Posted by Iguana1 on 8/10/2010 2:51:00 PM (view original):
I agree that stamina does have an affect on better FT% but I wouldn't necessarily take player C to make a FT over player A.
First off, the % can be pretty random when you're looking at such a few amount of FTs.
Player C (Gass) has shot 14/14 in games in which playing 19+ minutes. When playing 18 minutes or less he's 23/32, 72%.
Player A (Thompson) is 11/13, 85% when playing 19+ minutes. In games in which he's played less than 19 he's 13/23, 57%.
That's a combined 25/27, 93% for 19+ minutes vs. 36/55, 65% for under 19.
If I just judged on that arbitrary cutoff of 19 minutes per game, I'd assume the more minutes played the better FT shooter.
I noticed these damning statistics as well, and completely agree the limited individual totals leave much to be desired. However, only 2 players on that team have over 50 attempts on the season, therefore I selected the the larger sample size despite the lower data set (since a probability analysis with sample size of 1 is meaningless, and using 2 just seems lazy). The individuals were selected because they had both the similarities and differences needed to succinctly demonstrate my argument, and are not the only examples on which I based my conclusions.
Some of the players on my DIII team reflected these assumptions last season, but I am cautious of heavy dependence on them (individual game stats erased) despite allowing for a sample size of 4 individuals having overall attempts ranging from 58 to 89. F*** it, lets run the numbers,"a failed hypothesis is never in vain because it leaves you one step closer to the truth." Examples: (ST and FT are adjusted to reflect the totals at seasons end)
FT grade FT percent Min. per game Stamina
Player A B .753 23.9 79
Player B B- .793 15.8 89
Player C B- .696 26.9 85
Player D C .640 28.3 90
Player E C- .569 28.9 95
Outcome - The model correctly predicts the relationship between the variables/percentages of players A, B, and C. While the relationship between C and D exposes the crudeness of this exercise because each player has an advantage in one of the variables defined as being indicative of success; therefore, since neither variable is weighted, no outcome can be made by measuring the difference between FT percentages.
Interpretation - He must be Neo because he can see the matrix! Player B (with lowest min. per game & highest ST) had highest FT% amongst players with similar/better FT grades. Also, the relationships between the pairs of A & C and D & E can be analyzed simultaneously because in both instances the players with higher FT grades and lower minutes per game (A and D) had better FT% than the players with higher ST (C and E).
Conclusion - This is only the second (albeit the 2nd successfullish) testing of these parameters, and most likely doesn't mean s***. However, this becomes a touch more interesting given each player's position and eligibility: A (redshirt fr, sg), B (soph, sg), C (soph, sg), D (sen, c), and E (sen, c). Lastly, while I was doing this, the room darkened, techno music began playing in the background, and logical proofs slowly floated through the air, just like it happened to that kid in Little Man Tate and zerocool from Hackers. Seriously.
http://www.whatifsports.com/hd/TeamProfile/Ratings.aspx?tid=3068