One Offs vs. Trends
How much of what Virginia Tech showed against Old Dominion was real, and how much was an illusion?
Virginia Tech plays host to Purdue in a game that is, by the numbers, absolutely fascinating.
I want to hit on as many angles as possible, so let’s dive in.
Devil’s Advocate - Part 1
Nobody seems to like the two quarterback system. But let’s focus on the word system, which suggests a premeditated plan.
The issue is really about going into each game with a plan to play both Grant Wells and Kyron Drones. It ruins the flow, halts momentum gained by one or both quarterbacks, not to mention the team, and it impacts everyone’s confidence.
The criticism is all legit. The question is whether or not it’s just the price of admission.
The Virginia Tech starting quarterback got injured and missed time in 5 of the last 10 seasons. By those odds, Tech loses a starting QB once every two seasons.
Grant Wells did not miss any time due to injury in 2022, so, by the numbers, the Hokies are due for an injury at QB.
For the coach who believes there is a good chance he could lose his starter to injury at some point, what is the best move?
Roll the dice and play the starter until he can’t go, trusting the #2 guy only when you must
Play QB2 in meaningful snaps all season so he is ready should QB1 get injured (with the added benefit of limiting the injury chances for the starter by shrinking the denominator - QB snaps)
At this point, Drones is a talented guy with all the physical tools, but little game experience and a still developing feel for the offense. The two demerits can only be erased by playing him meaningful snaps.
Getting those snaps each week, a little at a time, seems like the best course of action for preparing Drones to play a full game. And the odds are, the Hokies are going to need him to play a full game at some point this season.
Devil’s Advocate - Part 2
The book on Purdue, with exactly one week’s worth of 2023 data, is this:
The Boilermakers cannot stop the pass
They are pretty solid against the run
Points are out there for the taking
A lot of analysts in Tech media circles have extracted the above to conclude that Wells should play the whole game against Purdue because he is the better passer.
But let’s turn this around. The Hokies have a guy in Drones who needs snaps. Simple handoffs or option plays have some value, but real development requires Tyler Bowen to actually use the entire playbook with Drones, not just a set package.
If Purdue’s pass defense is as bad as it looked last week, would it not therefore behoove the Hokies to give Drones a chance to air it out some this week?
The Game Plan
It is true that when the Hokies have the ball the best option is the pass. Tech’s success rate on passing plays is 52%, and the Purdue defense allows a 58% success rate. In contrast, those same numbers on rushing plays are just 28% and 34%.
The question is, will the Hokies switch from running the ball on 60% of plays against ODU to passing the ball on 60% of plays against Purdue?
The answer: not if they don’t have to.
Bowen needs to get the run game going. The Hokies’ rush play EPA currently stands at -0.18, a level of hideousness that exceeds even the depths of last season. Purdue is good, but not great, at stopping the run. And balanced teams are, in general, more difficult to defend.
So, an early stat to key on is Tech’s run/pass ratio. If it’s around 50% at halftime, the Hokies offensive coaches probably feel like they are in pretty good shape. If it is slanted decidedly in favor of the pass, the Hokies could be in trouble.
Regression to the Mean
At this point, we can work through the unknowns around these 2023 Virginia Tech Hokies with a series of binary questions centered on regression to the mean, or, more specifically, determining what the mean is this season:
Was Tech’s inability to run the football last week a one-off or the beginning of a trend?
Was the ability to generate turnovers and win the turnover margin a one-off or the beginning of a trend?
Was the strong special teams play we saw in week 1 a one-off or the beginning of a trend?
Were the poor gap fills by Tech’s linebackers a one-off or the beginning of a trend?
Was the willingness to throw the ball to the running backs coming out of the backfield a one-off or the beginning of a trend?
We should have answers to all of the above next week, but in-game, here are your keys, in order:
If running back touches are largely dependent on catching passes, that will be a sign that Bowen lacks confidence in the team’s ability to run the ball.
If Tech can generate pressure, via an early lead and a strong pass rush from the defensive line, we could be looking at a team that knows how to squeeze the opponent and force him to take low percentage chances, which often result in turnovers.
Tech played a lot of starters on special teams against ODU. If the Hokies do that again against Purdue, the team’s improved play (in comparison to last season) in this phase of the game is likely the beginning of a trend.
Will Johnson is getting serious reps at linebacker by the second quarter would be a red flag concerning the ability of the starters to fit the run.
Set the first half over/under on passes to the tailbacks at 2.5. Anything north of that should suggest the beginning of a trend.
Expectations
I took a peak at Bill Connelly’s score predictions early in the week. His SP+ driven model forecasted a 27-25 Virginia Tech win. This was back when Tech was a 2.5 point favorite.
The line is -3.0 now, and my model expects a 26-23 Hokie win. Notably, my model does not include SP+ metrics as variables.
Consensus across metrics is meaningful in score predictions late in the season. After one game? Not so much.
Still, this consensus is important for perception. If the college football world agrees with Vegas in expecting Pry’s Hokies to win by two to three points, and the Hokies win (or lose) by two touchdowns, we can expect a major reassessment of the team come Sunday.
On the whole, this Purdue team looks like a good matchup for the Hokies. Designed QB runs are not a major part of the Boilermaker offense, they will likely depend on the pass game to move the ball, and their defense is leaky.
We saw last week that the 2023 Hokies could play an imperfect game and win comfortably against a G5 opponent. Expectations for this week are another imperfect game and another win, albeit a much closer one.
That said, I would be less surprised by a double digit Tech win than the inverse. And it would be a good time to notch another convincing win.
Decision day approaches.
