Rowland: 10 thoughts on advanced statistics that shed light on UK football

Justin Rowland, Publisher
Cats Illustrated

We're only three games into Kentucky's 2017 season so stats have a limited value. It's very easy to weigh some numbers too heavily or fail to weigh other numbers heavily enough. And some numbers just won't hold over the long haul.

Nonetheless, snapshots in time are all we have to go on, so let's break things down.

GO PREMIUM, GET FREE MONTHS ADDED

Oonsebweylsuh1wwsl92
Oonsebweylsuh1wwsl92

Stephen Johnson's Total QBR against South Carolina was 92.8.

Total QBR, as we mentioned on the site yesterday, is one rating system that ESPN has used dating back to 2014 to rate a quarterback's performance over the course of a game or a season. It factors in a range of things, from the impact of a quarterback's running to his passing relative to replacement levels, and from value lost due to sacks to total action plays the player accounted for.

In short, 92.8 is a very good score. Through three games Johnson's Total QBR ranks No. 10 nationally out of all quarterbacks (84.3). For context, Andre Woodson's 2007 Total QBR was 67.6 (21st in the country) and that was during a season in which he passed for 40 touchdowns and 3,700 yards. Mike Hartline's 2010 Total QBR was 75.6.

That's not to say Johnson is a better quarterback than Woodson, Hartline or either of the two (you can debate that and form your own conclusions, I know better than to tell a Kentucky fan what to think about a quarterback wearing blue and white). And certain rating systems will make certain kinds of quarterbacks look better or worse. It's probably fair to say the Total QBR system is weighted in such a way that it is, overall, favorable to Johnson. For instance, he's not losing a lot of points for sacks (that hurts some other QB's in different systems), he's getting points for rushing yards, and he's not attempting a lot of passes (which also helps the number).

But regardless of all that, the Total QBR number should at least give everyone pause and lead to an appreciation of what Johnson is doing.

Oh, and Stephen Johnson's best Total QBR game for his career? Louisville, 2016. 97.8. That was the best single-game QB performance by any Kentucky quarterback since 2004, when the system came into existence.

Football Outsiders' F/+ rankings give Kentucky a 79% chance of making a bowl game.

According to the website, which uses advanced statistics across a wide range of categories to measure a team's strength, Kentucky is most likely to finish the 2017 season with seven wins. Here's their full projection chart:

0-2 wins: 0% (Kentucky already has three wins, hence two is impossible)

3 wins: 1%

4 wins: 5%

5 wins: 15%

6 wins: 26%

7 wins: 27%

8 wins: 17%

9 wins: 7%

10 wins: 2%

11-12 wins: 0%

Putting aside the possible objection that 11 or 12 wins is statistically impossible (because it is very unlikely, optimists), take a closer look at those numbers. Does Kentucky really have a 21% chance of not being bowl eligible even though the Cats are already 3-0? They only need to go 3-6 the rest of the way. That might seem like a high number but if you consider Kentucky started 5-1 and 4-2 in different years under Mark Stoops and failed to make a bowl game it's not that difficult to understand.

These numbers say there's a 79% chance UK wins at least six games, a 53% chance UK wins at least seven games, and a 26% chance that Kentucky wins at least eight games.

I do think that last number is low and my explanation in the next section will explain why this particular model isn't as friendly to Kentucky as it might otherwise be.

P6mrg7vvyqpa71qm82kr
P6mrg7vvyqpa71qm82kr

Football Outsiders' F/+ system also gives Kentucky a better than 50% chance in only TWO games the rest of the season.

These rankings (and you can read the full Kentucky summary here) give Kentucky a 77% chance of beating Eastern Michigan. In every other game remaining of the 2017 schedule, Kentucky has less than a 50% chance of winning. Here's the full presentation:

Florida: 33% chance of a UK win

Eastern Michigan: 77%

Missouri: 60%

Mississippi State: 22%

Tennessee: 43%

Ole Miss: 38%

Vanderbilt: 36%

Georgia: 23%

Louisville: 33%

So if you're a Kentucky fan and you're looking at those numbers you are probably, A) Thinking, "That's way off," B) Starting to worry, or C) You're a negative person severely damaged by years of UK struggles.

