OT: Super Bowl starters: less 4 and 5 stars than you would think

Submitted by OC Alum91 on February 5th, 2017 at 11:31 AM

Numerous stats studies been posted on this blog correlating being a college FB starter or an NFL draftee with recruiting evaluation stars.

However, for the Falcons and Pats.  These two teams have less 4 and 5 star ranked starters than expected.  (compared to percentages from the NFL draft, prior Super Bowls, and Pro Bowl).  Edit (originally  wrote "more starters were unranked or rated 2 stars (12 total) than 5 stars (three total)")

Julio Jones, Martellus Bennett, and Malcom Brown were the only starters who were rated 5 stars coming out of high school.

Alan Branch from UM was a 3 star.  Tom Brady was the only starter who was recruited before the star system was widespread--author retroactively gave him a 4 star.


Edit: title changed from "more 3 stars" to "less 4 & 5 stars"



February 5th, 2017 at 11:34 AM ^

Why does this need to be re-explained every year. There are so few 5 stars, only ~30 per year vs 1000s of 2-3 stars. The percentage of 5 stars success at any level is dramatically higher than 2-3 stars success rate


February 5th, 2017 at 4:01 PM ^

Right. If you can count them then it's fewer. If it comes is piles, heaps or mounds then it's less.

As in, people who get this wrong tend to have fewer IQ points compared to those who don't. But people who complain about this shit tend to be less fun to hang out with.

You can count IQ points but you can't count fun.

OC Alum91

February 5th, 2017 at 1:10 PM ^

I get that there are only a handful of 5 stars to go around.  I think a lot of people are missing my point (I forgot how sensitive everyone is about "starz hype" and didn't write the post clearly)

My observation is that these two NFL teams have won their conferences and made it to the pinnacle of theNFL, the Super Bowl, with less 4 and 5 star talent than other NFL teams.  I put out some data in response to  lhglrkwg  about the stars composition of Pro Bowl, Super Bowl, and NFL draft picks, and these two teams have less than average.

If you buy the fact that 4 & 5 stars, like NFL draft order, is a generally correlated with "talent" and better performance, then these two teams have defied that correlation, and have reached the pinnacle of the NFL with less "talent."  

Is it coaching, culture, luck?  




February 5th, 2017 at 2:08 PM ^

No, that's not what this means.

You are seeing that for some awesome accomplishment (getting to the Super Bowl, in this case) there is a 12-to-3 ratio of NR/2* to 5*s.

Your mistake is one of interpretation. Those numbers strongly SUPPORT recruiting rankings as correlating with success. I'd like to help you to see that.

The annual ratio of NR/2* players to 5* players is something like 6,000 to 30. Which is a 200:1 ratio.

So unless there are 200x more NR/2* players in the Super Bowl than 5*'s, you have made the wrong conclusion.

And there aren't 200x more. There are only 4x as many NR/2*'s as 5*'s. Which supports the predictive value of recruiting rankings.

And renders your conclusion as bad. Your conclusion is very bad.

OC Alum91

February 5th, 2017 at 7:52 PM ^

I think you may be misunderstanding.

Yes, I get your point about about Recruiting rankings correlating with success, and I agree.   I never said that Recruiting rankings aren't predicitve. That was someone else on another thread maybe last weak, last month or last year.  I 100% agree the Stars are predictive and correlate with future performance.

My obesrvation is that these two particular teams, the Pats and Falcons, have fewer 4& 5 star players than expected compared to other SuperBowl teams.  I am not using the general pool of high school recruits as the reference point.  Maybe the typical Super Bowl team has 15 players rated 4*/5* , these two teams only have 9 players rated4*/5*.  In my comparison, I have already weeded out all the other NR/2* players who never made it to the NFL, and am only comparing to other NFL teams.  

If I said these two teams have more players drafted in later rounds, than other NFL teams, would that be easier for everyone to accept?  If you buy that 5* is correlated with higher draft pick, then it's not that far of a leap.  But everyone on this board is so smart with the statistics, and quick to jump on a perceived misconception.

Yes, I agree that an NFL team will NOT be filled with 4* and 5* like Alabama and OSU.  I'm saying these two teams have even less 4* 5* than your typical NFL team or typical Super Bowl team.  

The average star rating of the Pats and Falcons this year is 2.8 and 2.5 (more 2 stars and unranked than 4 & 5 stars)

2012, Super Bowl teams  (Giants and Pats) average stars rating was 3.5 & 3.2 (on balance more 4 & 5 stars than 2 stars and unranked)

2013,  Super Bowl teams (Ravens & 49rs) average stars rating was 3.4 and 3.5 (on balance more 4 & 5 stars than 2 stars and unranked)

2014, Super Bowl teams (Seahawks and Broncos) average stars rating was 3.0 & 2.5 (one team had more 4& 5 stars, the other had more 2 stars and unranked)

2015 Super Bowl teams (Seahwaks and Pats) average stars rating was 2.8 and 3.0




February 5th, 2017 at 11:43 AM ^

3 out of 22 = 13.6% Not gonna spend time digging too deep into the stats, but that seems about right to me. 

Also, Alan Branch was not a 3 Star.

