24/7 Annual Post-Draft Analysis of Accuracy of Star Ratings
This article has some really interesting information:
- Star ratings really matter - 4/5 star recruits were 30x more likely to get drafted in the first round compared to all other recruits.
- Reading between the lines, it seems like 24/7 was low on some Michigan recruits that panned out (Morris, Turner)
- Recruiting services are getting more accurate over time, which the author attributes to wider availability of electronically timed drills for recruits and better quality HS game video.
it seems like 24/7 was low on some Michigan recruits that panned out (Morris, Turner)
We like to call that "player development"
Also, "scouting."
Or, in Spartan terms "Tunnel Vision."
Ugh, I had to log in to up vote this comment. It only took me 2 seconds.
:P
Important to remember that NFL draft position isn't the final word on how great a college player someone was - which I've always taken to be what the recruiting rankings are predicting. For example I'd argue Chase Winovich, who went 77th in the draft, was a more productive college player than Rashan Gary, who went 12th in the same draft. So even though Gary's higher star rating did end up corresponding to a higher draft pick, the recruiting services didn't get that one right in terms of measuring who was the recruit a team should have wanted more.
Started actually reading the article and it says that 247's goal is to "is to judge talent, and not college stats", which is news to me. Still going to leave the above post up as collectively this seems like a decent illustration of why recruiting rankings can be a general guide to the quality of player your team is getting but is really not a determiner of how helpful a player will be in winning games - since that's apparently not what 247 is even trying to measure.
Indeed you have stumbled upon one of the inefficiencies in the recruiting industrial complex. I believe all the services say this the way they do it. It explains how guys like Blake Corum and Denard weren't 5 stars out of high school. Not NFL sized. Not an obvious NFL position - you won't be a 5 star no matter your physical skills. Another good example is O lineman. When Michigan is pulling 250lb 6'6' dues playing TE in High School you shouldn't worry that they 3 or 4 stars. A lot of the ranking is just because they currently are undersized.
If you read their scouting reports, it's all about comps to NFL players and draft position. They've been pretty clear that 5* = likely 1st round pick and so on from there lately. Not sure if this is something they've talked about long, but they've been pretty clear about it lately
Its a guessing game because players are still growing, and some are late bloomers...so they'll tend to give stars based on athletic traits vs skill, and to determine skill there are so many variables based on scheme, coaching, playing proper position, etc.
Along these lines, I was listening to a good conversation on the CBS College Basketball podcast. They were talking about whether you'd rather have the #5 rated high school recruit or the #5 rated transfer portal recruit and they both basically said they'd rather have the transfer if you're trying to actually win games. The conversation eventually came around to the fact that high school star ratings in basketball are supposed to predict NBA viability, not college success. It made me think there is a opening out there for some service to start offering an alternative ranking that focuses on how productive you will be in your first two years of college.
That's actually not a bad idea. I'd really be interested if something like this came along.
Matt who does the hoops crootin writeups for this site absolutely rates guys by whether they're likely to be a huge asset to winning games vs just a crootin win.
He was very clear that Houstan was likely to be gone before he ever was a star in college. Believe he may have said the same about Papa Kante and some other recent Michigan targets.
Football players have to stay 3 years, so predicting who's gonna be killing it by year 2 has got to depend on opportunity. For example, Caleb Williams joined an OU that had 5⭐ "Heisman candidate" Spencer Rattler in front of him. You wouldn't have bet a lot he'd be starting year 1, right?
The Garys & Daxs should be 1st rd players by the time they've played 3 years. It's a real win if they're already performing like that before they're even draft eligible (i.e. like Caleb, Brock Bowers, and hopefully Will Johnson & Colston Loveland!)
Winovich certainly hasn’t had a better pro career than has Gary. Who had the better college career is open for debate, although I’d say Gary did, whose stats might not have been as gaudy but that’s in part because of how he was used in the defense. A fully unleashed Gary would have been fun to watch.
FYI: Gary was having a huge season last year until he got hurt.
No, you should want (and recruit accordingly) the better player.
If, for various reasons, a Wino outperforms a Gary while in college (or for another possible example Sainristil vs Dax), well that's gravy!