In my estimation the numbers above don't give Kentucky the credit they probably should be getting, but I understand why. The Southern Miss game was a statistical anomaly in that Kentucky was significantly outgained by the Golden Eagles. Those numbers are still impacting the computer models in a significant way. For instance, the F/+ model says Kentucky (or any team) only wins that Southern Miss game 7% of the time (that's 7 times out of 100) when the stats were as slanted to Southern Miss as they were.

But anyone who watched Kentucky-Southern Miss game knows the Cats basically controlled that game from the jump and the yardage really didn't matter, or at least never caused the game to be in serious doubt save for a small stretch when it was 14-10.


So what odds would I give Kentucky in each of their remaining games this season?

My own perception is just as prone to limitations and faulty weightings as a computer model that's fixed on one statistical anomaly.

Nonetheless, I would give Kentucky the following chances in each of its remaining games:

Florida: 51%

Eastern Michigan: 93%

Missouri: 75%

Mississippi State: 30%

Tennessee: 50%

Ole Miss: 55%

Vanderbilt: 50%

Georgia: 23% (I'll agree with the computer)

Louisville: 52%

That means I think Kentucky's toughest games from here on out are, in order of greatest difficult to least: Georgia, Mississippi State, Vanderbilt, Florida and Tennessee.

Where's the power?

Last year according to Football Outsiders' advanced offensive line (rushing) stats Kentucky's Power Success Rate was 6th in the country (80%). What does that mean? Four out of five times Kentucky had a 3rd or 4th down with 1 or 2 yards to go, they converted the short yardage situation. That rate is outstanding. When UK created short yardage third and fourth downs, they capitalized. They extended drives. They drove down the field.

This year Kentucky's power success rate through three games has plummeted to 57%, which is 95th in the country (down from 6th).

How concerning is that? Well, it depends on how much stock you put in the South Carolina game, when Kentucky's third down fortunes were reversed and the Cats largely dominated both sides of the ball in short yardage situations. If you believed that's an emerging trend or a sign of what's to come, then the 95th Power Success ranking will improve dramatically and the current ranking is just a poor small sample product.

Twio0leycqq2zsxisbwb
Twio0leycqq2zsxisbwb

But on the flip side, Kentucky has been stout. And I mean stout.

Kentucky's "opportunity rate" run defense ranks No. 4 in the entire country. Opportunity rate is the percentage of carries that gain at least five yards. UK's opportunity rate on defense is 24.7-percent, which means that less than one-quarter of the time have opponents rush attempts gone for five yards or more.

Last year Kentucky's run defense opportunity rate was 48.5-percent and that ranked 127th in the country.

In short, Kentucky's power run game for the whole season has been dramatically worse than it was a year ago...but Kentucky's run defense has more than made up for the difference.

It is my strong opinion that things even out for Kentucky on both sides over the course of the year (extremes rarely hold).

Pxk3h0wf1xzuwetgmh0i
Pxk3h0wf1xzuwetgmh0i

Sihiem King is making things happen

Sihiem King might be the clear number two behind Benny Snell but in at least one category he's Kentucky's No. 1. In the "Highlight Yards" category (that is, yards that are credited exclusively to the running back (see: definition here), King is easily Kentucky's No. 1 this year. He averages 6.1 "highlight yards" every time the offensive line takes care of its job up front, and that's 6.1 yards on top of what the line gets him. For comparison's sake, Benny Snell is averaging 2.6 "highlight yards" per opportunity.


Qtcvhzkoks21ttpzpfyp
Qtcvhzkoks21ttpzpfyp

Now some statistics relevant to Kentucky's passing game thus far

— Who has Stephen Johnson targeted the most this season?

Garrett Johnson: 22.1%

Tavin Richardson: 13.2%

C.J. Conrad: 11.8%

Charles Walker: 11.8%

Blake Bone: 10.3%

Kayaune Ross: 8.8%

Sihiem King: 5.9%

Benny Snell: 4.4%

Thus, Garrett Johnson has been Stephen Johnson's preferred target and that is not a surprise.

