Posted by rgerkin on 12/5/2013 4:08:00 PM (view original):
I'm pretty sure there are at least two kinds of two-point shots, because PE definitely improves two point percentage (at least somewhat), and I can't imagine that includes shots near the basket because then what is LP for? Mid-range jumpers are definitely the lowest percentage shot, but free throws are the highest so if you especially if you see yourself in bonus/double-bonus situations he could still be a really efficient scorer. However, at high D1 I'm not really sure what it takes in ATH/SPD/BH to be that kind of slasher player who make a good percentage and draws a lot of fouls. I would track his True Shooting Percentage (0.5* PTS / (FGA + 0.44*FTA)) and see how it turns out relative to your other players.
(warning - this has nothing to do with the tracking of the player and can be safely ignored by anyone interested in the tracking of the player)
ill absolutely track his TSP, that is what i am super curious to see myself :)
i am not saying there aren't two kinds of 2 point shots, but i do question the logic there. PE definitely improves 2 point percentage, including for bigs. but this could be true for all kinds of 2 point shots, lp certainly does it to, there is no reason they have to be mutually exclusive?
one thing i noticed in the early days, before lp mattered for guards, was that killer lp/per sfs always, always underperformed where i hoped. i had an olden days 80+ ath/spd/lp/per guy in d2 (which was way harder than getting one now), which is more or less the most impressive d2 sf i ever saw since you couldn't recruit 1 and 2 star d1 players. but he just wasnt that great on offense. that made me think lp OR per was used to determine the kind of shot, never both. i hoped a certain effect existed - namely, that there were multiple kinds of shots he could take, and not only did he just take a mix of them, but the ability to take a mix of them and improve through versatility, would matter. i never could find any evidence it did and it really looked like it didn't. further, it seemed like whatever the kinds of shots were, it was basically lp OR per that mattered, not both - never both - maybe per a teeny bit for bigs. but in these days, going from say 1 to 50 lp or per in a 90+ spd/per or ath/lp type player, it clearly helps. you'd think a big who had killer ath/lp and just took shots in the paint, if he added the lowest % mid range jumpers, wouldn't help him - when his mid range jumper shot ability was mostly based on per. but it does, so that would suggest lp matters for mid range jumpers, which sort of makes no sense, or there is just one kind of 2 point shot.
not really sure where i am going with that - but the increase of the importance of lp on high per/spd guards and per for high ath/lp guards did increase the complexity of scoring and make it a lot more interesting. in the simpler world, it didnt really seem to me like you could get bigs to take jumpers, or get guards to slash, or get sfs with both lp and per to really outperform a guy with one of the two. this makes me think there were not two kinds of 2pt shots back then - otherwise through some decision mechanism to decide what shot to take, bigs would decide to take jump shots while guards would decide to slash, and lp would have helped guards while per significantly helped bigs.
to me, i often wonder if there are "guard 2 point shots" and "big man 2 point shots", even more than i wonder if there are close to the basket and far from the basket two point shots. for some reason, high ath/spd guards with low per draw more fouls than high ath/lp bigs. the only things i can think of (but im certainly not sure i have them all) that could justify that difference are 1) two different equations, 2) speed causing the extra fouls to be drawns, and 3) being defended primarily by a guard, not a big, somehow results in more fouls being drawn.
with issue 2), its certainly logical - and for guards, im pretty sure it does. but i really have not noticed much difference in bigs offensively for spd, ive always considered it basically meaningless as far as that goes. this is something we could probably figure out between us. any interest in running a regression on fouls drawn per fga at the 5 based on speed? seems to me fouls drawn might be tough, so fta could be used, but care would have to be taken as a bunch of things muddy that water - the 5's ft shooting ability, the systems being run by both teams, etc...
with issue 3), the guards have all the ath/spd/def of the bigs, in many cases. so what, is it the SB? i am kind of skeptical. but a regression could probably be run on that, too - really, would be interesting to see by position - the correlation (or lack there of) of the SB of the defender, and the fouls drawn per shot attempt.
anyway, im not sure how much this matters, but if there are different formulas for 2pta of guards or bigs (or similarly, if the position of the player in question is part of the formula), that would be pretty interesting, at the least. also, ive wanted to ask you for like a year, if you would have any interest in looking at what plays into foul drawn - where are the closest correlations - because it seems to me lp/per dont play much role, and it would be pretty interesting to know if its really just ath/spd/bh? and how they compare in importance.
so back to the two kinds of 2s, if there are two kinds of 2s for bigs, how would that work? for starters, the complexity of making that work is so much higher than 1, my gut said admin probably went the simpler route, as is so often the case (you even used that argument to largely convince me of your position on tempo). if there are 2 kinds of shots, do all bigs take both kinds? if so, how do you figure what % of close shots vs far shots are taken? they cant all be the same for all bigs, that would make no sense, right? anyway, now if there is a jump shot and a lp shot, and bigs take more jump shots the higher level of ability they have (which seems about the only logical way i can think of, to figure out the %s of each), you would expect a curve where small amounts of per are absolutely useless, and then there is a discontinuity at some level. this is something that should be able to be verified, if you are up for the statistical work :)
my opinion is that per is not absolutely useless up to a point, and that while per is CLEARLY valuable to bigs in terms of fg% in this day and age - that spd is not - even if its on a 50 per big. to me a 90 ath/lp big is not good enough in high d1 to be a lead scorer. its one of the biggest imbalances in the game, IMO. but with 90 ath/lp and 50 per, now he can be a lead scorer - albeit probably not as effective in the upper high scoring levels (20 ppg or higher) as a guard - but still, hes viable there. however, i don't think he needs that spd. if there were two kinds of shots, you would expect that 50 per really only to help on the jump shot, and you would expect the jump shot to rely on spd and bh like 3s do. so, my feeling is that bigs are just taking 1 kind of shot, its just such an easier way to go, and from the way the ratings help him improve, i don't really see how it could be 2 shots for bigs.
one other thing that makes me thing guards and bigs take different 2s is the ratings that play in. spd and bh helps guards in terms of 2pt%, there is probably no disagreement there. low per guards with crap spd/bh take terrible 2s, while high spd/per guards can be very effective in terms of fg% just taking 2s. plus, guards with higher spd/per who do take 3s, when you back out the 3s from fga, clearly shoot better from 2pt land than their lower spd/per counterparts. how does spd help guards so much and bigs so little, just in the set of 2 point shots? there could be two totally different kinds of 2s but even then, within the set of just jump shots, its still true - why?
anyway, im sure i am not making as much sense here as i could. so in conclusion, my opinion is most likely there are different 2pt shooting formulas for players at different positions - minimally guards and bigs. if that is the case, it already gives flexibility in how 2 point shots work. i think i initially believed there were 2 kinds of shots because the way different players performed on 2s worked so differently. but now, i believe that is because different positions are figured differently (same goes for 3pt shots). now, if you already have different 2s and 3s for different positions, that is pretty complex - and we have a variable to control how many of each type is taken - which makes me really doubt they would take that next step and make different kinds of 2s for each position. figuring out how a center decides how many jump shots to take vs paint shots is pretty tricky, especially with 0 user input, and you have 5 positions to deal with... on top of that, i dont see observations that support that kind of system. if you did, id expect to see things like low amounts of low on guards or per on bigs having no impact, and then a major jump. id expect in the olden days for lp to actually matter for guards, and per to be meaningful for bigs. to me, it just doesn't really add up.