After this week's class, I put together a box plot of VT recruiting classes over the years:
What stood out to me was that Pry is raising the floor:
- The Minimum rated player in this year's class is a 0.8263 - that's the second-highest minimum a VT class has had in the modern recruiting era (the highest was in 2005, where a 0.8400 rated player was the minimum)
- The Lower Quartile (aka, the 25th percentile player rating) for this class is a 0.8547 - This is (coincidently, also) the second-highest lower quartile a VT class has had in the modern recruiting era (the highest was in 2018 - lower quartile was 0.8559).
- Finally, we are seeing a decrease in unranked players. I'm not really sure how to interpret these... more on that below.
Year Number of unranked players 2019 1 2020 3 2021 3 2022 0 2023 1
Some Caveats, and further discussion about the data:
- Data is pulled from 247 - I've read that 247 data before 2005 is not usable, and that it is significantly better around 2011 (I've read some people attribute this to when HUDL started becoming mainstream).
- I used the most recent available rating for players. This is relevant for transfers; some transfers have a transfer rating (which I used if available), other transfers do not (in which case, I used their high school rating).
- On the unranked players....
- The unranked players are all transfers, walk-ons, or special teams players (kickers, punters, etc).
- From 2005-2019, we only offered 1 unranked player according to 247 - my guess is that 247 just didn't track this until recently? Not sure though.
- The unranked players are excluded from the minimums in the charts, but not from the medians or quartiles. Candidly, I couldn't figure out how to get the QUARTILE function in google sheets to ignore unranked players without removing them from the data all together, and I didn't want to do that (yet). After fiddling with it for longer than I'd like to admit, I said, 'Ahh, fuck it,' and just left them in. I'm not sure how this will impact the data, but 3 unranked recruits is >10% of a 20-25 person class - it would probably make those last Fuente classes look better than they actually were.
- Of the 8 unranked players, Changa Hodge was probably the most successful/had the most name recognition (at least to me)
- Khalil Herbert's Transfer rating was a .9000. I don't remember that being a thing at the time.
Finally, this is one year of data. It's obviously too soon to know if this is a trend, but for now, I'm going to slide on my maroon and orange glass, be overly optimistic, and assume that there are more good things coming.
Forums:
DISCLAIMER: Forum topics may not have been written or edited by The Key Play staff.
Comments
The reason for the uptick in unranked players in 2020 is Covid. As far as calculations go, the only way to remedy this is to straight up kick them out of any calculation (On the class calculator, they show up as 0 points added to the team score instead of some small fractional amount). It's essentially "missing" data, so I think you did it right with your maths.
There weren't as many camps and HS games for all of the services to rank everybody, and that should be going away now. I expect we should at least have a composite grade for Mose Phillips III at some point (ON3 and Rivals have already graded him, and if 247 gives him a similar grade that should kick him up to around an 84 on the composite).
Not sure how they handle transfer grades, but I don't think they do that for FCS transfers. I'm guessing Da'Quan Felton will remain unranked.
Super interesting! Thanks for clearing up. Always appreciate your input.
I always appreciate the interactions between you two when it comes to visualizing and contextualizing numbers
Seems to me this is a good sign.
Next up: raise the ceiling, too, while keeping the floor high.
High floor = depth, I hope.
So... what you're saying is... the ceiling is the roof?
Yes!
Great job on this post.
I love what the staff did to close out this class. I like the trend!
God damn our ceiling has been low though the past 4 years
Yeap... It's been a trend I noticed over 2 years ago. I think the upper quartile is the number that best demonstrates the value of a recruiting class. Fwiw, the 2023 class does have the highest upper quartile of any class since 2019. But given how much we're lacking depth, I think raising the floor is a good thing.
I happened to look yesterday at the 2020 class. Holy smokes, only seven players I believe from that class are still on the roster. Virtually the entire top half of the class flamed out, and the bottom ranked guys that stayed ended up being starters or at least solid contributors.
This is a great visualization, I couldn't tell you the last time I used a boxplot. I was starting to feel like "Raise the Floor" was getting a little overused and becoming akin to an office buzzword. This shows that there really is something to the phrase and not entirely our O&M glasses taking over.
