In partnership with CBSSports.com
Online Now 392
You have no favorite boards.
The most viewed topics.
The most replied to topics.
The most up-voted topics.
The most down-voted topics.
The most up-voted posters.
The most down-voted posters.
The most followed posters.
First off I think it was great that they listened to the customers and made the tweak but I still cannot understand these rankings.
1. Anthony Jennings is a 4 Star on ESPN, 247, and Rivals and yet he is a 3 star on the Composite Rankings? huh?
2. In the 247 Rankings (not composite) how is Ohio State 2 spots ahead of us?? Can someone explain this one?
LSU 15 -4 Stars and 7- 3 Stars
Ohio St. 11 - 4 Stars and 7 - 3 Stars
Not bitching at all by the way, just trying to wrap my head around what formula could yield these results.
This post was edited by tigerinatx 20 months ago
I can explain pretty easily.
#1. Jennings just missed the cutoff index for 4-star. The cutoff is .8900 which produces 300ish 4 stars. Which is in line with what JC is looking for in the total for this stage. .8900 will probably rise to 350 by the end. Jennings is really close.
#2. Ohio State's 4 stars have much higher value than LSU's. We use the guassian distribution formula which places a great emphasis on the value of your team's highest rated recruiting. Ohio State also receives added points for having a 5 star. The key is to look at the ratings much more so than the star ratings. CHeck out the help icon on the team page for a full description of the team ranking formula. It is slightly quirky in the early stages (for example, comparing Team A with 22 commits vs. Team B with 12 commits), but it is an outstanding formula for Dec/Jan and NSD.
247Sports and BOL updates: http://twitter.com/sbterry247
"I guess you've just gotta find something you love to do and then... do it for the rest of your life." - Max Fisher
Lucky - tOSU does not have a 5 star on the 247 Rankings.
I still don't understand. We have 4 more 4 stars and a higher average per player.
Lucky, that still doesn't make much sense to me. For instance, Su'a Cravens has a 98 rating, and Vonn Bell also has a 98. One is the #1 S and #9 overall, one is the #2 S and #24 overall. You are saying that Cravens is worth more because he is higher ranked but that they are essentially the same value at a 98 rating. That seems quite silly to me.
"They're on our left, they're on our right, they're in front of us, they're behind us....They can't get away this time." Chesty Puller
No. I am not saying that. Both are worth the same in the formula, because both have a 98 rating.
My bad on the 5-star... looking at the wrong list.
From a simplistic explanation standpoint, forget the stars exists. They are only for illustration. The Team Ranking formula is based exclusively on "rating" (ie 98, 96, 89, etc). Secondly, the prospects at the top of your list, have a greater impact on your ranking (see the Gaussian bell curve info for details). The disparity between LSU and Ohio State is because the class is rated higher at the top. When LSU lands a few more of the elite types, those players will have MAJOR impact on their team ranking.
There is the true flaw IMO. That is why a team with a higher volume and a higher overall average can be ranked lower because the bell curve over-values higher rated players. That is the only way that could happen. the value of a 98 must be proportionally greater than a guy rated as a 97 and so on an so fourth.
Basically what you are saying is that it is like a Track and field scoring system to a certain degree. If you finish 1st which would be like a 100 rating then you get 10 points, whereas if you finish 3rd like a 97 you only get 6 points and if you finish 8th like a 90 which is still very good you still only get like 1 point.
Do I have it right now?
Thanks again for trying to clarify Lucky.
Ohio St gets enough credit for having a few "higher rated players (4stars)" to overtake us having 4 more high rated player (4 stars) than them and a higher player average. Do you not get credit for having 4 additional 4 stars (highly rated players)?
I'm just going to write this off as one of those things that I will not understand.
This post has been edited 3 times, most recently by tigerinatx 20 months ago
I don't think any of us without degrees in applied statistics will completely understand, but I think that's just the nature of the beast. I just googled "Gaussian bell curve" and my eyes glazed over after about 5 seconds.
This post was edited by Gravitiger 20 months ago
There are flaws to all formulas, but we do have one -- eliminating as much subjectivity as possible -- and we like the fact that our formula gives greater value for the high value targets. Nothing's perfect, but I really like what the existing formula does.
So where a guy ranks individually within his position or overall is not factored in at all?
