The Performance Gap
Part 2 in an Examination of the Virginia Tech Roster: On-field Performance
If you missed Part 1 in this series, I recommend going back and reading it before proceeding with this article.
As I mentioned in my analysis of the recruiting gap, this is my working formula for talent:
Talent = Potential (to perform at a consistently elite level) + Actual Performance
Today I’m going to do a deep dive on performance trends since 2015. There is a lot that goes into on-field performance. Coaching, game prep, strength and conditioning, nutrition, injury status, and experience are some of the key factors. Disentangling which factors are driving which results is not possible, given that many of those factors are not quantifiable. Still, we can draw some high level conclusions by looking at performance over time.
For this analysis I gathered PFF data spanning from 2015 to the present at a player level, focused only on the starters. I defined starters as the players who played the most snaps at each position. If two players were evenly split, I went with the player who was actually the intended first stringer for the majority of the season. For example, in 2022, Dorian Strong and Brion Murray have played essentially the same number of snaps. Strong was the clear starter before getting injured, so I included him in the dataset and excluded Murray. This analysis looks at starters only and offers no comments about depth. The metric of analysis is offensive or defensive grade.
Distribution
There are two ways to look at distribution. First up is a box-and-whisker plot of grades by position trended by year:
Keeping in mind that the most successful teams were, in descending order, 2016 (10 wins), 2017 (9 wins), and 2019 (8 wins), we can see that a tight distribution is associated with more wins. Median position group grade is also, not surprisingly, important. The highest median grades were 2016 (72.6) and 2019 (70.4). The 2019 team had a higher median and a tighter distribution than the 2017 team, but won one less game. What gives?
The 2019 team had two outliers - one above the top whisker (cornerbacks) and one below the bottom whisker (offensive line). Ryan Willis was the starting QB at the beginning of the season. He was quietly nursing an injury and the offensive line was struggling in the early part of the season, which cost the Hokies dearly in losses to Boston College and Duke. The switch to the more mobile Hendon Hooker (who was better in the read option game) masked some of the OL deficiencies, and that group improved as the season went on, primarily when Vice started playing freshmen like Luke Tenuta, Brian Hudson, and Doug Nester more.
Virginia Tech was 8-3 when it travelled to Charlottesville to play UVA with the Coastal Division title on the line. Caleb Farley was hurt and didn’t play in that game, or the bowl. His replacement, redshirt freshman Armani Chatman, gave up a number of big plays, and the Hokies ultimately lost the game as well as their other starting CB, Jermaine Waller, to injury. Playing with two backup corners against Kentucky and its WR-turned-QB in the bowl game, Tech gave up a late TD, and some critical pass completions, en-route to a disappointing loss.
When I switch the box-and-whisker to focus on year by position group, we see what I would think are some surprising results:
Quarterbacks have the highest median grade (73.6), followed by wide receivers (71.6). Both distributions are tight, with only one outlier between them (a negative outlier for WRs - 2022). The RB and CB positions have been completely hit or miss since 2015, and the TE room has been a mess.
First, on the running backs, Adam Lechtenberg was the Hokies RB coach for only two seasons (2020-21), but those were the two highest grading years, by far. I wish Brent Pry had retained him. He took two transfers with potential, Khalil Herbert and Raheem Blackshear, and put them both in the NFL.
I find the cornerback numbers shocking. Tech has become a shell of its former DBU self. One excellent year, two good years, and five bad years. No consistency. And that is with, what we all considered to be, good coaches (Torrian Gray, Brian Mitchell, and Ryan Smith). Not surprising is that the three best years were, in descending order: 2019, 2017, and 2016. Tech has to be very good at this position in order to win games. Derek Jones might be the most important assistant on the staff over the next few years.
Finally, the tight end room. What can I say? Even with Dalton Keene and James Mitchell, these numbers are quite poor. I should note that Keene never achieved a grade of 65 in any of his three seasons in Blacksburg (his range was 60.2 to 63.8). The negative outlier is 2022, and that grade is just atrocious. So long as the starters at any position grade out that poorly, Tech will continue to struggle.
Trends and Associations
When I trended performance over time by position group, I found some very close associations. In two cases, I would posit the associations are really dependencies - one position group impacting another. In a third, the common link was the passing game. And the final association, well, I see no connection at all.
Dependencies
Offensively, the data show how dependent the running backs are on the offensive line:
Running backs have been outperforming their offensive lines for decades in Blacksburg. Still, there are limits. The Zohn Burden years were pretty stagnant, but the RB grades increased along with the OL grades in 2020, and the difference between the two position groups increased in 2020-21, suggesting very well coached running backs. The fan base had a generally favorable opinion of Vance Vice, but his OL unit got worse every year from 2016-19.
Defensively, the numbers support a dependency between the linebackers and the defensive line. If the line is getting mauled, the linebackers will be less free to make clean stops.
As the graph demonstrates, the two units were tied at the hip during Bud Foster’s final years as defensive coordinator and inside linebackers coach. The shift to Justin Hamilton at DC in 2020, along with changes in linebackers coaches in both 2020 and 2021, appear to have decoupled the two units somewhat. However, the arrival of Brent Pry and Chris Marve has once again tied the performance of the two units closely to one another.
The Passing Game
Looking at WR and TE grades over time really hammers home the importance of the passing game coordinator, a role currently filled by QB coach Brad Glenn.
Beyond the association between the two position groups, I was surprised to see that the receivers graded out just as highly under Jafar Williams as they did under Holmon Wiggins (who moved on to Alabama, where he coached the 2020 Heisman Trophy winner, DeVonta Smith). I also find the negative trend for the TEs somewhat perplexing. I thought James Shibest was a pretty good coach, and I feel the same way about Tyler Bowen. Still, that graph is concerning. The tight ends are just not consistently performing well.
A Head Scratcher
Here is a chicken and egg conundrum - quarterbacks and safeties. I cannot imagine these two positions having any impact on one another, but take a look at this graph:
That’s a pretty tight association up until 2021, and the 2016-17 seasons are almost pinpoint exact. What should we take from this? Nothing. Well, almost nothing. I do think the post 2017 trend lines are meaningful. Play at the safety position has trended down according to both PFF and the eye test. Quarterback is another story entirely.
Brad Cornelsen got a bad wrap for his “failure” to develop quarterbacks. I’m not sure that is entirely well deserved. Despite the revolving door of VT quarterbacks, the trend was positive from 2017 through 2021. And on the topic of Hendon Hooker, let’s quickly review his grades by season:
The anomaly was the 2020 season - the Covid year. There was no spring practice that year, and Hooker missed most of the summer practices due to an injury. Any practices he could participate in were likely limited, due to so many players being out with Covid. Training was also curtailed, and Hooker was noticeably thinner that season. So, is it any wonder that he was the same player in 2020 that he was in 2019, from a PFF offensive grade standpoint. Had 2020 been a normal season (i.e., there was no Covid), I have no doubt his grade would have been around 79.5 for the year, if not better, given all the talent around him. That would have meant that he consistently improved his PFF offensive grade by about five points every year. Kudos to Tennessee for helping Hendon reach his full potential, but he was on track to do that at Virginia Tech, if not for Covid.