In another thread, rpotts12 asked a question about how recruiting ratings are given to players, and I gave an answer that was (at best) half researched/half pulled out of my ass (edit: it wound up being at least 50% incorrect).
BUT, during those 5 minutes of googling, I managed to find 247's definitions for 'high 3-star' vs 'mid 3-star' vs a 'low 3-star':
Now, let's look at the quality of our past few recruiting classes:
Next, take the definitions from 247, and append them to the table above. You see that:
This might be a 'no duh' moment for for a lot of you, but matching this qualitative definition to the numbers was a bit of an 'aha' moment for me that truly contextualized how bad Fuente's last two full classes were.
Do you have similar descriptions for below that? Seeing those averages in '20 and '21 feels like a kick in the nuts and a "if they were BELOW P5 starter level, no wonder we did how we did this year."
that's 80-83. "potential contributor" pretty much means it's anyone's guess
I meant the averages (far right column). that were below .8000. Means there must have been a lot of players below .80 to pull the averages down so much.
ah i getcha -- typically unrated players or specialists like kicker, long snapper, etc
edit: that's why the median is so much higher than the average fwiw
Reread the OP and the first comment - due to COVID, there were a lot of unranked players. 3 or 4 players in the class are unranked (given a zero) and the whole class average is brought down. It's one of the reasons I prefer using median over average.
To illustrate this:
Two 2-stars and one 5-star will average out to be ~.8000 ([.7+.7+1]/3 = .8)
One unranked player, one 2-star, and one 5-star will be a .5667 ([0 + .7 +1]/3 = .5667)
Makes sense, and I wondered that, so maybe give that one a pass. But how much was the '21 class affected by that though?
probably as much or more than the 20 class because the '20 class already had their senior tape and was mostly signed by the time the pandy hit, but '21 class was going into their senior year and lots of places either restricted athletics or canceled entirely
edit: on a more granular level specific to VT: the 20 class suffered due to turnover in our backroom recruiting personnel and the staff electing to put all of the eggs in the '21 TX2VT basket.... the '21 class sucked because we didn't close on any of our most important recruits (dematrius pls) and settled for plan c signees up and down the class
Ahhh, that makes sense for the 20 vs 21 classes, I had the impacts reversed in my head.
Bob Schtick, CJ Mcray, and Will Ross were all unranked (source)
Somewhat unrelated question - but feel like you might know this. We've talked about recruiting misses (players rated at a certain level not meeting the descriptions for their respective rating). Do you know if any of the recruiting ranking outlets has done an evaluation of rating vs college performance to evaluate the percentage of "misses" and even potentially to see if the ratings have gotten better over time (so fewer misses over time, or about the same (especially given the wider availability of video nowadays)).
Feel like that would be an interesting analysis.
Another might be for all the players who have entered the portal their transfer ratings compared to their recruiting rating as a similar evaluation of recruiting rating vs college performance.
Not really... There's a few things I've seen:
It's just a tough thing to do. How do you determine if someone was a contributor? Number of snaps played? What if they're playing more snaps than expected because the person ahead of them is injured? What about players who were great in college but are tweeners and don't fit in the NFL? Just think about Tebow
and Cam Newtoneasily2one of the best college QBs of the last 20 years. Nothing in the NFL.EDIT: Speak of the devil... look what the athletic just dropped: https://theathletic.com/3128181/2022/05/23/recruiting-revisited-ranking-...
Also, here's a piece that's VT focused on the 2018 class, with links back to earlier classes.
Here's the scale they use over at the Athletic (created/popularized by Max Olson):
Bro cam newton won an mvp and started in the super bowl after going 15-1 that year lol
LMAO completely forgot about that... All I could think of is how the patriots got him for cheap and it went nowhere.
Good points. I guess some things definitely lend themselves better to metrics than others.
I like doing evaluations to determine if there's been improvement over time. I assume the sheer quantity of players also makes the evaluation difficult. Add to that if a player transfers up or down between P5, G5, or even FCS, the metrics wouldn't be equivalent anyway (the analysis part of my brain appears to have turned on 😂).
There's also not much of a market for it. 247/ON3/Rivals own the space in between high school and college. Every major media outlet has plenty of draft coverage, which is the space between college and NFL. But who cares about ranking the players who graduate but can't play in the NFL?
