That supplies correlation ... but says nothing about causation. How about all the years that Texas or Tennessee as recruited ridiculously well. What has been their pay-off?
Coach Bell going back and forth with Hawknation on twitter.
- Thread starter Wasnt-drunk-didnt-troll
- Start date
Other determining factors include ... how well a players fits within a particular scheme, how quickly a player can learn a particular scheme, and how easy it is for players to execute within a given scheme. Some of these latter factors I mention have definite overlap with the latter factor you mentioned.
You seem reasonable and I totally agree that using Star Ratings at a micro level (i.e. individual players) will yield several outliers. Iowa's evaluation methods tend to be pretty good at spotting these outliers. Iowa's development process is also above average.
Jon is talking about using Star Ratings at the macro level. You are purposely misrepresenting Jon's correlative point: the higher your team's Average Star Rating, the more likely it is that you have won a Championship. He never said it was the 'cause'. He simply stated the correlation is there. You laid it out yourself. The best players tend to be those who are the most sought after (there are clearly exceptions). The most sought after (by your own admission) tend to have higher Star Ratings. Compare that to Championship teams. Championship teams tend to have better players, which is why they win Championships. Better players tend to have been the most sought after. The most sought after tend to have had higher Star Ratings. Why on earth are you arguing that a positive correlation between a team's Average Star Rating and chances that team has of winning a Championship does not exist? People intuitively know this, even if they don't have facts in front of them. Teams of good players tend to do better, thereby validating the Star Rating as an indicator of potential success.
I'm not arguing that. I'm responding to the headline (I didn't read the article, because I've read that article 100 times). I'm arguing that stars do not, in fact, matter. The thing the rankings attempt to measure matters. Recruiting matters (NO ONE would argue otherwise), but the rankings are imperfect.
That distinction matters because people get so caught up on the imperfect attempt to capture one imperfect measure of one factor influencing a team's success that they dismiss results on the field based on it.
The same goes for analytics stuff. People get so caught up on stuff that attempts to predict results that they value it higher than the result itself.
Last edited:
For those of you talking about how stars don't matter I would strongly encourage you to read the original article. No one is saying stars are the end all be all but actual math and science was used and it was proven they do in fact matter.
They don't. Having better players does. I've read that article and seen that math 100 times.
You know what else would be highly correlative and predictive? Whether a team has won a national championship in the 30 or so years previous to that season. But those trophies don't play. They just reflect the same thing as star rankings: prestige and past success, allowing you to have a better chance of getting the players you want. And, like star rankings, it would be imperfect.
Those things don't matter. They reflect something that matters. That's it.
Those things don't matter. They reflect something that matters. That's it.
Yes they do. As stated, it isn't an opinion. There is actual data being used. What data can you show that says they don't? You can say there are teams that recruit well that don't win the championship, too which I will respond there is only one champion a year and it is always won by a team that recruits well.
A quick statement just because you're hitting on an issue that relates to one of my pet peeves. You can ONLY truly prove something if you have PERFECT information or PERFECT data. Thus, one of the only areas I know where "proof" is the norm is in in the field of mathematics. Actual real, practicing mathematicians spend their entire careers continuing to "build" the language of mathematics by proving propositions about mathematical objects, mathematical syntax, or mathematical structures.
The standard form of such a proof is suppose you have two statements ... statement P and statement Q. A common scenario in math is that we might contend that the 2 statements P and Q are equivalent. To prove such an equivalence we'd have to verify that if P implies Q ... we must also find that Q implies P.
Otherwise, if you ever have uncertainty or imperfect data/information ... then you're in the murky realm of solving what are called "inverse problems." It turns out that for any data, there is an infinity of different models that can "fit" that data. If you had access to perfect data ... you'd ultimately try to add more and more constraints and more and more data until you'd be able to uniquely specify THE SINGLE model that "correctly" represents the data. Unfortunately, if you do not have perfect data, then you never can truly divorce yourself from the infinity of possible models that can describe your data.
For this aforementioned reason ... my bullshit meter starts pinging to 11 whenever anybody claims that they've "proven" anything.
You see, even in science, you have to be ultra careful about claiming "proof." Scientists can tell you A LOT about what things CANNOT be true ... that is to say that they can DISPROVE things (this reflects Popper's perspective on demarcation in science). Furthermore, scientists can "puzzle" fit ... and find models that are CONSISTENT with data (a view that Kuhn was very comfortable with). However, responsible scientists rarely claim that they've "proven" something tangible about the world around them. Of course, this isn't to say that scientists haven't uncovered "knowledge" about the world around them. Scientists have been able to generate specific descriptions about the world around that no experiment has ever been able to disprove. We're not just talking about a few experiments ... we're talking about thousands to millions of varying experiments that have been incapable of finding fault with the given descriptions.
I've read the 247 sports article about recruiting ... and there were many faults in their reasoning. I'm just one person ... finding fault with their article just be considering a few hypothetical scenarios. Suppose you take thousands of critical minds breaking down that same article ... they'll find many more faults. Thus, I'm sorry but that article doesn't PROVE a darn thing.
