End of the Regular Season: Let's Take a Look at the Numbers

[walksalone]

All the games with below average rush attempts, we either lost or were very close to losing...

 

[I am I]

[i'm also curious as to what the longest streak of run plays called was this season (in a situation where the game wasn't yet decided]

Yeah, that's the Full Beck I think people think "he goes to". But 'Id venture that he was pretty consistent this year. Didn't seem to have a "Wisconsin before half" this year.

 

[It'sNotAFakeID]

The problem is you're taking all of these statistics out of context. What do the numbers say when the game score is equal- +- one score? What about interaction effects with score difference and time remaining, I bet you'd get more than just a main effect. What about proportion of run to pass compared to number of plays run?

Teams throw more when they're behind. Teams run more when they're ahead. You try to account for this anecdotally, but who are you to say when time management was important or not? You also fail to account for how you determined significance.

I appreciate what you're trying to do, but you're creating a strawman argument.

I note the problems with the analysis. Several times, in fact. I think this line came up a few times: "the argument that can be made that we had to pass more because we were behind and you need to preserve clock if you are to come back. The running game doesn't preserve clock, so you must pass."

I mention that it's time consuming to do so. But I've got a break coming up, so maybe I can give a more in depth look. Perhaps there is an interaction between the two groups and say, the score difference in the game. If so, the main effects are obviously misleading and the stuff above means jack sh*t.

Thanks!

 

[Huskerzoo]

The problem is you're taking all of these statistics out of context. What do the numbers say when the game score is equal- +- one score? What about interaction effects with score difference and time remaining, I bet you'd get more than just a main effect. What about proportion of run to pass compared to number of plays run?

Teams throw more when they're behind. Teams run more when they're ahead. You try to account for this anecdotally, but who are you to say when time management was important or not? You also fail to account for how you determined significance.

I appreciate what you're trying to do, but you're creating a strawman argument.

I note the problems with the analysis. Several times, in fact. I think this line came up a few times: "the argument that can be made that we had to pass more because we were behind and you need to preserve clock if you are to come back. The running game doesn't preserve clock, so you must pass."

I mention that it's time consuming to do so. But I've got a break coming up, so maybe I can give a more in depth look. Perhaps there is an interaction between the two groups and say, the score difference in the game. If so, the main effects are obviously misleading and the stuff above means jack sh*t.

Thanks!

You using SPSS? If so, I'd be happy to play a bit with it as well, but couldn't do it as much until January.

 

[RedDenver]

Thanks for the extensive analysis! I'm not sure what conclusions can really be drawn from the stats here, but I'm still looking through them.

When you say the results were "statistically significant", what do you mean? Is that just your opinion or is there a specific stat test you're using?

 

[It'sNotAFakeID]

The problem is you're taking all of these statistics out of context. What do the numbers say when the game score is equal- +- one score? What about interaction effects with score difference and time remaining, I bet you'd get more than just a main effect. What about proportion of run to pass compared to number of plays run?

Teams throw more when they're behind. Teams run more when they're ahead. You try to account for this anecdotally, but who are you to say when time management was important or not? You also fail to account for how you determined significance.

I appreciate what you're trying to do, but you're creating a strawman argument.

I note the problems with the analysis. Several times, in fact. I think this line came up a few times: "the argument that can be made that we had to pass more because we were behind and you need to preserve clock if you are to come back. The running game doesn't preserve clock, so you must pass."

I mention that it's time consuming to do so. But I've got a break coming up, so maybe I can give a more in depth look. Perhaps there is an interaction between the two groups and say, the score difference in the game. If so, the main effects are obviously misleading and the stuff above means jack sh*t.

Thanks!

You using SPSS? If so, I'd be happy to play a bit with it as well, but couldn't do it as much until January.

I am using SPSS.

Really, it's just taking a look at this season. Although, if we wanted to, we could take a look at Tim Beck's time as the signal caller here. Has he been here two or three years? If it's two, if you wanted to do last season then that would be wonderful.

Here's what I'm planning on doing:

ESPN has play by play information for each game. I'm separating by game situation, such as tied, up 1 score, up 2 scores, up 3+ scores, down 1 score, down 2 scores, and down 3+ scores. 1 score is anywhere from 1-8, 2 scores is anywhere from 9-16, and 3 scores is anywhere from 17 up. I'm then keeping track of:

Rush attempts

Rush yards

Yards per carry

Rush attempt to touchdown % (i.e. if we ran the ball 10 times and had two rushing TDs, this number would be 20%)

Pass attempts

Pass completions

Completion %

Pass yards

Yards per completion

Pass completion to touchdown % (same as the rushing)

Turnovers

Penalty yards

for both NU and the opponent. All opponent stats are placed with respect to where Nebraska stands in the game. So if Nebraska is down 7, all plays that happen while Nebraska is down 7 go under the down 1 score category.

