PDA

View Full Version : Steph Curry: Points per shot



JasonJohnHorn
02-10-2016, 11:12 PM
I realize points-per-shot is a greatly flawed stat, and certainly can't be taken into consideration without context.

But Curry is curry scoring about 1.5 points per shot. As far as I know, that is the highest the league has ever seen.

Jordan's was pretty impressive. As a rookie Jordan his 1.42 points per shot. On his career he had 1.31.


Right now, though, Curry is getting 1.51. There are some low-volume spot-up shooters, like Korver, how have hit 1.5 a game. But has a high-volume shooter ever had such an astronomically high number?


It is all the more impressive given that he doesn't draw as many fouls a game (Jordan would draw 5 or 6 a game; Curry gets 3 or 4).

HandsOnTheWheel
02-10-2016, 11:39 PM
:yawn:

tredigs
02-10-2016, 11:45 PM
Why is it a greatly flawed stat?

aman_13
02-11-2016, 12:09 AM
It's not surprising, he shoots a lot of threes and makes a lot.

nastynice
02-11-2016, 02:58 AM
It is all the more impressive given that he doesn't draw as many fouls a game (Jordan would draw 5 or 6 a game; Curry gets 3 or 4).

Yea, definitely. Because if there's one flaw with points per shot, its the idea (I'm guessing, I'm not sure exactly how its calculated) that if someone hits 2 free throws, they get 2pts/0 shots, even tho the reality is that the free throws DID come at the price of a possession, but its not accounted for. Curry barely gets to the free throw line.

Usually when you see someone close to double points than shots, you'll notice they had a **** ton of free throws, meaning the shot count doesn't account for how many possessions they actually used to get those points. But with curry, its like 5-6 free throws, if that, and he still damn near doubles his shot total in points.

For a stat to inherently put a certain player at a disadvantage, yet that player still putting up historic numbers regarding that stat, damn this boy lit!!!

Gander13SM
02-11-2016, 03:48 AM
Why is it a greatly flawed stat?

I'm guessing he feels that way because it doesn't take into account possessions used by free throws? I mean the points per shot formula is pretty simple, it's just Points/FGA.

Points per possession would be the better stat to use, no?

tredigs
02-11-2016, 11:22 AM
^No it'll include free throws also. Just globs them all together and basically says, "if you're using an offensive possession to attempt to shoot the basketball, this is how many points you are averaging". So, getting the ball blocked or whether it's a 2 or 3 or FT is just all combined.

As long as you're aware of a players general role/volume, then PPS is actually probably the most pure way of ranking scoring ability. I'm thinking points per possession is more useful as a team stat, not with PPS for individuals (because if you're not using the offensive possession to shoot at all, your # would go down with PPP, but it's unchanged with PPS). I think I've actually been saying "PPP" in a couple threads earlier this week where I'm meaning to say PPS.

houstonfan
02-11-2016, 11:49 AM
Hardens is 1.47! Gotta love those free throws.

Jeffy25
02-11-2016, 11:49 AM
I don't see the flaw in the stat. It tells you exactly what it is.

How you use it can be flawed of course.

FraziersKnicks
02-11-2016, 11:59 AM
Player A shoots 0-5 but gets fouled 5 times and makes all 10 free throws. 10 points, 10 possessions. A PPS of 2.

Player B shoots 5-5 but no free throws. 10 points, 5 possessions. A PPS of 2.

Both score 10 points but player B uses 5 possessions whilst player A uses 10. However both have the same PPS score, when it's clear player B is a more efficient option for his team.

Obviously that's an extreme example but that's where I see the fault.

valade16
02-11-2016, 12:03 PM
Let's not forget that 3-pointers greatly help a person's ability to raise their PPS.

IndyRealist
02-11-2016, 12:04 PM
Points per shot is flawed because it doesn't take into account possessions used on free throws but still awards the points, so in essence it overly rewards people who go to the line.

However it is very easy to calculate and doesn't rely on guesstimation of posessions or physically counting from video. You just have to understand what you're looking at, like all stats.

