'17 NC DE Matthew Butler (UT Signee)

#576
#576
TennMan101 said:
You're right. The composite ranking takes the value of the four major recruiting sites and gives you a median output. The 247composite is the most accurate analysis of a player you can get, in my opinion
Click to expand...

Unless one of the four services has someone really skewed , one way or the other.
 
#577
#577
TennMan101 said:
Not true, IMO. Not only do they give you their take on a player, but they also give you the median of all four sites. That is something none of the other sites do!
Click to expand...

Actually, they give you the mean. The median would be the number of stars at which half the sites are higher than and half the sites are lower than.
 
#579
#579
drvenner said:
Actually, consensus implies agreement. And, that is not the case with many recruits.
Click to expand...

I've been misusing the word then. Electoral college...democratic...majority...tiebreaker...whatever works. :boredom:
 
  • Like
Reactions: 1 person
#583
#583
bignewt said:
Thats my problem with the composite is that at least one out of the four hoses just about every recruit if i had to choose id say 247 and rivals are the most accurate. Just my opinion, gracias!
Click to expand...

I couldn't agree more with this...I would like to see them add a condition to their algorithm that eliminates one of the four from a given players composite if that sites ranking of said player is significantly different from the other three...it wouldn't even be that difficult to program in...

Just this year you have Locklear (83 on 247, high 3 to 4 star everywhere else, mid 3 on the composite), Labruzza (72 on ESPN, high 3 to 4 star everywhere else, not a four on the composite) and Murphy (low 3 star on Scout, high 3 to high 4 everywhere else, not a four on the composite)...

One bad ranking will tank a players composite...and it frustrates me because I believe in a lot of cases (like Labruzzas ESPN ranking) the ranking is low because the site hasn't actually evaluated the kid...these sites don't have the resources to really look at every kid in depth so sometimes they just slap a rating on them
 
  • Like
Reactions: 1 person
#584
#584
thunder5 said:
I couldn't agree more with this...I would like to see them add a condition to their algorithm that eliminates one of the four from a given players composite if that sites ranking of said player is significantly different from the other three...it wouldn't even be that difficult to program in...

Just this year you have Locklear (83 on 247, high 3 to 4 star everywhere else, mid 3 on the composite), Labruzza (72 on ESPN, high 3 to 4 star everywhere else, not a four on the composite) and Murphy (low 3 star on Scout, high 3 to high 4 everywhere else, not a four on the composite)...

One bad ranking will tank a players composite...and it frustrates me because I believe in a lot of cases (like Labruzzas ESPN ranking) the ranking is low because the site hasn't actually evaluated the kid...these sites don't have the resources to really look at every kid in depth so sometimes they just slap a rating on them
Click to expand...

When the ratings are limited to 3 vs 4 it's hard to determine what an outlier is. I think the key is in interpretation. People (fans) use the rankings for far more than is justified by the way they are developed. It's like arguing over whether the water is cold or cool or tepid. There just isn't enough discrimination in the ratings to cut it that fine.
 
#585
#585
drvenner said:
Unless one of the four services has someone really skewed , one way or the other.
Click to expand...

drvenner said:
Actually, they give you the mean. The median would be the number of stars at which half the sites are higher than and half the sites are lower than.
Click to expand...

drvenner said:
Actually, consensus implies agreement. And, that is not the case with many recruits.
Click to expand...

Geez man, win a World Series and all the sudden you're Mr. know it all :)
 
  • Like
Reactions: 2 people
#586
#586
I watch 45 seconds of Hudl highlights then make a definitive opinion from which I will never be swayed about the kid's talent and future position.

And I will post that opinion in every thread.
 
  • Like
Reactions: 4 people
#588
#588
247 composite would be better if the formula had a way to factor out a ranking that is far removed from the other sites. Write an algorithm using standard deviation and drop one of the rankings of it is way off from the other 3. I think that is important, especially considering they use ESPN. ESPN is way late to being proper evaluations of some players. And even then they can be half-assed evaluations many times.
 
  • Like
Reactions: 2 people
#590
#590
therickbol said:
247 composite would be better if the formula had a way to factor out a ranking that is far removed from the other sites. Write an algorithm using standard deviation and drop one of the rankings of it is way off from the other 3. I think that is important, especially considering they use ESPN. ESPN is way late to being proper evaluations of some players. And even then they can be half-assed evaluations many times.
Click to expand...

This ^^^^^
 
#593
#593
I remember when rankings from polls and computer analyses were used to determine who would play for #1. There waa a journalist poll and a coaches poll and several computer "polls". Whenever a computer ranking disagreed with the human rankings, everyone said the computer algorithm was flawed and should be omittes. Maybe it was the only one that was right (you never really know) , but our emotions surpass whatever analytical results don't agree with them. It's really an interesting phenomenon.
 
#594
#594
thunder5 said:
i couldn't agree more with this...i would like to see them add a condition to their algorithm that eliminates one of the four from a given players composite if that sites ranking of said player is significantly different from the other three...it wouldn't even be that difficult to program in
Click to expand...

therickbol said:
247 composite would be better if the formula had a way to factor out a ranking that is far removed from the other sites. Write an algorithm using standard deviation and drop one of the rankings of it is way off from the other 3. I think that is important, especially considering they use espn. Espn is way late to being proper evaluations of some players. And even then they can be half-assed evaluations many times.
Click to expand...
😎
 
#595
#595
VolsnCards5 said:
Newb....I don't even need 30 seconds
Click to expand...

and the ladies be like...
dsppntmnt.gif
 
  • Like
Reactions: 2 people
#597
#597
therickbol said:
247 composite would be better if the formula had a way to factor out a ranking that is far removed from the other sites. Write an algorithm using standard deviation and drop one of the rankings of it is way off from the other 3. I think that is important, especially considering they use ESPN. ESPN is way late to being proper evaluations of some players. And even then they can be half-assed evaluations many times.
Click to expand...

SPC, I love it!!
 
New posts
You must log in or register to reply here.

VN Store



Back
Top