Parallels Abound as the Hokies Prepare to Play Duke, This Time in Football
Parallels Abound as the Hokies Prepare to Play Duke, This Time in Football
The 2022 Blue Devils are exactly what Tech fans hoped the Hokies would be. And the outlook for Saturday's game looks similar to a certain March basketball game. Might we see a similar result?
During the first media timeout at the ACC tournament championship game last spring, Mike Young told his team that if they attempted to out-shoot Duke in a high scoring game, they would lose. Despite their hot start, they allowed a similarly hot start to Coach K’s team. The team needed to play better defense and execute on what Brent Pry would call “complimentary ball”.
Seven months later, the Hokies football team, losers of six straight, takes on a high scoring Duke team that looks nothing like the beaten down group that Tech curb-stomped in Justin Fuente’s final game as head coach. Duke is solid everywhere, and great at many positions. Their 6-3 record, but for a couple three-point losses, could easily be 8-1.
Virginia Tech is 2-7, but they are not a bad team. A bad team is one with bad players who play badly. More accurately, we could say that Virginia Tech is a team with decent players who are consistently inconsistent. In other words, their play does not match their ability. Rarely does it all fall apart across the board, but conversely, the team has yet to play solid football in all aspects for 60 minutes. To beat Duke, they will need to do just that, and perhaps get a bit lucky.
Jealousy
When I gaze at the preview viz, the guys in blue look a whole lot like what I was expecting to see from Virginia Tech this year. And it wasn’t just me:
I didn’t think they would win the ODU game by four touchdowns, and obviously they didn’t, but I also didn’t expect Duke to be any good this season. And yet, here we are, Duke enters the game with three times as many wins, and they are 9.5 point favorites. Tech beat these guys 46-17 last year! Two first year coaches walk in, but only one rebuild walks out.
The Duke Offense
What I find most impressive about Duke is their rushing attack. It makes the whole offense go:
Every offensive x-factor favors Duke, and outside of field position, it’s all because they can run the football and Virginia Tech cannot. The offensive line alone is responsible for nearly a yard more per run than Tech’s. The Blue Devils run the ball (53.3% run rate) to stay ahead of the chains, and they are rarely forced to throw (28.5% pass down rate). It’s amazing how a passing game can open up when the run game is clicking. Finally, like most opponents, Duke scores nearly a point and a half more than VT every time they get inside the opposing 40-yard line (points per opportunity).
A comparison of advanced offensive stats sheds further light on how Duke’s offense has been successful:
It is not so much that Duke’s offense is more explosive than Virginia Tech’s - it’s not - but it does a much better job of consistently moving forward. Most telling is the offensive stuff rate. Virginia Tech’s running backs are held to no gain or a loss on 19.2% of runs, compared to only 11.6% for Duke. Staying ahead of the chains is the primary driver of Duke’s seven percentage point edge in success rate, which is defined as follows:
Gain 50% of the yards required to score or get a fresh set of downs on first down
Gain 70% of the yards required on second down
Gain 100% of the yards required on third or fourth down
A bad play for Virginia Tech’s offense is a 1-2 yard loss, while a bad play for Duke’s is a 1-2 yard gain. Neither is a successful play, but a five-yard run by Duke on 2nd and 8 would be, while an eight-yard pass completion by VT on 2nd and 12 would not be. All of this culminates in Duke having a PPA (predicted points added, which is an EPA, or expected points added, model created by collegefootballdata.com) three times higher than Virginia Tech. In short, Duke’s motion results in a lot more action than Virginia Tech’s.
And, as if that weren’t bad enough, Duke also generates a lot more motion:
Duke’s median percentage of yards gained (55.7%) is more than double that of Virginia Tech (22.3%). Translation - Tucker Holloway will have to be Georgia Tech-level awesome just to keep the field position battle even. And notice the curious fact that Virginia Tech has not had one drive this year that gains between 77.7% and 98.4% of available yards. That stat hurts doubly coming off a pair of one-point defeats.
