Bye Week Advanced Stats: By the Down

The Hogs are on bye this week, but the team is hard at work figuring out what went wrong offensively against TCU and how to fix it, so I thought we could do some of the same.

Here’s a question I had after the game: what went worse, early downs or third downs? The Hogs were just 4 of 14 on third downs, but that number rarely tells the whole story. Imagine two scenarios: in the first, the Hogs face a 3rd-and-1 and get stuffed at the line. In the second, the Hogs face a 3rd-and-14 and throw a 12-yard pass. In both cases, the offense failed, but only in the first case can you really blame the third-down offense for what went wrong.

In order to break it down further, we’re meeting two new advanced stats: leverage rate and adjusted third down conversion rate. What I like about these two is that they are completely independent of each other; one helps explain how well the offense avoids must-pass situations, while the other helps explain how well they do on third down regardless of what happens on first and second down.

Leverage rate

One of the objectives of an efficiency-based offense like Arkansas’ is to stay on schedule. In order to measure this, we break all plays up into standard downs and passing downs. Most plays are standard downs. Here’s the definition of a passing down:

• 2nd and 8+
• 3rd and 4+
• 4th and 4+
• Any down/distance when trailing by 1 score with less than 2 minutes left in the game

Obviously, the offense should seek to avoid such situations. Enter leverage rate: the percentage of total plays that come on standard downs. The higher the number, the more on-schedule the offense remained. Since the vast majority of an offense’s plays come on first or second down, this is a good indicator of early-down success; that is, how well the offense avoided the dreaded third-and-long.

Here’s the data from Saturday:

About 65-70% is normal. Arkansas was on the low end of the normal range, while TCU had a good day avoiding must-pass situations. But the difference isn’t huge, suggesting that Arkansas didn’t lose this game on first and second down.

Let’s go further. Early down conversion rate measures the percentage of made first downs that come before third down. It’s not a particularly reliable stat because it’s influenced by so many factors, but here it is:

Notice the end result: TCU averaged a half-yard less to go on third down than Arkansas (6.9 to 6.4). That certainly helps TCU, but that isn’t a huge difference. It doesn’t scream 28-7. The Hogs could easily have overcome that.

Arkansas actually won first down:

Notice the pass numbers: just a 14% success rate, yet positive EV? That’s because of Austin Allen’s 49-yard touchdown pass to Jonathan Nance. On other first down passes, Allen was 2 of 6 for minus-2 yards.

Nance’s catch earned him Offensive Player of the Game honors. He contributed 4 points (EV) to the Hog offense, easily the most of any player:

I’ve copied this idea from MGoBlog, and they report these figures each game as “Nance +4” and “Allen -6” in order to provide a shorthand version of how each player performed. I’ll keep track of player EV for the entire season.

Adjusted third down conversion rate

Okay, so we’ve seen the early-downs data: the Hogs had a 65% leverage rate and averaged 6.9 yards to go on third down. The leverage rate is ideally a bit higher, but not terrible. So how’d the Hogs do on third down?

Back to the example at the top of the 3rd-and-1 versus the 3rd-and-14, it’s important to understand that how well a team does on third down often depends on the situations the offense is put in based on earlier downs. The easiest way to do this is to compare a team’s actual third down conversation rate to some kind of projected third down conversion rate, basically to see how well the team compared to how a normal team would do.

Thanks to a wealth of data, we can do this. We know, for example, that 3rd-and-1 is converted 73% of the time in the FBS. So if a team faces 10 3rd-and-1s in their game, they should convert about 7 of them. Plugging this data in for all 14 of Arkansas’ third down attempts will provide a projected third down conversion rate that we can then compare to the actual.

The difference between projected and actual, then, is the adjusted third down conversion rate.

And there’s your game. Arkansas should have converted 6 of its 14 chances: instead it converted 4. TCU should have converted 7 of its 14 chances: instead it converted 10. That massive swing of 5 conversions was too much to overcome. TCU’s two turnovers more than canceled out the fact that the Frogs were slightly better on early downs, but Arkansas going minus-37% in adjusted third downs plus two missed field goals sealed the deal.

The defense wasn’t as disruptive as is ideal, but they played well given the circumstances. Half (5) of TCU’s third down conversions were 3rd-and-1.

Also notice the distances. The Hogs struggled most on third-and-medium, which was actually a strength in 2015 and 2016. In 2015, the Hogs were 2nd in the FBS in Passing Down Success Rate. In 2016, they were 11th. So far this year? 63rd, and that’s with the Florida A&M included in the numbers. That’s not going to cut it, especially when the run game isn’t where it needs to be.

Struggles on third-and-pass downs indicate a lack of rapport between a quarterback and his top target, which makes sense given turnover at receiver. In 2015, Brandon Allen favored Hunter Henry (and Drew Morgan) on pass downs. Last year, Austin seemed to like Keon Hatcher when he needed someone to catch the ball. Against TCU, O’Grady, Deon Stewart, Nance, Jeremy Patton, and Austin Cantrell all received third-down targets, with Nance and Stewart (2 each) the only guys to be targeted multiple times. Basically, Allen has little faith in his receivers, which is why he’s spreading it out. Also, zero targets for Jared Cornelius isn’t good. The lack of work with Allen this fall may be to blame. He needs to get 100% healthy and on the same page with Allen, and fast. Bielema has also said the staff may trim the number of receivers who get reps with the 1s. That’s a good move, but Coach, once again it’s September and we’re still making tweaks that should have been made during fall camp.

Up next: Texas A&M. A Hog win would help get things back on track as the SEC schedule begins to heat up.