 

[RedDenver]

I am using SPSS.

Really, it's just taking a look at this season. Although, if we wanted to, we could take a look at Tim Beck's time as the signal caller here. Has he been here two or three years? If it's two, if you wanted to do last season then that would be wonderful.

Here's what I'm planning on doing:

ESPN has play by play information for each game. I'm separating by game situation, such as tied, up 1 score, up 2 scores, up 3+ scores, down 1 score, down 2 scores, and down 3+ scores. 1 score is anywhere from 1-8, 2 scores is anywhere from 9-16, and 3 scores is anywhere from 17 up. I'm then keeping track of:

Rush attempts

Rush yards

Yards per carry

Rush attempt to touchdown % (i.e. if we ran the ball 10 times and had two rushing TDs, this number would be 20%)

Pass attempts

Pass completions

Completion %

Pass yards

Yards per completion

Pass completion to touchdown % (same as the rushing)

Turnovers

Penalty yards

for both NU and the opponent. All opponent stats are placed with respect to where Nebraska stands in the game. So if Nebraska is down 7, all plays that happen while Nebraska is down 7 go under the down 1 score category.

Wow. That's quite the under taking. Thanks for posting what you've done.

 

[It'sNotAFakeID]

Thanks for the extensive analysis! I'm not sure what conclusions can really be drawn from the stats here, but I'm still looking through them.

When you say the results were "statistically significant", what do you mean? Is that just your opinion or is there a specific stat test you're using?

The test is an F-test which measures effect/error. The significance test is a p-value which tests if that F value is significantly greater than 0. I'm using .05 as that cutoff. So if p is less than .05, then the F is significantly greater than 0 and there is an effect. There's less than a 5% chance of committing a Type I error or a false alarm: saying there's an effect when there really isn't an effect.

But never mind that above post. Huskerzoo pointed out that I should have considered how far Nebraska was ahead or behind. Which is true, teams tend to run more when they are farther ahead and so the difference in the score is a confounding variable. What I'm doing now is controlling for it by using it as a second independent variable and taking a look at the interaction between the outcome of the game and how we called plays when we were in various situations.

I may have 2013's season done by later tonight.

 

[Enhance]

I think if you look at Beck's pass/run ratio, you don't really see anything that pops out as poor play-calling. Where Beck struggles, and this is arguably the most challenging part of being an OC, is knowing what plays to call in what situations. This is where I think Beck needs the most improvement. There was one game I remember from earlier in this year where, after what I believe was a successful touchdown drive featuring quite a bit of downhill running, we came out the next drive calling a reverse and a couple of pass plays. We then punted, IIRC. All I know for sure is that whatever we had been doing that was successful, we got away from it.

Now, you obviously have to keep the defense on it's toes. You also have to be creative as an OC and mix things up. But, I sometimes feel we do too much, and it shows. There are plenty of examples I could spout off, but one of my biggest gripes is repetition and getting good at certain things. Do we have enough repetitions in practice to get everybody on the same page with all the different wrinkles to our offense? Sometimes I wonder if we do. IMHO there has to be certain parts of every offense that the offense can rely on in tough situations. What are we really good at and what do we feel confident will work when our backs are against the wall? I feel like Nebraska doesn't know what that is sometimes.

 

[RedDenver]

Thanks for the extensive analysis! I'm not sure what conclusions can really be drawn from the stats here, but I'm still looking through them.

When you say the results were "statistically significant", what do you mean? Is that just your opinion or is there a specific stat test you're using?

The test is an F-test which measures effect/error. The significance test is a p-value which tests if that F value is significantly greater than 0. I'm using .05 as that cutoff. So if p is less than .05, then the F is significantly greater than 0 and there is an effect. There's less than a 5% chance of committing a Type I error or a false alarm: saying there's an effect when there really isn't an effect.

But never mind that above post. Huskerzoo pointed out that I should have considered how far Nebraska was ahead or behind. Which is true, teams tend to run more when they are farther ahead and so the difference in the score is a confounding variable. What I'm doing now is controlling for it by using it as a second independent variable and taking a look at the interaction between the outcome of the game and how we called plays when we were in various situations.

I may have 2013's season done by later tonight.

