
October 11th, 2015 at 2:28 PM ^
It's not really showing you anything meaningful about a team's performance, other than the fact that they usually have the lead and can grind out the clock.
October 11th, 2015 at 2:29 PM ^
Its important and points out what a good job Rudock is doing despite his shortcomings. He is great at controlling the offense and tempo. They play very crisply. Has there been one moment of panic and a TO needed in any of these games? Not really. Another area Jake has been very good is in his ball handling skills, play action etc. This will pay off later in the year when he hits something downfield or pulls the ball out on the FB belly...and pitches to Drake going full speed around the end.
October 11th, 2015 at 2:30 PM ^
Sent from MGoBlog HD for iPhone & iPad
October 11th, 2015 at 2:30 PM ^
I think a lot of it has to do with the fact that were a run-first team. The offense may be sustaining longer drives, but that's probably a product of the clock not stopping that often when were on offense, even though the total amount of real-world time that we have the ball might be about the same per possession.
Having the defense constantly force short possessions on the other team also helps.
October 11th, 2015 at 2:30 PM ^
Time of possession is a symptom not a cause.
October 11th, 2015 at 2:36 PM ^
At the beginning of the game, TOP is the result of sustaining a long drive, and the result of the defense getting off the field. But I suspect TOP becomes a significant contributing factor as the game goes on. In the 2nd half, the other team gets demoralized, you keep their offense from chances, and you just wear them out. Somewhere in the 3rd quarter, TOP shifts from not only being a result of keeping the ball, it actually contributes in some way. Or so I think.
October 11th, 2015 at 3:06 PM ^
Concur. Football is a game played by people - not statsitics. Intensity, ryhthem, psychological, domination, being worn down, are diffcult, maybe impossible, to quanitify but as you suggest I think they have a culminitive effect over the course of a game. I understand the argument that TOP is a proxy for other measurables - so why not a proxy for the immeasurables?
October 11th, 2015 at 4:51 PM ^
Then count number of plays run from scrimmage, not time of possession. Indiana when they still had first- or second-string quarterbacks alive would demoralize opponents by scampering to the line and running another play at ludicrous speed. That doesn't eat up a lot of clock but it has a significant football effect. So does making the other team's offense stand on the sideline for 15 or 20 minutes while you keep getting 1st downs.
Why not use it as a proxy? Because it sucks as a proxy. And other things you can measure are better as a proxy than Time of Possession:
- % of your possessions where you had a multi-score lead at the end.
- Number of plays from scrimmage
- Number of first downs
- Actual time elapsed between possessions
- Clock used while with a multi-score lead.
The top one would need a season's worth of stats before the noise starts to leave the numbers, but it would also be of use in assessing a team's ability to limit possessions for the other team after a lead is established.
Measuring success at limiting possessions (whether to salt away a game or to keep your defense from having to run too many plays, or to raise the variance of the game if you're an underdog, etc.) would have some value, even if it's not as predictable (meaning good at predicting future success) as stats like yards per play or points per drive or percentage of available yards gained or turnover %. For example Michigan may have planned on 4 really cool blitzes that they thought would work against the team they're playing, and the fewer plays the defense runs the larger % of total plays are those special ones they drew up (which won't be as effective once they're used up). The more plays a defense faces, the greater % of those plays are going to be the base defense, and the more comfortable the offense will be with defeating it.
In Michigan's case I think our amazing depth everywhere but linebacker has made our defense especially capable of performing even as long drives persist. How many times this year has a drive gotten a key 3rd down conversion or something due to dumbass luck or a crappy penalty or whatever, then the defense shuts that down too? Our players don't get as tired because Michigan has the luxury of guys to rotate in who can perform as well as the starters.
October 11th, 2015 at 4:32 PM ^
since Michigan has been blowing teams out in the first half and coasting to the finish line that doesn't apply. If Michigan was winning narrow games in the fourth quarter there could be a case. That's not really what's happening.
October 11th, 2015 at 8:19 PM ^
Exactly. TOP is probably a telling stat specifically for teams like Michigan most of the time. However, even in those cases, other stats are more telling. If TOP is important for Michigan because having an advantage there means our run game is successfully marching down the field 7 yards at a time, then why not just look at stats about our running game?
Also, the OP keeps missing some huge factors in TOP. Yes, we might have a TOP advantage because we're sustaining drives and the other team isn't. But there are other factors. M hitting a deep bomb is going to ruin our TOP for that drive but is clearly a good thing for the team. Any defensive turnover deep in the opponents' territory or brought back for a TD or a special teams TD hurts our TOP also, despite clearly being great plays.
TL;DR version: Chesson's TD kickoff return hurt our TOP. That shows how useless a stat it is right there.
October 11th, 2015 at 2:39 PM ^
October 11th, 2015 at 2:40 PM ^
October 11th, 2015 at 3:03 PM ^
This. It's all about the number of snaps. Michigan wants to have more snaps (more opportunities) to score than the opponent. This becomes even more important against really good offenses, where really want to minimize the number of plays they run.
We can limit it in several ways:
-Slowing our offensive tempo
-Making our first downs
-Creating turnovers
note how all three result in increased Time of Posession.
October 11th, 2015 at 3:48 PM ^
October 11th, 2015 at 8:47 PM ^
October 11th, 2015 at 2:40 PM ^
TOP is a response variable, not a predictor.