So you're saying we could corner the market on ranking the Eurasia players?
Only if you're telling me that there are in fact Eurasian players
Oh it exists, at least according to this obscure website I found.
https://eirball.ie/eurasia-american-football/
I really should have stated we could corner the market on ranking players that decide to go to the CFL, European Football League, or any other overseas league, but I didn't want to have to type all of that.
Well, I look at it as the rating outlets wanting to figure out if their ratings are good or could be improved. I mean, individual schools may be interested in terms of their own scouting and matching the rating outlets, and also on whether they can improve their scouting rating. Fans/boosters/administration would be interested from the perspective of whether their coaches are scouting/recruiting well (also basing that decision on how many players have been misses). I mean, it feels like we had more than our fair share of misses over the Fuente tenure (based on performance here and at other schools players have transferred to after Tech, HH notwithstanding).
247 is pretty damn accurate. We hear about the busts or the misses because they're notable, but the reality is that for the past 15 years, it's been impossible to build a team on 'diamonds in the rough'.
Awesome - thanks for the update!!!
That year for VA top ranked recruits was really a bust wasn't it for everyone.
CJ MCray was a transfer from marshall so dont think he counted in the signing class rankings that season?
I think even as a transfer, if he was unrated going into college, the point still applies, right?
In my numbers, I pulled ratings for every recruit, transfer, and JUCO player.
Class of 2021 had 7 incoming transfers. 2 were unranked, 1 had a high school grade but no transfer ranking. The other 4 had transfer rankings.
Awwesome - thanks for doing the math with the examples for me - helps a lot!
As in below a low 3-star? That would be a 2-star:
LINK
Yeah, so I'm assuming you don't want a lot of those on your team (don't punters and kickers get rated low though, even if they're good?). Basically, having our AVERAGE be below .80 seems shit-tastic.
Adding on to this... of our 6 transfers into the program (so far):
Pry is starting to raise our ceiling
Recall: Fu' & Co. sold the very same narrative.
Frank's bottom 25 or so were NOT A.c.c. quality.
So, they started out by raising the Jamerson floor and building bottoms up.
Maybe?
May even be true now for Pryimtime.
I cannot say that that is dead 'rong, right?
However, I can say... that IF/when your floor guys are in the game?
LOL... someone was prolly getting 'floored' itself.
As it is really tough to find a Tony Paige or someone akin in the social-needia video selfie era.
💯
b.street
Can someone help?
Ill try...
Recall: Fu' & Co. sold the very same narrative.
Frank's bottom 25 or so were NOT A.c.c. quality. Fuente said Beamers classes had a low bottom, not enough acc caliber players
So, they started out by raising the Jamerson floor and building bottoms up.
So Fuente tried to raise the recruiting floor and build from the bottom up
Maybe? Maybe ehhh
May even be true now for Pryimtime.
Deion Sanders is trying it too at ColoradoPry=PryimtimeI cannot say that that is dead 'rong, right? Its not a bad strategy
However, I can say... that IF/when your floor guys are in the game? its not a bad strategy until your bottom guys become starters across the board
LOL... someone was prolly getting 'floored' itself. Fuente wasnt doing his job well
As it is really tough to find a Tony Paige or someone akin in the social-needia video selfie era. Hard to find diamond in the rough players like Tony Paige in this generation
I took the Pryimtime to be Pry and he's having to build from the floor up also. Which makes sense since we're not getting 4 and 5 *s.
You're 100% right I thought it was a typo lol thats how far Pry and Prime are in my mind
I took it this way also, and that Pry's complaint that the floor guys were not ACC caliber may be more true than Fuente's.
Also that Pry might be actually doing what he said.
Although in my opinion, it wasn't that Fuente recruited players that were not ACC caliber, at least not at first. It's that as time went on, the floor for our recruits kept dropping until it was below ACC caliber. And in some cases, he tried to compensate for lack of quality recruits overall by over-recruiting some positions because he could get higher rated players (the reasoning is speculation on my part, but the over-recruiting clearly happened as evidenced by RB recruiting, for example).
The amount of talent that he ran off is astounding. The 17 18 and 19 classes had some decent top end guys, but the floor kept getting lower, people were ran out of the program, and the 2020 and 2021 classes were just so bad overall.
I am impressed with the translation.