FanPost

1st Down Performance in Losses, Variance in Our Offense, and Other Matters

This google doc lists every offensive play that Georgia ran in its first five games last year. In other words, it shows our 1-4 start. The list covers 355 plays. Above the list of plays are two tables. One table shows UGA's yards per play and standard deviation for each of those first five games individually. It also shows the number of possessions we had in each game, and the number of three and outs we have.

The second table shows UGA's performance on every down for all five of those games combined, and then again for the four losses only (excluding the UL-Lafayette game, in other words).

Some quick analysis after the jump.

Of course, I'd like to get more games compiled into the sheet, not only for Georgia but for some other teams as well. But you'll notice that in each of those first five games, our standard deviation on yards per play was a little over 8. Whether we won big, lost a heartbreaker, or lost by two possessions when the opponent pulled away late; whether we averaged only 5 yards per play or 7 yards per play (hello, AJ Green!); no matter what, in each game our variance was the same. (Standard deviation is the square root of variance, so both numbers go up and down together). 

In the first four games (when we went 1-3, hammering ULL but losing to SC, Ark, and Miss St), notice that our yards per play hovered just above 5. It was at its best in the ULL game, but still wasn't great. (Perhaps that could have offered us some warning, that we blew out a nothing opponent 55-7, yet only averaged 5.6 yards per play?) Then, when AJ came back for the Colorado game we surged to over 7 yards per play (and, I'm sure, we maintained something up around there for most of our games afterward, excepting the Liberty Bowl). Yet no matter what, our variance was similar. We always had a standard deviation of about 8.5 in each of those games.  So, in the first 4 games where our yards per play was only about 5.5, that means that over 2/3 of our plays fell in the range of -3 to 14 yards. In the Colorado game, our yards per play jumped up to 7, but the standard deviation remained the same, so 2/3 of our plays fell in the range of -1.5 to 15.5 yards. 

I said in the other thread that I suspected that a typical standard deviation is something higher than Boise State's 5.9 from Saturday night, but way below the Dawgs' 13.75. With just these five games from one team having been compiled, I'm even more sure now that I'm right (the sample size is small, yes, but it also isn't really that small. The odds that a "normal" standard deviation is way off from 8.5, but our first five games just happened to cluster around that as an abnormal number, are pretty low). 

So, looking at four of our stinker games from last year, we had a similar yards per play to the Boise game, but a much lower and more normal standard deviation (variance). Saturday night was an anamoly, folks. We do not normally do the feast or famine thing like we did with Boise on Saturday night. What conclusions we should draw from that are up for discussion, but I don't think that the data is going to support the idea that we have traded points for a "lot" of 3-and-outs.

That said, do notice that every game can have pretty different results as far as three-and-outs are concerned. But *overall* you can't have a "normal" standard deviation and also have a whole bunch of 3 and outs. Over he course of a season, if you're gaining yards like Georgia gained (after AJ came back especially), and you have a standard deviation that is fairly "normal," then that provides you a statistical prediction regarding how many three and outs you're going to have. And it won't be many. But, the one's you *do* have may cluster in odd ways, because that's just how random variance works. All your bad plays don't follow good plays. Sometimes you have streaks, and that's true of *all* teams.

Looking at the 3nOut column, we notice a few things. First of all, the ULL game really should be 12 possessions and *zero* 3 and outs. All 3 of the three and outs came on the last three possessions of that game, in the 4th quarter, up 55 to 7, starters pulled, and largely running the ball with Munzenmeier over and over again. The bottom line is, we got 3 and outs way less than 20 percent of the time. Closer to 10 percent of the time. That's good!

The one exception was a very, very ugly exception. The South Carolina game. We only had 9 possessions, but we had *five* three and outs. It's no wonder why that game stands out as an exemplar of what's wrong, and particularly as an exemplar of the idea that our offense isn't giving our defense a chance. Because in that particular game, those things are undoubtedly true. 

But, again, there are going to be individual games where the stats "clump." That said, for so many to clump into a nine-possession stretch, as happened in the Carolina game, is really freaky. It's worthwhile to look at what we did in that game offensively that was *different* than normal, to see what might have gone wrong. But it doesn't really make sense to look at that game as an example of our *normal* kinds of problems, because that game was a far from normal result. 

(Also, we *completed* at least 2 "first and bombs" in that game, so 1st and bomb couldn't have been the problem.)

Looking over at the second table, our 1st down offensive numbers were *even better* in those first five games (and even better still in the four *losses*) than they were over the course of the year! In other words, we really bossed teams around on first down. I don't think there's anything "wrong" with our subsequent numbers on 2nd and 3rd down, either, though that's probably the next specific thing that should be looked into.

Our 1st down offense is so good that it *is* tempting to conclude that there's something *bad* about being good on 1st down. But that really can only be said as a joke, not as a serious concern. Scoring points correlates to wins, and gaining yards correlates to points. If you gain a lot of yards but are *also* losing games, then something is definitely wrong. And it may be something with your *offense*, too. Issues of timing, play-calling, toughness, etc., may all be holding you back from getting the score when you need it most, or something like that. That's reasonable. But what wouldn't be reasonable would be to somehow think that first down play-calling *is* itself the problem. I'm sorry, but there ain't nothing wrong with gaining 10 yards on first down! 

X
Log In Sign Up

forgot?
Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Dawg Sports

You must be a member of Dawg Sports to participate.

We have our own Community Guidelines at Dawg Sports. You should read them.

Join Dawg Sports

You must be a member of Dawg Sports to participate.

We have our own Community Guidelines at Dawg Sports. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker