by bigh0rt on Sat Jan 26, 2013 12:36 am
Med, I was able to find the following data, which I think we can make even more accurate tendencies, AS WELL AS accurate shooting ratings...
Shooting Attempts per 40 minutes with the following breakdowns: At Rim, 3-9 ft, 10 - 15 ft, 16 - 23 ft, 3 PT, with FG% for each range! I also have the league averages for each of these, which will allow us to create a median score, as well as max scores. We can use attempts as a basis for tendencies, and FG% for the shooting ratings (for those who have played enough games). I have them all in an Excel file, currently, sorted alphabetically by player first name (I believe this is how Roster Editor sorts, too).
For example, the #1 shooter by volume from Inside is Nikola Pekovic (7.4 per 40 min), and he shoots 63.1% from there, while the league average is 2.4 per 40 min, at 64.3%. So for Pekovic, he would have a very high Inside Shot Tendency, with an average Inside Shot Rating. LeBron James, meanwhile, takes 6.8 shots per 40 min Inside, making 78.0% of them, which is a superior % (9th among players who have played 15+ G). So LeBron would have a high Inside Shot Tendency (again, he is 6.8 compared to 2.4 league avg), and a high Inside Shot Rating (his 78% compared to 64.3% league avg).
In my preliminary formulas, James would receive a 94 Inside Shot Rating (FG% x 1.2), Pekovic would receive a 76 Inside Shot Rating, and the League Average would produce a 77 Inside Shot Rating, which seems pretty fair to me. It can be tweaked, of course, but League Average, in my opinion, should probably be around 75 Rating, no? (Other random Examples: Luol Deng 80, Jameer Nelson 74, Jerry Stackhouse 67, Mike Conley 60) Again, if this results in an overall increase in Inside Shot Rating for players, that's what the Sliders are for. At least individually the ratings will be accurate, then Sliders will adjust them globally across the board. Also, with proper Tendencies, a general increase shouldn't dramatically affect the way the CPU plays. Even if a player has a nice Inside Shot Rating, if their Tendency and Sliders dictate they only shoot their 2.4 times per 40 minutes, that's the happy medium we need to find.
If this all seems too honky for you, I may start working on my own roster with these ratings to see how it plays. I truly believe that using data is the best way to produce the most accurate results. It's a considerable amount of work and tweaking to find the right mix, but once you do, you're head and shoulders above everyone else. Your roster's there visually already, and I think we could completely change the way it plays for the better as well. Let me know.