APBRmetrics

Guys, I am quite new to your board but have been hearing about advanced statistics for quite some time, and stumbled accross this. I am a coach at a small Division 1 college, and have just conducted an examination (not sure if I am scientific enough for a "study") to determine how the 5 shooting statistics compare in terms of "comparative rarity". Basically, it boils down to the following question: We have been saying for years that for example 50% FG shooting is "good". Well, that follows with the unasked question: How good is 50% shooting or 40% from the 3P line, etc.

So what I did was the following:
I plugged every significant player (418) (I ommitted walkons or players with negligible statistics) for the last 4 years of our conference into a spreadsheet using his shooting/scoring data. After calculating the the FG, 3P, FT, eFR, & TS% (along with all shots/game data for my own purposes), and then calculated percentile scores for each list of data - to determine using the same block of data, "levels of good". While I am sure I may have made a mistake or two because I am not much of a statistics (the academic sense) guy, here is what I found:

Percentile Equivalency
% FG 3P FT eFG TS%
98 0.606 0.500 0.901 0.627 0.652
95 0.568 0.434 0.853 0.592 0.622
90 0.536 0.407 0.806 0.566 0.602
85 0.513 0.393 0.790 0.547 0.586
80 0.498 0.382 0.772 0.533 0.572
75 0.479 0.374 0.758 0.525 0.559
70 0.468 0.363 0.739 0.516 0.552
65 0.455 0.358 0.725 0.517 0.549
60 0.442 0.350 0.703 0.500 0.539
55 0.433 0.333 0.691 0.490 0.530
50 0.423 0.325 0.673 0.481 0.524
45 0.417 0.316 0.657 0.475 0.513
40 0.408 0.308 0.639 0.468 0.505
35 0.401 0.295 0.629 0.459 0.497
30 0.392 0.284 0.607 0.450 0.489
25 0.385 0.273 0.583 0.439 0.480
20 0.374 0.250 0.569 0.427 0.473
15 0.363 0.233 0.539 0.412 0.462
10 0.342 0.208 0.498 0.391 0.441
5 0.313 0.150 0.433 0.372 0.414

Mean 0.438 0.338 0.676 0.490 0.531


Median 0.421 0.294 0.667 0.479 0.520


I am interested in feedback, academic/mathematic criticisms, or ways to exapnd this study and make it more usable. However, I have felt like for years, that we have spoken in terms of a vague understanding based off of experience. I simply wanted to obtain an answer for myself - how good is a number? While there is no hard scale, I may have assisted myself by at least creating a comparative scale. Now keep in mind, this is not for pro players, but simply a usable study for me personally in low-NCAA I basketball. If necessary and with assistance, I can post my entire excel document for examination. If I did make a mistake, I would want to correct it.

Here is that table plotted, for reference:



Whenever we talk about field goal percentages, we always need to consider the usage level of the player and the shot types/guardedness as mitigating factors.

Thanks for sharing your data. Your main question is "how good" are these various stat level achievements. Percentile rankings of players are interesting but I'd also suggest converting the various stat level achievements into points per game above the mean given the players frequency for a specific statistic compared to the mean frequency or looking at the value difference if they were at mean frequency and then also having a measure of the value gain or loss from that because the player's frequency was actually different from the mean.