Statistical Look at 2015 Texas Offense and the 2016 Version: Part 1 - Our Worst Stat

jerrod-heard-ncaa-football-rice-texas1-850x560.0.jpg

This is the first of two posts. In this Part 1, I'm going to discuss one of the key stats that explains why Texas was close to dead ass last in total offense in 2015. In the next post, I'm going to use statistical analysis around our stats this year and Gilbert’s history to make some educated (read: wild-ass) guesses at what our offensive production will be within 50 yards/game or so next year based on a simulating a few hundred thousand games.

If you need any charts to be at little bigger, right click and open the image in a new tab. At the end of this post I also included some bonus charts.

Part 1: The Worst Stat for the Texas Offense in 2015

Take a look at these FBS rankings of the 2015 Texas offense, which comprise the usual building blocks for measuring offensive efficiency (unadjusted for anything):

Yards per Play: 58th – 5.7 Yards per Play

Rushing Yards per Attempt: 21st – 5.1 Yards per Rush
Passing Yards per Attempt: 76th – 6.9 Yards per Pass Attempt
Points per Play: 68th
Turnovers Lost: 10th

I would say, while not great, those numbers are middle of the pack. In the last six years, we have only averaged more yards per play than we did this year one time (in 2012). To emphasize… this 2015 team averaged more yards per play than the team that went to the national championship in 2009-10. How then can we explain a ranking of #92 by yards per game, the second-worst yard per game numbers since 2009?

# of Offensive Plays per Game: 117th – 65 offensive plays per game

This stat reflected the undisputed #1 issue for the Texas offense this year, which is why I'm focusing on it. I realize that none of us are thrilled with the offensive efficiency in general (especially the passing game), but you have to plug the biggest leak first.

To put this in perspective, if Texas ran 10 more plays a game (giving us 3 more than the national average), using the existing efficiency numbers above it would have been right about the #45 ranked offense in total offense per game. Not great, but almost top third and cutting your ranking in half.

The two most obvious factors that help determine your play count are pace & converting 3rd downs.

Pace: So how about that up-tempo offense we were supposed to be installing?

It's obvious from the eyeball test, but it didn't really happen.

I took the overall number of plays per minute of T.O.P. and estimated the amount of time people took between ball being spotted and the snap, making an adjustment for incomplete pass clock stoppages and the time it takes to run a play.

This chart shows the Big 12 time between snaps... if you took 25 minus these numbers, it should approximate remaining time on the play clock. The two green bars to the right are Sterlin Gilbert's last two seasons (2015 with Tulsa and 2014 with Bowling Green).

[IMG]

We may be middle of the pack in the FBS, but we are only faster than Iowa State and Kansas State in our own conference. By the way, check out how fast Gilbert's last two teams were snapping the ball. This is the most obvious and predictable change we can expect in the offense (hint, hint: for the next predictive post).

Pace isn't everything, but it's a major controllable factor that is highly correlated to overall play count. There is not a single team who snaps it in under 15 seconds that had a below-average play count.

Just for fun, the simple math of snapping the ball 5 seconds faster on average (the difference between Texas and Tulsa) means you could get 15-20 extra plays per game (in a vacuum and 50/50 T.O.P.).

A final note on Pace: Some of you probably rightfully are thinking that playing at a quicker pace sounds stupid because it just means that our defense will face more plays. That’s likely to be the case, as there’s a very strong correlation. This post is about why our offense had bad total stats.

That said, when you have a situation like Texas did this year (and is likely to have next year) and you average more yards per play than the defense gives up per play, you absolutely want to create as many plays and as many drives in the game as you can. When you have an advantage on a "per play" basis, the optimal decision is to maximize the number of plays and drives in the game…that’s just math. Pace is the piece you can most easily control. More bites at the apple means per play efficiency should start to win out over things that have a little more randomness to them…like third downs…

3rd Down Conversions: What the heck was going on?