It's a testament to talent assessment & coaching, which for the last few years (2020 excepted) this staff has demonstrated at the highest level!
But if the crootin services are essentially trying to predict which HS players will ultimately be invited to the Combine, and which will become high draft picks, well you want as many players like that on your team as you can get.
I mean - you're implying they should be rating players on whether if the 5⭐ lined up alongside them draws triple teams or pull a hammy, these are the guys who'll really step up.
I say that's the coaching staff's job.
I read through the article and I still don't understand how they came up with their grades. They use some type of point system that wasn't explained very well and there is no breakdown of the scoring. I wish somebody would just take the time and calculate what the expected draft position is based on recruit rating (e.g. a .9500 rating is expected to be drafted with the 61st pick in the draft). Everything I see is looking at star rating with no distinguishing amongst ALL 4 stars, and Draft Round with no distinguishing before the 1st and last picks in a round. Perhaps this is a good off season project for Seth???
It is kind of interesting tho a bit apples to oranges, but at the end of the day nobody really pays much attention to how effectively college recruiting sites mock the NFL draft 3-6 years ahead of time.
If you want to win championships in college you have to hit the recruiting lottery more often than not. Top 5 classes should be your average not your high point. If making the playoff and getting whooped is your plan there are plenty of opportunities.
There are a couple of exceptions:
Clemson - They had a historic run of QBs playing above their rankings paired with a 20-6 run, at one point, in one score games.
TCU (last year) - Because of course we have to be on the losing end of the exception to the rule.
This current classes talent level, if sustained for years to come, will put us up there with OSU, Georgia, and Bama.
Only 2 teams have won playoff games without dominating the recruiting world (TCU and Clemson).
I like this current class and I trust the evaluations of the current staff, but if you look at the average ranking per recruit it's really not at the historical Alabama / Georgia / OSU level.
Yeah you are right. Wedging ourselves between the football factories and the rest of the unwashed commoners is probably our best hope.
If you look at Clemson' average player ranking around 5-10 years ago, it was pretty high. The catch is they had smaller sized classes compared to Bama, OSU, etc. When the recruiting services rank the classes, the quantity of players in a class plays a fairly big role in where the class is ranked.
Yup. I always look at average player ranking more than overall class ranking.
1. Star rankings do matter in the aggregate, not necessarily individually. Meaning a service can get a singular player wrong but it's highly unlikely to get a whole bunch wrong so a team like the currrently ranked #1 recruiting squad in the country Michigan Wolverines is FAR more likely to be good going forward than say, oh I dunno, the 40th ranked MSU Spartans.
2. That being said while star ranking matter as an indicator for potential NFL drafting, coaching matters even more when looking at college team performance. That's why a team with a middling recruiting class in 2020 and 2021 like Michigan can throttle a team chocked full of elite five star recruits like Ohio State two years running.
I like this analysis. Particularly point #1.
Agreed with everything mGrow. I will add rivalry motivation can also be a great equalizer as well.
-MSU beating us through the years (This actually supports your points as well)
-U of M whooping OSU but OSU almost beating Georgia then Georgia destroying TCU.
I hope they are stuck with MeLLLLLLL for the entirety of that 10yr $95 million contract.
You mean MeLLLLLLLLL. That's nine Ls and a "Me" (short for "Meh").
Recruiting service presents data to say recruiting services are better than ever, an annual event.
if only there were a way to $ub$cribe to these services.....
“4/5 star recruits were 30x more likely to get drafted in the first round compared to all other recruits.”
Yet Blue@LSU’s diary stats say that 40% of 5-stars don’t even get drafted.
That's what the data from the 2006-2017 classes showed. It does look like they are getting better over time, though ESPN jumps around quite a bit. Here's a graph:
For some reason, both of them took a drop with the 2015 class. I'm not sure what happened there.
I’m curious what percentage of the average NFL roster is comprised of 5-, 4-, and 3-star HS recruits.
I'm afraid that's above my pay grade.
But don't get me wrong. I'm not a stargazer by any means. I just like to look at data and see how well it jives with our expectations, without any value judgements. When it comes to recruiting, I'm always going to trust the coaches over the services.