Va Azul

February 5th, 2017 at 11:44 AM ^

Kahneman has a book, "Thinking Fast and Slow", that everyone who wants to revisit this topic should read (or at least be familiar with other sources of cognitive bias regarding failure to properly weight baseline probabilities in making decisions / inferences)


February 5th, 2017 at 12:18 PM ^

I highly endorse the reading of "Thinking Fast and Slow." This is one of the best books ever written for understanding how human beings think and when and how they think wrongly. Even better, it is completely accessible to people who aren't schooled in human psychology. Wonderful, wonderful book.


February 5th, 2017 at 11:49 AM ^

There are very few 5 stars. The odds of a 5 star being a success are much higher than that if a 3 star. If you want proof, google it. This gets addressed every year

OC Alum91

February 5th, 2017 at 12:31 PM ^

Maybe you're missing my point, or maybe I'm missing yours.

Based on the correlations of 5 stars and 4 stars to draft status, I would guess that there would be a higher number of 4 or 5 stars starters.  I don't have exact numbers, but something like 50% of 1st round draft picks were 4 or 5 stars.  

I get it that there are few 5 stars.  But if you include 4 & 5 stars, you have a large pool of very talented high school recruits.  

Eyeball test looking down the list, on these two rosters there are disproprionately fewer 4-5 stars.  Out of 44 spots, only 9 (20%)  starters were 4 or 5 stars.

2017 ProBowl.   80 players.  35 (43%) were 4 or 5 stars.  Only 19 were 2 star or unranked

2015 Draft.  251 players.  86 (34%) were 4 or 5 stars.  Only 52 were 2star or unranked..

2017 Super Bowl.  44 starters.  8  (29%) were 4 or 5 stars.  Only 12 are 2 star or unranked 

I think I worded the post badly.  My main observation is that it feels like there are less 4 and 5 stars in this Super Bowl than you would expect.

Average rank of Pats is 2.8 stars, Falcons 2.5 stars.

In 2012 the Super Bowl teams were 3.5 stars/3.2 stars.

2013:  3.4/3.5.  

2014:  3.0/2.5

2015: 2.8/3.0

(source: http://j.tinyurl.com/h82dbpb)

Maybe I should have titled it "less 4 or 5 stars than you would expect"  





February 5th, 2017 at 11:51 AM ^

In addition to the obvious point about there being far fewer four and five stars, it also seems like there are plenty of decent to good NFL teams that make their living in the 3rd to 6th rounds of the draft, which seems to be where a lot of the better but still less heralded players in college seem to find themselves. There is somthing to be said for that strategy quite honestly, and something to be said for scouts who can pick these kids out from what is a very large pool otherwise. 


February 5th, 2017 at 12:09 PM ^

Not only that, but there are consistently good teams (Steelers, Patriots, etc.) who are always picking at the bottom of each round because of their on-field success. So not only do the Patriots generally pick well in the later rounds, but they're picking at the bottom of each round. So those 5-stars who realize their talent in college go early, while the Patriots (when they win the Super Bowl) are picking at #32, #64, etc. The Steelers are picking (roughly) at #28, #56, etc. The Packers are picking at #26, #52, and so on.


February 5th, 2017 at 11:52 AM ^

It just goes to show you that your fate isn't sealed based on high school performance, but there is a predictive advantage of having more stars. Sometimes players get better, sometimes they live up to their potential and sometimes they get worse. And five stars have better rankings than unranked players. True story, even though I was an unranked player coming out of high school, I did win a couple IM sports championship tee shirts. Not that I was good or anything, but I got the shirts.


February 5th, 2017 at 12:00 PM ^

The worst teams draft the best college players so a lot of the five stars go off the board in first half of it's round and play for lousy teams. I think that great coaches make great players and teams. Lousy coaches can draft great talent and not get measurably better. (Any Lions fans out there?)


February 5th, 2017 at 12:21 PM ^

Yeah, the topic is a dead horse. Comes up every year and the folks who obsess over high school recruiting get defensive while most folks recognize it for what it is -- everyone understands the rankings/stars and laughs at espn when they run ads for the Draft and show clips of Brady (rather than half the 1st Round busts)

Dee Hart playing today?? He was 5-stars and went to Alabama, so presumably this is a rhetorical question.


February 5th, 2017 at 12:37 PM ^

I can't believe they sorted the list by first name!

Not by number of stars! Not by position!...by first name!...that is as useful/interesting as sorting it randomly!

Going through that list really made me want to punch the author in the face.


February 5th, 2017 at 12:38 PM ^

Just to reiterate what everyone else is saying here about the baseline for five stars versus three stars, let's try a silly thought experiment:
Let's say there are 32 five star prospects every year and every single one is such a success that every year 32 five stars are drafted in the first round. Then, because they are awesome, they each have a 10 year career as an NFL starter. In this *wildly* optimistic scenario, each team would have only 10 five star starters on its roster at any one time. That means that less than half of the starters in the NFL would be five stars.
And these assumptions are outrageously unrealistic: the average NFL career is less than four years (not not less than four years as a starter either, just less than four years), and being a five star is not, in real life, an iron-clad guarantee of being drafted or playing in the NFL, let alone being a starter there. Yet even with these incredible assumptions, you're still going to have more non-five stars starting for NFL teams than five stars, just because there aren't enough five stars around.
Anyway, once you start walking the numbers back to a more reasonable baseline, it's not really shocking to see "only" a handful of five stars in any given Super Bowl. Indeed, we should be more impressed that of the thousands and thousands of high school football players, scouting services were able to identify those small handful of players most likely to be in the Super Bowl. It's a testament to the *effectiveness* of these services that we see so many five stars starting in the NFL.