As dbrocket rightly highlights ... statements saying that "stars matter" or "stars don't matter" are sensational exaggerations. Recruiting certainly matters. However, the PROBLEM deals with how do you QUANTIFY recruiting? How do you rigorously ascertain the correct metrics? How do you rigorously manage uncertainties? These are interesting questions that these sports-related websites fail to address.
I have many friends who are information scientists ... and many of them tell me about how they generate model of things that they have no understanding about from huge quantities of data. I ask them how they can trust their validation procedures of their models ... when they don't even understand what they're trying to describe. You see, the problem with most validation procedures is that they're reliant on circular logic. They essentially use their own models/algorithms to test themselves. If the results fall within an empirically "reasonable" set of bounds ... they "trust" the models/algorithms. The response my friends supply .... "we have to try something." The suggestion being that something is better than nothing. I agree with this sort of premise IF the algorithms are constantly being iterated upon ... improved and updated to account for new data and new understanding. However, many algorithms and models are used ... and they're used as unquestioned black-boxes.
Yes all data in the real world is "dirty". We do not live in a vacuum. Yes, in all studies there is a certain amount of error that is accepted. There is a data set that actually shows that if you want to win championships you need to have high level recruits on your team. Can you offer any data that shows championships can be won by teams that do not have extremely high rated classes?
Layman's terms:
Everyone loves the feel good story about hard work beating talent when talent doesn't work hard. What doesn't get brought up is when "talent" works hard, it beats the piss out of the other guy. The star system is not always right and it is not an end all be all but in today's day and age it's probably more reliable than it's ever been as a predictor.
Everyone loves the feel good story about hard work beating talent when talent doesn't work hard. What doesn't get brought up is when "talent" works hard, it beats the piss out of the other guy. The star system is not always right and it is not an end all be all but in today's day and age it's probably more reliable than it's ever been as a predictor.
But, again, I could just as easily use the criteria "has won a national championship in the 30 years prior" and be close to 100% accurate, especially in the same period as the recruiting rankings. And it's a fairly small group that was won a NC in the last 30 years.
Edit: 1990 was the last time a team won a CFB NC which had not won in the 30 years prior. It was a split NC, not outright (neither Colorado or Ga. Tech had won).
Last edited:
I may have missed it in the article but when looking at the big ten why are the number of rated players different (assume it is having player not rated or walk ons now on scholarship). Most teams on the list have 85 players, a few in the upper 70s with Iowa the least rated players at 78. That alone skews the big ten rankings, assume if we had 7 more 2 or 3 star players we would be over Wisconsin in that ranking.
What makes anyone feel like Iowa should be able to recruit at least ten 4 & 5 star recruits every year. I don't think in 100 plus years of Iowa football existence we have accomplished this once. If we did it would be under Eve in the 50s. I lived through these times and the University would never allow the coaches now to recruit that kind of student. I watched as many all Americans didn't ever see the inside of a classroom.
They matter to a fan because a fan doesn't see and evaluate all or most of the prospects.
They matter not at all to college coaches who are evaluating prospects based on their experience and knowledge. To think otherwise is just silly
I can see both sides of the fence with this. The best kids we've ever had he....that went on to solid NFL careers....were very lowly regarded star wise. Blake larsen...perhaps the most highly recruited kid we have ever landed..5 star...couldn't find the field if you pointed him to it.
I see what Bell is saying but the teams in the final 4 this year were loaded with 4 and 5 star guys.....but so was Michigan who had the same record as us....primarily 2 and 3 star guys.
Inexact science
I see what Bell is saying but the teams in the final 4 this year were loaded with 4 and 5 star guys.....but so was Michigan who had the same record as us....primarily 2 and 3 star guys.
Inexact science
Exactly. To the coaches, what matters is whether they got the guys they most wanted - or, how many of them they got. If we got the top 20-25 guys on our coaches' board, that's a phenomenal year. It doesn't get better. It might not be a top 10 class nationally (especially if they were rock solid commits early on), because they're evaluating the players and star rankings largely measure the demand for the players.
I can see Bell's point. First, no coach worth a damn chooses who to recruit based on stars. They do their own analysis.
Second, there's a lot more to consider than simply talent. Epic talent with a cancerous demeanor or no work ethic is bad.
So, for him, stars don't matter.
Second, there's a lot more to consider than simply talent. Epic talent with a cancerous demeanor or no work ethic is bad.
So, for him, stars don't matter.
Don’t matter....non football people evaluating. Any person with a morsel of football sense can see the Top 20 kids (5*). Absolute crap shoot after that. The average star rating for last years super bowl for the starters was under 3. Websites need to glamourize it to make money
working against us:
Minimal population area.
Weather.
The landscape..... corn and beans ..... beans and corn. And oh yeah lots of piggies.
Most all teams in the US get pretty good exposure now as compared to 1970s/1980s. Cable, internet, 24/7 round the clock as much as one wants.
But at least our nickname is not Gophers or Cornhuskers (when will they change that just for image sake?).
Kind of like stock pickers. Sometimes you find an underfollowed one that goes on to be a multi-multi bagger.
How well does a rural kid get evaluated anyway? Don't know what Drew Ott was but damn he had a motor.