Don't F-tests test the likelihood of data being being drawn from some distribution? Are you comparing the run/pass plays from each game against a normal distribution with the mean and variance based on the plays for the whole season?

I'll take a look at the new results when you post them. Also, if you're looking at confounding variables, then down and distance might be just as significant as score margin.

 

[It'sNotAFakeID]

Thanks for the extensive analysis! I'm not sure what conclusions can really be drawn from the stats here, but I'm still looking through them.

When you say the results were "statistically significant", what do you mean? Is that just your opinion or is there a specific stat test you're using?

The test is an F-test which measures effect/error. The significance test is a p-value which tests if that F value is significantly greater than 0. I'm using .05 as that cutoff. So if p is less than .05, then the F is significantly greater than 0 and there is an effect. There's less than a 5% chance of committing a Type I error or a false alarm: saying there's an effect when there really isn't an effect.

But never mind that above post. Huskerzoo pointed out that I should have considered how far Nebraska was ahead or behind. Which is true, teams tend to run more when they are farther ahead and so the difference in the score is a confounding variable. What I'm doing now is controlling for it by using it as a second independent variable and taking a look at the interaction between the outcome of the game and how we called plays when we were in various situations.

I may have 2013's season done by later tonight.

Don't F-tests test the likelihood of data being being drawn from some distribution? Are you comparing the run/pass plays from each game against a normal distribution with the mean and variance based on the plays for the whole season?

I'll take a look at the new results when you post them. Also, if you're looking at confounding variables, then down and distance might be just as significant as score margin.

Exactly, and that'll be another thing I'll look at. But for now I'm looking at scoring margin.

 

[Guy Chamberlin]

You don't have to look much farther than turnovers to realize that the best laid plans of an offensive coordinator fall apart when you no longer have the ball.

We had one of the best seasons for a running back in Husker history, and two of the best wide receivers we've ever had. And they had a lot of great plays. I just think that it's hard for an offense to gel with three different quarterback, crucial turnovers, an AWOL defense the first half-season and your own special teams pinning you inside the ten. I'm not even mentioning the offensive line injuries, because I thought the reserves actually stepped up pretty big against some tough Big 10 defenses. Abdullah got his yards, and protection for Armstrong and Kellogg didn't appear to be the problem.

For all our turnovers and QB inexperience, I thought our offense kept its cool much of the season, and kept us in contention.

 

[Enhance]

With the turnovers, it's important to add this. There's a veteran on this team that had a turnover in every single game Nebraska lost this year. It was, also, the most valuable player to the team.

Ameer Abdullah.

It's not just a quarterback inconsistency thing, though, that definitely will create miscommunication and lead to turnovers at times. It's a systemic failure on account of the whole team. Nebraska fails miserably at securing the ball and creating turnovers on a consistent basis. It even shows up in the most valuable player.

 

[walksalone]

[i'm also curious as to what the longest streak of run plays called was this season (in a situation where the game wasn't yet decided]

Yeah, that's the Full Beck I think people think "he goes to". But 'Id venture that he was pretty consistent this year. Didn't seem to have a "Wisconsin before half" this year.

Which is also why I'm baffled as to why there weren't more run plays called. I know the line got beat up as the year went on, but we had 3 backs that could have been rotated to try to keep them as "fresh" as possible. And concievably, wouldn't running the ball eat a little more clock, giving the D a chance to catch a breath.

 

[Guy Chamberlin]

The Nebraska offense had 541 rushes and 4.9 yards per attempt.

and 362 passes for 6.6 yards per attempt.

While you may not call Beck a "run first" OC, we ran considerably more than we passed, and generally stayed with it when it was working. Since our pass plays averaged more yards, and we fumbled as much as we threw interceptions, it's hard to argue that it was either daring or reckless to pass as much as we did.

To really analyze this you'd have to consider games like UCLA and Minnesota where Beck called consecutive passing sequences in successful first half scoring drives. Good rushing plays were mixed in, too. I don't think there was often an "abandoning" of the running game as much as opposing defenses made second half adjustments and neither facet of the offense was working as designed. Yet a certain segment is frustrated by two incomplete passes, and for some reason doesn't consider two failed rushing plays to be the same play-calling problem. An OC has to keep a defense guessing, especially when the offensive line isn't quite strong enough to announce its intentions and impose its will.

While the first few games of the season made me look forward to an Abdullah/Cross/Newby platoon for the rest of the year, it became hard to argue against Abdullah being given the ball as much as possible. Like Burkhead and the great NFL backs, Ameer seems to get stronger as the game goes on.