Most people realize this was an issue for the 2015 Texas offense. There is a degree of randomness in third down conversion rates. It is decently correlated to yards per play, but there are always outliers. This year, Texas was one of those outliers. Whether that means that our play-calling was abysmal (maybe), we were just unlucky (possible), or our players just suddenly forgot how to play football when the downs shifted to third down (less likely), it was another big factor in us not getting enough plays run on offense.

Third Down Conversion Percentage: 101st in the FBS
Percentage of Drives with a 3 and Out: 114th in the FBS
Plays per Drive: 122nd in the FBS

Those first two metrics aren’t as correlated as you may think. And 3 & Out percentage is actually probably worse than third down conversion rate, because it negates any opportunity to have a drive that goes longer than 3 plays. We averaged 4.8 offensive plays per drive, while the FBS average was 5.6. If we could just get the FBS average plays per drive that would have added over 11 plays per game.

Here is a little table that shows 3 groups against each other: 2015 Texas, Gilbert’s last two teams, and a few teams next to Texas in the third down conversion percentage basement.

Look in that bottom group. All of those teams, despite having some of worst third down percentages in the country, figured out a way to way run more plays per game than Texas. They either had less 3 & outs or played with more pace. The point is, you can be bad at some… but if you are bad at all of them, you’re screwed.

As for the "cause", it wasn’t the defenses we faced, so what was it (coaching, bad luck, or players shitting a brick)? Honestly, this is a hard one to peg without going to the tape on each one and making an assessment. In lieu of a single specific explanation I will just present you with a bunch of charts. On each, I put Texas vs. FBS average first, but also included Gilbert’s last two teams against the FBS in the same scenario.

Average Yards per Play by Down:

From the first chart you can see that our yards per play on first down were much better than on any other down. In fact, the difference between yards per play on first and third downs (1.6 for Texas) was the 15th largest in the FBS. Clearly, we weren’t an offensive marvel on third down. Gilbert’s last two teams on the other hand, saw an increase from first down to third down. Not necessarily predictive, but it doesn’t make you feel bad.

Conversion Percentage by 3rd Down Distance:

The first chart truly shows where Texas came up short on third down this year. FBS-average or so for 1-3 yards, 4-6 yards, and 10+ yards, but way below average on 7-9 yards. I’ll leave it to the reader to speculate why. Gilbert’s last two teams have basically been at or above average at every down and distance. Again, can’t guarantee that’s a trend that will continue, but it certainly shouldn’t worry you.

Play Selection by 3rd Down Distance:

It’s important to note that Texas ran the ball 67% of the time this year, so you would expect to see more runs, relatively speaking. That said, we ran the ball a huge amount of times in third and long situations. It’s also worth noting that all QB runs (including sacks) are counted in these numbers.

A final note on Third Downs: Historical analysis of college football says that third downs are one of those things that is just pretty unpredictable. Sure, there’s enough of a correlation with yards per play to predict it, but the range of potential outcomes is still so large that it’s not reliable. What makes that frustrating is that it determines so much of the outcome of a game or season for an offense (and tired defense).

Here is what I can tell you with total certainty:

Texas was an absolute outlier this year in terms of third down % vs. yards per play, whatever the cause may have been.
Sterlin Gilbert has never been the (co-)OC of a team that has done worse on third down than Texas was this year.

For now, that’s gotta be enough.

Conclusions/TL;DR:

Texas ran so few plays this year that while its yard per play average was actually very good compared to the recent era of Texas football, the total offense numbers were abysmal. Some of the factors to alleviate that are very controllable (pace), while some have an extra degree of complication to them (third downs). Regardless, there is reason to be optimistic, as I’ll cover in my next post predicting our yards/game next year.

Bonus charts:

[IMG]

(Interestingly, despite all the coaching adjustment complaints, we averaged more yards per play in the second half (5.9) than in the first half (5.5))

[IMG]