Get Big or Get Out?
Get Big or Get Out?
Was lack of commitment to strength and conditioning a factor in freshmen entering the transfer portal?
Today’s topic is the transfer portal, and before I launch into a quantitative analysis of the strength and conditioning angle with regard to transfers out, I wanted to share an awesome Substack that I recently discovered.
If you’re big into college football history - the kind of person who used to love listening to Ivan Maisel and Beano Cook on the ESPNU College Football podcast more than a decade ago (Editor’s note: has it really been that long?) - you will definitely want to subscribe to Football Archeology. New articles come out daily, and they are extremely engaging, with lots of great historical images.
On with the show
I was surprised to see Cam Johnson enter the transfer portal. He’s coming off a pretty major injury, but seemed in the hunt to get serious playing time, if not this year, then certainly next. One thing that stuck out about Johnson is his size - 165 lbs. when he signed his letter of intent and 165 lbs. in the spring of 2023, according to Hokiesports.com. That got me thinking about what makes a guy, not even one year into college, decide to transfer out. There are a lot of fit-related factors (homesick, sick family member far away, etc.) that are beyond quantification, but there are variables that we can analyze numerically.
I created a little dataset that, among other factors, included:
The number of offensive/defensive snaps played in 2022
The weights for each player in 2023 and 2022 (according to 247 Sports)
Transfer Portal Status
I noticed that the guys who portaled out had half the year-on-year absolute value difference in weight as the guys from that class who remain on the team (7.3 vs. 14.0 lbs.). Considering weight change as a proxy measure, that begs the question, are guys who don’t fully buy in to the strength and conditioning program being pushed out?
Secondarily, I wondered if the number of snaps played in 2022 was associated with the absolute value difference in weight. In theory, guys who played more snaps would miss Friday lift sessions, which would make it harder to gain weight and get stronger.
Obviously, there are some serious limits to consider. While most high school players need to gain weight to get on the field in college, some don’t. Some need to lose weight, and the best players are usually close to their ideal weight when they first step foot on campus. So, perhaps looking at both snap counts and weight changes together can explain portal entrants. For example, a player who didn’t see the field and whose weight did not change could be judged as neither particularly talented nor a close adherent to the S&C program. Were these guys more likely to transfer?
Regressing our way to answers
First, I will state the obvious - the following regressions are not sufficiently powered to uncover anything outside of a very clear association (such is the challenge of analyzing decisions related to second year coaches). So, for simplicity’s sake, when a regression fails to reach statistical significance, I’m going to assert the blanket conclusion of on a case by case basis. Ideally, I would like to see all of the following regressions fail. There are other ways to measure commitment, many of which cannot be quantified, and I would hope that they are given sufficient weight when these decisions are made.
Is weight change (a proxy for S&C commitment) associated with entrance into the transfer portal?
Result: On a case by case basis (no statistical significance)
The Class of 2022 shows no connection between weight change from the moment of commitment until the spring of 2023 on the decision to enter the transfer portal. Statistically speaking, it wasn’t even close.
Nothing to see here folks. Or is there? Let’s keep digging.
Is the number of snaps played in 2022 associated with the absolute difference in weight?
Result: On a case by case basis (no statistical significance)
A little bit closer here. In the neighborhood, if not the ballpark.
Admittedly, we’re talking about a big neighborhood (low R Squared and high P-value), but let’s pull some more threads. We’ve gone from a P-value 0.437 to 0.182. Another 0.135 and we’ll hit statistical significance.
Is there an association between the players who needed to change their weight and the absolute difference in weight?
Result: On a case by case basis (no statistical significance)
For this one, I created a subjective dummy variable (0 or 1) to flag the players a reasonable person would say needed to gain or lose weight. For example, the aforementioned Cam Johnson needed to change his weight (gain) as did Xavier Chaplin (lose), but all three tight ends are in the acceptable range.
No dice again. With an R Squared approaching 0.1 and a P-value of 0.161, we’ve taken a quarter step forward, but still a ways off from statistical significance. Let’s see if we can get any traction with a multiple regression.
Is there an association between players who needed to change their weight, the absolute difference in weight change, and entrance into the transfer portal?
Result: On a case by case basis (no statistical significance)
In so many words, nope!
Not only did we not get within sniffing range of statistical significance, the R Squared number is barely greater than zero, meaning, the two variables explain about 3% of the difference between the players that portaled out and those who stayed.
Is there an association between the absolute change in weight, snaps played, and the decision to enter the transfer portal?
Result: On a case by case basis (no statistical significance)
Finally, let’s consider weight change and snaps played. If a guy didn’t play and didn’t get his body right, one could hypothesize that either the coaching staff would push him out or the player would be buried on the depth chart, see the writing on the wall, and hit the exits of his own accord.
Once again, the data do not allow us to move beyond the null hypothesis that there is no association between these factors.
Where does this leave us?
One of the problems with scientific research these days is that non-confirmatory studies are rarely published. However, it is just as important to report when you do not find an association as when you do. Perhaps next offseason I will revisit the players in this class and see if their weight changes from their senior year of high school through to August 2023 were associated with their playing time during the season.
For now, I am actually happy with these findings. I would hate to see a staff so wedded to players reaching the ideal weight in year one on campus that they push out guys who struggle to pack on (or lose) weight. Case in point: Dorian Strong. He’s still skinny, and he’s still an excellent player.
On Deck
In case you missed it, Will Stewart had a really enlightening conversation this week with Kevin and Robyn Jones from Triumph NIL. Among many topics, they talked about how NIL can play a role in great players choosing to forego early entry into the draft. Stewart noted that Michael Vick would have preferred to have stayed at Tech another year, having just figured out college (Editor’s note: I was just figuring out college midway through my third year at Tech, too). Jones added that he would have stayed for his senior season had there been NIL in 2004. Can you imagine the 2004 team with KJ in the backfield? That very well could have been a National Championship team.
Anyway, I’m really curious about the math behind such stay/go decisions. If you’re going to be the #1 pick, you have to go. But, at what point does the value of staying match the value of leaving. As the TSL Podcast interview makes clear, personal brand building is a big part of NIL, and one has to figure it can have major professional ramifications (both in regard to where a player is taken in the draft and what kind of non-NFL offers guys get). This could end up being a multi-part investigation. We’ll see. Should be interesting. Stay tuned!