Those are not inconsistent data points. Very few college players are drafted, especially guys who were less than three stars. The class of guys who are below 4* players is pretty big set; if only 2% of those guys get drafted then a 60% rate for 5 stars would be 30x more likely.
It’s also possible that the draft rate for 4* is higher than 5* because the latter are a smaller group more prone to being impacted by players that bust. Then when you aggregate 4 & 5* players you get a draft rate of more than 60%.
so 30x must mean there was only 1 non-4/5 star drafted out of 31 first round picks. Yes, that Michigan math/engineering degree is paying off where it counts most.
I sincerely hope it is one of you, my mGoBrethren, who consistently calls out Ari Wasserman on The Athletic (and twitter?) about how he keeps shilling for the "Stars Matter Foundation". dead dot emoji, every time.
I wish it were me. Hats off.
That's a good one. I troll him but not as consistently. The 'how it started...how it's going' of his podcast before and after the 2021 OSU game is still one of my all time Schadenfreude moments.
I think for college football fans it's better to compare star rankings to the All-American teams of even teams in the playoffs. There are a lot of lower-rated QBs I'd take over Will Levis and Anthony Richardson on Saturdays.
Speaking of low... Oluwatimi was a 2 star and ranked as the #3300 recruit nationally per 247.
Star ratings certainly matter, though they obviously aren't a guarantee. They could be a bit of a self fulfilling prophesy, one can imagine. The tiny handful of higher ceiling prospects also have the best chance of reaching their potential. They usually go to the bigger programs, with the better facilities and program momentum, elite strength and conditioning and technique coaches, playing in the bigger games alongside elite talent, etc, etc.
Non-rhetorical question: why does it bother people that recruiting rankings are, when aggregated in a large enough sample, basically accurate?
It’s not particularly surprising (or offensive) that the fastest, most explosive athlete at age 17 is very likely to be the fastest and most explosive at age 21. It would be weird if the opposite were true.
Even if recruiting rankings are really good these days, doesn’t mean that coaching isn’t important. Five stars still require coaching, S&C training and general development - just ask Zach Harrison about that.
I was so bummed by us not getting Harrison. I was also incredibly underwhelmed by his play on the field.
Isn't one of the knocks on recruiting rankings that they're looking for NFL, not necessarily College, traits?
Id like to just mention that the metric of success is often framed by the draft; that is, a player rates highly is a success if they are drafted in these analytical pieces.
However, it is important to remember - and this can often be more readily apparent in basketball - that there are some excellent *college* football players who do not translate to the NFL. While a 5 star who does not get drafted is seen as a failure through the lens of some discourse, some factor could preclude their effectiveness in the NFL, but also see them have productive college careers.
Again, this is often more apparent in basketball. Players like Xavier Simpson, MAAR, possibly Dug - in fact, I see fans hoping we attain *non* NBA caliber players so they stick around, with the idea that they can still be excellent college players. A more holistic assessment would take this into account when judging stars, but then again, stars are often seen as a measure of a players pro potential.
Absolutely, in the portal era ncaa hoops i don't want anymore 1&dones. All but the Bancheros are not contributing as much as 3rd & 4th year players who may never be NBA starters, or even play in the NBA at all.
Very very few kids leave college ready to step on a major D1 football field and shine. Look at Will Johnson everyone knew he had all the talent and ability in the world but it took months (which is small compared to some) to finally be able to know the ins and outs of college CB. Development of the top talent signed is the key. Also I think a big big overlooked area is recruiting to your scheme. I think development and recruiting to scheme is what has really changed for Harbaugh at Michigan. They know who they want and a lot of times that is in agreement with top talent from these recruiting services but sometimes it is not. I trust this staff more then any other staff I have watched in my life time to find the kids they want and then develop them to contribute to winning championships. In between there is a lot of ground where things can go sideways. For the most part I think they are killing it in these areas.
One things that stars don't measure is heart. That's what turns walk-ons and 2-star HS seniors into NFL draft pick.
To be fair, what turns walk-ons and 2-stars into NFL draft picks is generally growth spurts and filling out their frame with muscle.
And heart. And want-to. And love of the game. And it doesn't hurt to be a coach's kid!