In order of success rates (i.e. targets that have gone for catches):

Conrad: 75%

Walker: 62.5%

Bone: 57.1%

Johnson: 53.3%

Richardson: 44.4%

King: 50%

Ross: 33.3%

But most significant is the yards per target statistic. This is self-explanatory. How many yards does Kentucky's average per pass attempt to a given receiver?

Conrad: 20.1 YPT

Bone: 14.6 YPT

Walker: 9 YPT

King: 7.8 YPT

Johnson: 7 YPT

Snell: 6 YPT

Richardson: 4.9 YPT

Ross: 3.7 YPT

Charles Walker is good at what he does on punt returns.

Fans might be excited to see what Lynn Bowden can do on punt returns but through eight punt return situations Charles Walker has Kentucky ranked No. 1 in the nation with a 100-percent punt return success rate.

What exactly does that mean? Well, to make a long story short, punt return efficiency is measured by, "average values generated per punt based on the field position of the punt team and the field position at the conclusion of the play."

That's obviously not all Walker and has to do with the team punting the ball. But even though Walker has -1 punt return yards on one attempt with seven fair catches, he's putting Kentucky in a good position on the field relative to where punts are taking off and landing.

Kda2soqax2y9pwrpxwba
Kda2soqax2y9pwrpxwba

Other notable advanced stats through three games

— Kentucky has been much more likely to run the ball on "passing downs" than most other teams in the country. The national average is 33.8-percent runs on pass downs. Kentucky has run the ball 46.9-percent on passing downs. Digging deeper, Kentucky has run the ball 17 times on 3rd and 4 or more already this season. They are averaging 7.33 yards per carry when they run the ball in 3rd and 4-6 yard situations (which is very good; that means they're averaging a first down in those situations), but only 6.5 yards per carry in 3rd and 7-9 situations.

What does this mean?

Kentucky obviously does not trust the efficiency and success of its passing game when Eddie Gran knows that defenses are expecting a pass. He is calculating that Kentucky has a better chance of running the ball for a first down in 3rd and medium/long, quite often, as opposed to passing for it in those situations.

Given that Kentucky's sack rate rates very high nationally, and Johnson is a mobile quarterback, it stands to reason that he does not have a great amount of confidence in the Cats' receivers in those situations.

— In 2016, Denzil Ware was one of the SEC's leading "stuff" men on defense. A "stuff" is defined as a run that is stopped at or behind the line of scrimmage (that includes TFL's but also no gainers). Surprisingly, Kentucky's leader in "stuffs" this year is cornerback Derrick Baity. He has three through three games.

— Ordinarily it's not a good thing if your team's leading tacklers are a free safety and a cornerback. But in Kentucky's situation, don't judge too quickly. Darius West (19) and Derrick Baity (16.5) are Kentucky's leading tacklers through three games.

But let's dig into that a little bit. Where are those tackles taking place? If a free safety is making tackles deep in the secondary, that's not a good sign for the defense. The "success rate" stat for defensive players is helpful here. The lower the success rate number for a defender, the better. A lower number means that more often, a defender is making his tackles, on average, closer to the line of scrimmage.

Here are Kentucky's defensive leaders by success rate (i.e. those players whose average tackle is closest to the line of scrimmage). We'll limit qualifying players to those with 4 or more tackles.

Naquez Pringle: 0.0%

Adrian Middleton: 14.3%

Matt Elam: 20%

Jordan Jones: 23.1%

Josh Allen: 31.3%

Courtney Love: 34.8%

Denzil Ware: 42.9%

You would expect defensive linemen then linebackers and then defensive backs to rate in that order because of how they are aligned.

The only outliers are Mike Edwards (his 58.3% is very good for a safety), Kendall Randolph (his 57.1% is also good but not surprising for a nickel) and Chris Westry (his 70.0% is notably worse than counterpart Baity's 57.9%).

It's also surprising that sophomore defensive lineman T.J. Carter only has 2.5 tackles through three games considering the extensive game action he saw last year. But the line has played so well and rotated so many players that can't be regarded as a real failure until a larger sample is available.

Lanttcc55ifjjqlbwozx
Lanttcc55ifjjqlbwozx

What to Read Next