You don't set out and say "We want to run the ball, and have a defense that forces a lot of three and outs.... Therefore we must crank up our time of possession to do so."
You set out and say "We want to keep the ball away from the other team as much as possible to limit their scoring oportunities.... Therefore we need to focus on running the ball and having a strong defense that forces three and outs and turnovers."
As the great 21st century philosopher Jack White once said "You can't take the effect and make it the cause."
October 11th, 2015 at 4:50 PM ^
October 11th, 2015 at 2:44 PM ^
I actually was wondering a simliar thing but on the use of Total tackles as a statistic and a correclation between both good pass defenses (with incompletions as another potential axis) and use them as potentialmetric.. If I had the time, I'd probably be able to devote a diary, but that likely won't happen at least for the next couple weeks if I have free time then. Anyways, Michigan is 7th in the country with fewest total tackles at 304. Teams above them are Boston College, Florida State, LSU, Syracuse, Rutgers, and Pittsburgh leading the pack with only 272 tackles so far this season. Obviously if a teams runs a lot of plays towards the sidelines, a run out of bounds doesn't count towards a tackle. But even there seems to be some inverse correlation between Total tackles made and how good a defense is. This also likely plays into the Time of Possession metric as well.
October 11th, 2015 at 2:55 PM ^
October 11th, 2015 at 3:04 PM ^
and ability to control the ball via the running game. It's an indicator of ball control and not a determinant of victory.
In the spread offense, it's an irrelevant issue primarily because scoring efficiency is a higher aim than controlling the ball and tempo of the contest. If your defense is dominant, and you run the ball effectively you limit your opponents possessions and that limit is reflected by the amount of time they had the ball to score.
If your team scores a preponderance of its points through special teams play and defense, the time your offense has the ball is not a significant factor.
For Michigan, TOP is a relevant stat because it's a factor in both aims of this team, suffocating defense and grinding offense.
October 11th, 2015 at 3:25 PM ^
October 11th, 2015 at 3:40 PM ^
October 11th, 2015 at 4:16 PM ^
As said, TOP is a symptom not a cause. But even as symptoms go, I'd suggest that total plays is a better measure than time of possession. If you're talking about wearing down a defense, I don't think there's much difference between running the play clock way down between snaps and going quickly. The physical strain comes during the play, not on the minutes that are ticking away between them. But a greater quantity of plays defended is going to tend to wear down a defense more, yes.
October 11th, 2015 at 4:30 PM ^
(via)
You are making the classic blunder of those who use TOP as a metric: you're supposing that the cart is pulling the horse.
Teams that win time of possession consistently win football games for a very good reason: being ahead by a multi-score margin for most of the game means teams will try to burn clock as part of the strategy to limit possessions and therefore opportunities for the other team to score. Teams that are behind will do the opposite: they'll try to get either chunk yardage, out of bounds, first downs or incompletes in order to preserve clock and maximize possessions.
Put it this way: if you're down two touchdowns, do you try to "win time of possession" by going for 4- and 5-yard runs and letting the clock run down so your defense is fresher? Only if you're Kirk Ferentz. Time of Possession is therefore a mostly useless stat, since it is just measuring which team was trying to run the clock out, which is almost always the winning team.
You COULD create a statistic of use that is situational, however. IE you could track how much time a team runs off the clock when they're up by two or more scores in the 4th quarter or, like, at least 21 points before that, as a measure of their ability to put games away. But that should be separated out from other stats and used only to judge a team for when it's trying to salt away versus when its offense is operating normally.
It would be interesting to know things like how much time off the clock in a 4th quarter scenario is equal in value to a touchdown immediately. But since the value of a possession that starts on your own 25 is far less than that of a touchdown, the number of possessions you would have to remove from the game to make it equal to a touchdown would have be to rather large. I bet if you're up 17 points at the beginning of the 4th quarter, a 12-minute non-scoring drive or a 10-minute field goal drive is about equal to a touchdown drive that scores immediately. In other words, scoring points wins games, not playing keepaway.
October 11th, 2015 at 4:57 PM ^
October 11th, 2015 at 5:07 PM ^
The chance of converting those 3rd and longs was actually not very good for Michigan's favor, but the 70 seconds run off the clock on 1st and 2nd down did marginally improve Michigan's chances of winning. Given how good Michigan's defense was that year, it's hard to point to any 40 second runoff with zero yards that would be more valuable than Michigan's base offensive performance.
October 11th, 2015 at 5:51 PM ^
While there is some correlation, I get that TOP is a result or byproduct of doing the right things. Specifically, I think the metric of "plays run" is probably a more valuable one. The key is wearing out and demoralizing an opponent. TOP isn't really it, especially if another team has a quick strike offense. More important is that you're keeping the ball out of their hands.
What brought this to mind yesterday was the long, sustained drive (7 plus minutes?) culminating with a Green TD. In this particular instance, it wasn't about winning . . . it was about keeping NW off the field to preserve the shutout.
Regardless, I get what all of you are saying at a deeper level than I did before.
October 11th, 2015 at 6:15 PM ^
TOP is a by-product of our success so far, not a determiner of it.
Our D forces a lot of 3 and outs. Our O is steady, but not explosive. Those things add up to a TOP advantage.
I don't think there is a deliberate or necessary attempt to slow things down to "protect our Defense". We are not playing to keep the other team's offense off the field. If our O was capable of some more quick explosive scores, our coaches would gladly take that.