Did I say it was nothing but basketball or working out?? I said he clearly hasn't taken much of a break from basketball... It's the offseason, guys are free to have a life away from basketball (as they are during the season as well)!!
yeah well it pretty much came off as him being around basketball or working out in some fashion for the majority of his time
because it is supposed to be "true" shooting percentage, not scoring ability per possession. Let us to following your thinking to talk about TS% as scoring ability per possession. Player A and player B each has 8 possessions, Player A. 8 possessions. he made 8 of 8 3 pointers, scored 24 points. Player B. 8 Possessions, he made 6 of 6 3 pointers, got 6 FTs from fouls beyond 3 pointer line and made 6 of 6 FTs. He scored 24 points. both players have 24 points in 8 possessions. Same scoring efficiency per possession. Not so in your TS%, player A has TS% of 150% and player B has only 139%.
So then what should be the factor if not .44? It sounds to me like you should use efg if you don't like including free throws
Harden had 9.4 drives/gm and 7.2 pts off drives/gm Lowry 9.5, 5.4 Lebron 9.2, 7.1 Irving 8.9, 6.5 You think Lowry, Lebron, and Irving were out of shape last season as well?
That could be complicated, I honestly do not know. One reasonable way is to use data from all previous seasons and do fitting and regression iteratively to get that factor, and make adjustment each year after adding new data from that season. Need to normalize too.
harden did drive to rim less compared to the previous year when he had his mvp caliber season. comparing his drives to other players doesn't really have anything to do with someone being out of shape
Harden last season had 1 drive per game and less than 1 pt per game compared to his players mvp season are you are 100% sure it's because he's out of shape and not because of any other factor like him carrying a heavier offensive load or dwight not willing to play pnr or shooters around him not making their threes?
Never said its cause he was out of shape. I don't even know the exact numbers but from everything I saw he did go to the basket less so I guess that was accurate. There could be a variety of factors but was just pointing out comparing him to others doesn't have anything to do with his conditioning
There's an estimation involved in the formula so you will definitely be able to contrive a situation where the % doesn't reflect actual performance. In the example you cited, you said that player B scored 24 points in 8 possessions. To do that it would mean that he hit 6 3 pointers and he was fouled while shooting 3's on the two other possessions (and made all 6 FT). You could also contrive a similar situation by getting and1's on 3 point shots. That's a pretty unlikely scenario and certainly wouldn't occur over a span of more than a single game or so. Typically if a player has 6 FT then it will occur over 3 possessions rather than 2. That's the more likely scenario and that's how the estimate is factoring it. Under that scenario, you'd have player A with 24 points on 8 possessions and player B with 24 points on 9 possessions. That's the difference that's reflected by the TS% that you cited. The point is that scenario that you're quoting isn't likely to happen and it definitely won't happen over the course of very many games. FG% has similar misleading numbers but those do occur over the course of the whole season. For example, Dwight Howard had a fg% of 62% last season and he scored 0.208 points per possession that he was on the floor. Steph Curry shot 50.4% from the floor last year and scored 0.408 points per possession that he was on the floor. So Curry scored almost twice as many points per possession as Dwight yet Dwight had a FG% almost 10 points higher than Curry. That's a real world example that occured last year. Can you cite a real example of where TS% is off over the course of an entire season?
exactly, it's an estimation, presumably based on historical data. ideally, you would have all the data for and1's and fouls on 3's, and then you could get an exact number, but you're certainly not getting that for past players. now, considering how much play-by-play data basketball-reference has, you would actually think that a perfectly calculated TS% would be possible, but i don't think that's what they report. harden's numbers last year were 837 free throws and 56 and1's. ignoring everything else, that would mean he used 390.5 possessions for 837 free throws, which would be a .467 multiplier. then it's a matter of 3 point fouls and technicals. if got fouled on 30 3's and took 21 tech's, his muliplier would be 365/837 = 0.436 15 3's and 10 techs? 378/837 = .452 so he's presumably somewhere in the vicinity of .44. now one could argue whether technicals should count. it's obviously valuable to have a good shooter for techs, but for a shooting efficiency stat it doesn't make sense to count something you didn't produce with your offensive skills. but it's not going to have a big effect. a quick glance at curry would say that he would have a slightly lower multiplier than harden because he had a slightly higher and1 ratio and probably was fouled on more 3's and took more techs but even a fairly conservative estimate of 0.39 would only increase his TS% from 66.9 to 67.6. so the estimate isn't going to render TS% meaningless. it'll just be slightly off even in somewhat extreme cases. it certainly wouldn't make it less valuable than FG%.