IndyRealist
02-11-2016, 12:06 PM
Let's not forget that 3-pointers greatly help a person's ability to raise their PPS.

Compensated for the fact that 3s are a lower percentage shot, so you're getting the value of the 3 assuming more misses. eFG does the same thing.

valade16
02-11-2016, 12:46 PM
Compensated for the fact that 3s are a lower percentage shot, so you're getting the value of the 3 assuming more misses. eFG does the same thing.

Not that much lower of a shot that it does not help a person's PPS. Any % above 33% is adding more PPS than anyone shooting 50% from 2. Considering basically nearly every guard can shoot 33% from 3 it helps.

It's not even a knock on Curry or trying to belittle his PPS, because the 3 helps his PPS only because he's so mind blowingly good at it.

But if we looked at say Jerry West or Rick Barry, their PPS would almost certainly be higher if they had played with a 3 point line.

Vee-Rex
02-11-2016, 12:58 PM
Player A shoots 0-5 but gets fouled 5 times and makes all 10 free throws. 10 points, 10 possessions. A PPS of 2.

Player B shoots 5-5 but no free throws. 10 points, 5 possessions. A PPS of 2.

Both score 10 points but player B uses 5 possessions whilst player A uses 10. However both have the same PPS score, when it's clear player B is a more efficient option for his team.

Obviously that's an extreme example but that's where I see the fault.

This.

And that's not counting And-Ones, technical free throws, and 3-point plays (3 free throws from a foul at the 3-point line). If you look at PPS from a very basic view, it looks great, but when you add in all these variations, some flaws pop up.

Another flaw - you cannot use PPS to measure a player's offensive impact because it only measures the points per shot, not how an offense may suffer from having a specific player on the court.

Deandre Jordan leads the entire league in PPS at 1.95, yet no one sane will regard him as a top offensive player in the league. He's efficient because of alley oops and putbacks. PPS obviously is a good way (not great, because of the above-mentioned flaws) to measure a guy's points per shot (hence the name), but trying to use it to determine the better offensive player is when this flaw pops up.

tredigs
02-11-2016, 01:24 PM
Let's not forget that 3-pointers greatly help a person's ability to raise their PPS.

If they're good at them. For a Lebron or Kobe or Jordan who are career 33% from three shooters, it's negligible. But yeah, outside of layups/dunks, as a general rule it's going to be more advantageous for a solid+ 3pt shooter to always seek that shot on offense over a mid-range shot. If you're actually attacking the rim and seeking contact, that's a different story, because for most a FT is going to be their highest % shot.

For Curry, even though he's a very solid 50% career from 2 on ~9 attempts a game (up to 57% now. Both better averages/peak than Kobe from 2, and the peak is also a better 2pt% than MJ's best), he's up over 45% on 11 attempts a game from THREE (the first time he's averaged more 3's than 2's. It's the equivalent to being a 68% shooter from 2). Helps that he's also the best foul shooter ever along with Nash and Price, so you can't afford to send him to the line. Anyway, all that combined is what's making for one of the top scoring seasons ever.

Agreed with you that this PPS stat (same with TS%) isn't entirely fair to compare to pre 1980 elite scorers who could have also hit 3's. But, you can compare it to any of the bigs from those days (Wilt/Kareem, etc) who would not have taken 3's, and all players after '80 (which brings in most of rest of the top scorers ever: Bird/Nique/Jordan/Malone/Shaq/Kobe/Dirk/T-Mac/Lebron/Durant etc).

IndyRealist
02-11-2016, 01:25 PM
Not that much lower of a shot that it does not help a person's PPS. Any % above 33% is adding more PPS than anyone shooting 50% from 2. Considering basically nearly every guard can shoot 33% from 3 it helps.

It's not even a knock on Curry or trying to belittle his PPS, because the 3 helps his PPS only because he's so mind blowingly good at it.

But if we looked at say Jerry West or Rick Barry, their PPS would almost certainly be higher if they had played with a 3 point line.

That is because 3s are inherently more valuable than 2s unless you're at the rim. That is why teams are either shooting 3s or getting to the rim, those are just the better shots. Welcome to the world of analytics.