No I agree that all systems are flawwed....and I love the steps y'all are taking to try and bring as much objectivity to the rankings as possible but at some point it is all kind of a futile effort because every piece of data that factors into any formula is based on subjectivity.
Not at all.
This post was edited by Lucky 20 months ago
Seems like the biggest thing affecting LSU ranking on both 247Sports and 247Composite is Logan Stokes, Kennard Swanson and Lewis Neal. But what I really like is all 3 are pretty close on both sets of rankings where there really isn't any disparity.
I had a couple of emails asking this so wanted to go ahead and post in case anyone else had that question.
Okay, I disagree with that approach but I still think y'all are the best, and thanks again for answering all our questions.
I appreciate that. Again, no way to be perfect. However, we can definitely improve in the state of LA and I think you will see that improvement with the latest round of hires.
Some things I don't understand about the 247 Ranking LIsts:
1. Why do the all (247 rankings and composhit) Junior College Links under Player Rankings show the 247 Top 247 list; (when did Nkemdiche enroll in JC?)
2. Why do all (247 rankings and composhit) of the State Links under Player Rankings show the 247 Top 247 list; (when did Nkemdiche move to Louisiana, Texas, Georgia and Florida?)
3. How is Tevin Lawson the 27th ranked player in La. on the 247 Louisiana List with a 247 score of 88, when not one, not two, not three, not four, not five, but six players with a score of 87 are ahead of him in the 247 La. rankings; (algorithm problems?)
4. How can anybody take the ESPN rankings seriously? Christopher Taylor as the #8 ranked player in La. and a top 10 guard nationally, his best offers are CUSA and Sunbelt teams (no interest from major conference affiliated teams); or
5. Melvin Jones as the 65th best "Athlete", looks like someone saw film of him playing QB (and said, "meh"); and on top of this - ESPN historically does very little, if any, updating of rankings once their lists are published. (Old saying in comp. sci. = "Garbage IN, Garbage out")
It's nice to break out the shiny new toy (sarcasm), but it would be appreciated if the underlying product functioned properly.
Thank god for the info and updates from Shea and Sonny, because so far 247 "meh".
Disagree to some extent. It appears that Melvin Jones', Rashard Robinson's, and Josh Boutte's ridiculously low position rankings on ESPN have a bigger negative effect in the composhit rankings.
Your coverage of LA is already better than Scout's.
"That was Death Valley. It's where opponent's dreams come to die." - Miles
Hats off to tech guys, in less than 15 minutes from the posting of this complaint, the problem is being fixed. Thank you.
We aren't perfect but we will always listen to the people that make our Network what it is and that is you the subscribers!
My problems with the whole star system rating is probably the same as many people but maybe I am alone on this....We see LSU or ( insert your team here) with a lets say an 11th team ranking . We came in second last year and are ranked first in polls this year. This repeats over many years. there is such a difference in where you place vs what you recruiting ramkings over years are at. I could do the math on rankings last 4 years vs last years finish but I wont because that goes agaisnt this argument.That goes for up and down rankings preseaon and end of year, like Oklahoma pre season # 1 in polls last year. I think most people like me look at recrutiong class and wanna know where we are gonna finish without to much math. Like last year we were number 15 in recruiting this year composite were are number 11th...utoh not looking good.......Now this is apple to oranges in so many ways but its still the nature of the beast.
There is so many things that figure in like talent evaluation ,some are better then others.How well kids are coached up.The system they are in.The bias by some evaluators.The list goes on and on.I think most people want a system that predicts a winner or at least gets close. While geaux 247 mathmatically may be far superior to my ramblings......They will not be close on what kind of team most schools have in the future and thats the difference in stats vs reality. This is not a fault of geaux 247 rankings there math is pretty good....The fault is mine for my perceptions,wich is why I listen to S&S and evreyone else on this site,watch videos and read lots vs believing in a star system....
I don't think this is quite right Sonny.
The gaussian formula gives greater importance to the higher rated players and less to the lower rated players.
Therefore our three lowest players would have a very small effect.
The biggest reason for our low ranking is us not having any 5 stars or really high ranked players. We have a lot of depth in this class but limited star power according to the sites.
I think we'll see the team ranking change tons before NSD. Players will continue to be evaluated, we'll add some highly rated guys and other teams will start to take lower rated players to fill their class when they miss on the guys they want. I'll wait until NSD before I decide if I like the formula.
“There is no off switch on a tiger”
247Sports In partnership with CBS Sports