The original author who defined the TS% is lack of basic scientific training. 1) TS% is not "true". 2) TS% has nothing to do "shooting percentage". To define the percentage of a player's score points to the maximum points he could score in those shooting opportunities, the correct formula should be like: points/[2*(2 pointer FGAs)+3*(3 pointer FGAs)+FTAs] or points/[2*(total FGAs-3 pointer FGAs)+3*(3pointer FGAs)+FTAs] it can be simplified to points/[2*(total FGAs) + 3pointer FGAs + FTAs] FTA is the free throw attempt. It considers all 2pointers, 3 pointers, FTs including technicals, and has minimum value of 0 and maximum value of 100%, has no artifact or estimation. Anyway, all these discussions have nothing to do with shooting percentage. This is time consuming, I will stop here.
it's a person's efficiency converted to an equivalent 2 point field goal percentage, to put it on the same scale as what we are historically used to. your formula considers 2 pointers and 3 pointers to be equal. a person who only shot 2's at 40% would have the same efficiency as someone who only only shot 3's at 40%, which makes the formula equivalent to FG%. and as best as i can tell, you're formula accounts for FTA the same way TS% does, except if the multiplier was 0.50 instead of 0.44, and i believe using 0.50 is actually one version of TS%, although as we can see from harden's stats this season the 0.44 number seems to be more accurate.
Except ts% does have something to do with shooting percentage. The idea is to adjust fg% to account for 3s and freethrows. Pretty simple concept. The issue is that you don't know the number of posessions used on the ft attempts so they estimate it. Is it completely accurate? No but over the course of a season it is close. As I asked earlier, do you know of any real examples where you feel ts% isn't accurate?
In all cases where you don't shoot any FGs, but make all your FTs. You can actually exceed 100%. maxing out at 113.6%. :grin: What do I win?
So, in real world stats, the way this plays out is when you actually shoot 2 free throws in a possession, your FTs count significantly more than they should -- peaking at a possible 113.6% TS%. Thus, guys like Dwight Howard and DJ who shoot a statistically significant amount of intentional foul shots (hack-a-shaq), actually have a higher TS% than they should. Should Hack-a-Shaqs (you don't have the ball) even count in a stat like TS%? If so, their multiplier is, of course, .5 not .44. The key is shooting foul And-1s drives that .44 (plus technical fouls). But for Hack-a-Shaq players, the .5 is the main driver. TSA - True Shooting Attempts (which really means Estimated Shooting Attempts) TSA = FGA + 0.44 * FTA TSA - True Shooting Attempts for Hack-a-Shaq Players TSA = FGA + (0.5 * Hack-a-Shaq-FTA) + (0.44 * Remaining-FTA) When using that formula, and assuming 25% of Dwights FTA were Hack-a-Shaq fouls, Dwight's TS% falls from .603 to .598 = Harden's TS%. He probably shot more than 25% intentional fouls, certainly DeAndre Jordan did. THAT IS A STATISTICALLY BIG DIFFERENCE. Both Dwight and Jordan drop out of the Top 10 players in TS%. bottomline: Hack a Shaq players are significant outliers to the .44 multiplier, as they are the only players who are fouled intentionally when they don't even have the ball. They shouldn't be given that .44 multiplier on Hack-a-Shaq foul shots. So, Harden's TS% is actually better than Dwight and DJ if the stat would actually do math correctly for FTAs awarded on Hack-a-Shaq fouls. .
deandre had 37 and1's and 619 fta. i'll assume he wasn't fouled on any 3's and didn't shoot any technicals. he probably got flagrantly fouled but i'll ignore that. given that, then he used 291 possessions for his 619 fta, making his multiplier 0.47. that would account for about 2.3% more possessions for jordan and that would actually knock about 1.4% off of his TS% and put it at 61.4%. the same calculations for dwight actually end up with him with a multiplier of 0.45, so not far off. his and1's, predictably high, seem to cancel out the hack-a-dwight.
Cool...When we actually know And-1s, then we can show how much the added free throws due to Hack-a-Shaq are inflating their TS%. When you know the true And-1s, you actually don't need to back-into the multiplier, we know it -- it is .5. True Shot Attempts becomes: TSA = FGA + (.5 * (FTA - And1s)) <a href="http://www.nbaminer.com/four-point-plays-and-one/">Here's the And1s for each player</a> (add Harden's 4pt plays to And1, since they count the same). And only Harden shot Techs, so those get subtracted from FTA like And1s do. I think I saw where Harden shot 56 techs, but can't find it now. Howard drops from .604 to .600 to drop out of Top 10 and Harden moves up from #11 to #9, with a .601. Jordan has a massive drop from .628 to .614 btw: While analyzing the players And1s and formulas, it is also clear the .44 overly awards most all players, because their true multipliers are much closer to .5 (ie., the closer your And1s are to zero, the closer your multiplier is to .5 -- disregarding technicals). It's like the aggressive scorers of the NBA are skewing the multiplier in favor of the other players who can never get to the line for And1s