The Duke Defense
Both defenses grade out higher than one would think given the results on the field. Two out of Duke’s three starting defensive units grade better than Virginia Tech’s, but the Hokie defense comes in with the overall edge (81.4 to 79.4). In ACC play, that has resulted in Duke allowing 26 points per game vs. Virginia Tech’s 27.7 ppg allowed. For the season, the Blue Devils allow two points per game less than the Hokies (23.1 vs 25.1). Duke and Virginia Tech have played three common opponents, allowing the following:
North Carolina - 38 points (VT allowed 41)
Miami - 21 points (VT allowed 20)
Boston College - 31 points (VT allowed 10)
That is a net one point edge to Duke in games against Miami and UNC, but the Boston College game is curious. Tech only allowed 10 to BC and was playing without Alan Tisdale. Meanwhile, last week, Duke allowed 31 to a BC team playing at home with a backup quarterback. I’m not convinced that that is anything more than an outlier, but you can bet Tyler Bowen and the rest of the offensive coaches studied that tape very closely this week.
One final similarity between the two defenses - Duke’s points allowed per opportunity (4.0) is similar to Virginia Tech’s (4.2), and both are higher than one would expect given the PFF grades.
If the Blue Devils defense breaks slightly less often, it does bend more (allowing a median of 40.2% of available yards to be gained by the opposing offense, compared to 28.1% by VT’s defense). That could be the difference between a couple long William Ross field goals and a couple short Peter Moore punts (or the difference between the most likely and best case scenarios - more to come on that below):
Legendary sports journalist Natalie Merchant foresaw all of this back in 1995, and she captured Hokie Nation’s feelings aptly with just one word: jealousy.
Gut Feelings
I launched the multivariate score prediction model after the North Carolina game, and prior to the Pitt game I programmed it with data current at the time to look ahead to the rest of the games on Virginia Tech’s schedule.
At the time, based on results through the first five games of the season, it predicted a 39-19 Duke win. Fast forward to today and the model is calling the game 36-30 for Duke. What changed?
Duke’s defense for one. The Blue Devils allowed 19 ppg through their first five games, but they have given up, on average, 29.3 points in each of their last four games. Their advanced stat profile on defense just is not what it was a month ago.
Things are more complicated with Virginia Tech. The Hokies have lost all four games since they played UNC, and the model even picked them to win a couple of those. Here are the predictions vs. the actual results:
Pitt 49, VT 17 (Pitt won 45-29)
VT 34, Miami 33 (Miami won 20-14)
NC State 37, VT 24 (NC State won 22-21)
VT 20, Georgia Tech 12 (Georgia Tech won 28-27)
While the model picked the Hokies to go 2-2 in those games, the net result of the score picks was for Tech to be outscored by 36 points (9 per game). The actual result was a net difference of 24 points (Tech got outscored by 6 points per game). So, despite the losses, Tech overperformed a little just as Duke’s defense began to loosen up.
The remaining difference, I think, lies in the ELO rating. Remember, ELO lags a little, so Duke is still being penalized for it’s pre-2022 performances. However, another way of looking at that is perhaps Duke has played much better this season, but the roster still comprises many of the players who struggled so much last year, and some of those underlying weaknesses might remain. Fifteen hundred is generally the ELO dividing line between good and bad, and Duke, although they have climbed more than 400 points since last year, is still at only 1433.
All of this is to say that I think the model is correct, directionally speaking, but with Malachi Thomas and Dorian Strong out for Virginia Tech, I would consider that 36-30 prediction closer to the best case scenario than the most likely outcome.
FPI currently forecasts 7.6 wins for Duke this season, and the Blue Devils will end the year at Pitt then home against Wake Forest, followed by a bowl. That means FPI is counting on a Duke win this week. Combine that with the fact that VT has not scored 30 points this season, but has scored 27 in two different games, and I am expecting something more like a 38-27 Duke win.
Last spring, Mike Young got his team to settle down and play complimentary ball. Brent Pry will try, and may, in keeping with the recent trend, succeed for a quarter (or even two!), but these gridiron Hokies are both less talented and less experienced than those hardwood Hokies. That will be the difference on Saturday.
Record outright: 5-4
Record against the spread: 4-5
Record over/under: 3-6