EricLA
Cronus
- Joined
- Aug 24, 2011
- Messages
- 15,141
- Reaction Score
- 82,947
OK... so clearly I had some extra free time over the holidays and decided to do my own composite player AND composite class ranking.
The first thing I did was to list out the rankings for each player by each of the 4 services that I could find - Hoopgurlz, ASGR, Prospect Nation, and Blue Star. For the individual player composite, it was simple - I took the average of the 4 rankings services for each player. Fortunately there was enough data that there were only a few ties in the entire list of 188 players (Recee' Caldwell and Gabby Green tied at #6, and Mikayla Venson and Chatrice3 White tied at # 23). This was the easy part. I will put the top 50 below in the next post.
A lot of discussion occurred for players like Kia Nurse who are not American, and therefore are not ranked in either HG's class or PN's class. PN was easier - They have her listed under their "foreign player" category. And she's listed as 5 stars. So to be conservative, I put her behind the lowest 5-star player on the American players list - at #11.
Hoopgurlz has her listed as a 3 star player. Clearly they've not evaluated her in any meaningful way. So I took the average of the other 3 services, who had either rated or ranked her, and put her in as a composite ranking of 12. FWIW I believe she's one of the top 5 kids in the entire class, but that's my opinion, and not empirical enough for my analysis. There were a few other foreign players NOT ranked by either HG or PN so I used the same formula to come up with their composite ranking for each school they decided to attend.
I get that there was a lot of discussion the last time a poster compiled their version of the composite rankings. with the exception of ranking the foreign players, I probably used formulas similar to the other posters.
Doing the class rankings proved a bit more challenging. Clearly, a team who landed even one recruit of the level of Wilson, Mitchell or Turner (the top 3 kids) should have a stronger class than a team who landed 4 kids outside the top 75 or 100, since kids in the top 5, top 10, top 20 etc. are much more impact players than kids outside the top 50, 75, or 100.
But how much better was a kid ranked 25 or 30 over a kid ranked 45 or 50? So I used a sliding scale, weighting the top kids with a larger amount of points and decreasing from there. So kids in the top 10 had a larger sliding scale decreasing from 1-10, same for 11-20. Once the players were ranked outside the top 20, the sliding scale was much smaller and consistent all the way down to the 190th ranked player.
In the case of lists like HG and PN who don't list anyone outside the top 100, I again used the average of the other 2 services to rank players from 101-190 so when I took the average of the 4 services, it would be a consistent number.
So for example, total points for the top player in the class (Wilson) was 1,180. whereas the #10 player in the class (Drummer) received 830, and so on.
Clearly no one list is perfect. I'm sure people will have opinions on whether I should or should not have used the data and assumptions I did, and feel free to give your input. I will make tweaks in the future when I attempt this again.
The first thing I did was to list out the rankings for each player by each of the 4 services that I could find - Hoopgurlz, ASGR, Prospect Nation, and Blue Star. For the individual player composite, it was simple - I took the average of the 4 rankings services for each player. Fortunately there was enough data that there were only a few ties in the entire list of 188 players (Recee' Caldwell and Gabby Green tied at #6, and Mikayla Venson and Chatrice3 White tied at # 23). This was the easy part. I will put the top 50 below in the next post.
A lot of discussion occurred for players like Kia Nurse who are not American, and therefore are not ranked in either HG's class or PN's class. PN was easier - They have her listed under their "foreign player" category. And she's listed as 5 stars. So to be conservative, I put her behind the lowest 5-star player on the American players list - at #11.
Hoopgurlz has her listed as a 3 star player. Clearly they've not evaluated her in any meaningful way. So I took the average of the other 3 services, who had either rated or ranked her, and put her in as a composite ranking of 12. FWIW I believe she's one of the top 5 kids in the entire class, but that's my opinion, and not empirical enough for my analysis. There were a few other foreign players NOT ranked by either HG or PN so I used the same formula to come up with their composite ranking for each school they decided to attend.
I get that there was a lot of discussion the last time a poster compiled their version of the composite rankings. with the exception of ranking the foreign players, I probably used formulas similar to the other posters.
Doing the class rankings proved a bit more challenging. Clearly, a team who landed even one recruit of the level of Wilson, Mitchell or Turner (the top 3 kids) should have a stronger class than a team who landed 4 kids outside the top 75 or 100, since kids in the top 5, top 10, top 20 etc. are much more impact players than kids outside the top 50, 75, or 100.
But how much better was a kid ranked 25 or 30 over a kid ranked 45 or 50? So I used a sliding scale, weighting the top kids with a larger amount of points and decreasing from there. So kids in the top 10 had a larger sliding scale decreasing from 1-10, same for 11-20. Once the players were ranked outside the top 20, the sliding scale was much smaller and consistent all the way down to the 190th ranked player.
In the case of lists like HG and PN who don't list anyone outside the top 100, I again used the average of the other 2 services to rank players from 101-190 so when I took the average of the 4 services, it would be a consistent number.
So for example, total points for the top player in the class (Wilson) was 1,180. whereas the #10 player in the class (Drummer) received 830, and so on.
Clearly no one list is perfect. I'm sure people will have opinions on whether I should or should not have used the data and assumptions I did, and feel free to give your input. I will make tweaks in the future when I attempt this again.