SLY WILLIAMS
02-11-2016, 01:45 PM
Points Per Shot has never been a perfect stat but it is still a good stat.

A guy (Gallinari) the main board often ignores is 5th in the league right behind Durant at 1.48 PPS

http://espn.go.com/nba/statistics/player/_/stat/field-goals/sort/scoringEfficiency

tredigs
02-11-2016, 01:47 PM
Player A shoots 0-5 but gets fouled 5 times and makes all 10 free throws. 10 points, 10 possessions. A PPS of 2.

Player B shoots 5-5 but no free throws. 10 points, 5 possessions. A PPS of 2.

Both score 10 points but player B uses 5 possessions whilst player A uses 10. However both have the same PPS score, when it's clear player B is a more efficient option for his team.

Obviously that's an extreme example but that's where I see the fault.

No, I don't see how that's not a 2 PPS in your 2nd example. It's a 1 PPS. 10 shots. 10 points. 1 PPS.




This.

And that's not counting And-Ones, technical free throws, and 3-point plays (3 free throws from a foul at the 3-point line). If you look at PPS from a very basic view, it looks great, but when you add in all these variations, some flaws pop up.

Another flaw - you cannot use PPS to measure a player's offensive impact because it only measures the points per shot, not how an offense may suffer from having a specific player on the court.

Deandre Jordan leads the entire league in PPS at 1.95, yet no one sane will regard him as a top offensive player in the league. He's efficient because of alley oops and putbacks. PPS obviously is a good way (not great, because of the above-mentioned flaws) to measure a guy's points per shot (hence the name), but trying to use it to determine the better offensive player is when this flaw pops up.

It's a scoring efficiency metric, but yes obviously you still need to apply context. You wouldn't put a Deandre Jordan or Dwight Howard who score less than 15 points a game in the same category as the leagues leading scorers. It also has nothing to do with overall offensive ability because it's not telling you how much attention that player is drawing from the defense, or anything about his playing ability. Basically, it's a good stat to compare the scoring ability of similar volume scorers from 1980 on (or bigs before then).

JasonJohnHorn
02-11-2016, 01:57 PM
I don't see the flaw in the stat. It tells you exactly what it is.

How you use it can be flawed of course.

I used to feel the same way... but....


Points per possessions is better.


The problem is this.

Player one takes 18 shots and gets 30 points. He also has 5 FTs a game, so lets say one is a plus-1 and two are shots that didn't go down, which means that hesed 20 possessions to shoot himself.

1.5 points per possession.
1.6 per shots.


Player 2 score 33 points on 20 shots, but he shoots 10 FTs a game. So let's say 4 additional possessions. 20 shots, 24 possessions.

His points-per-shot will look AS good, but his percentages are lower and he needs more possessions. So he's really getting 1.3 points per possession, which is far less impressive than 1.5, though his points-per-shot is the same.


It also doesn't take into account turnovers and assists.


It's a good indicator, but it needs context.

tredigs
02-11-2016, 02:14 PM
Lol I need to take a nap and revisit this. I'm too hungover for this right now. Maybe I'm thinking about something wrong.

Chronz
02-11-2016, 02:15 PM
^No it'll include free throws also. Just globs them all together and basically says, "if you're using an offensive possession to attempt to shoot the basketball, this is how many points you are averaging". So, getting the ball blocked or whether it's a 2 or 3 or FT is just all combined.

As long as you're aware of a players general role/volume, then PPS is actually probably the most pure way of ranking scoring ability. I'm thinking points per possession is more useful as a team stat, not with PPS for individuals (because if you're not using the offensive possession to shoot at all, your # would go down with PPP, but it's unchanged with PPS). I think I've actually been saying "PPP" in a couple threads earlier this week where I'm meaning to say PPS.

It doesn't include ft/a tho. Thats the reason why its the worst of all the barometers for efficiency, it only pretends to measure efficiency. I've been telling JJH this for years but he insists on ignoring the best tools available, just as he insists offensive rebounding is a part of defense even tho coaches have all but abandoned chasing them to actually improve defensively. Weird old school logic.

Hawkeye15
02-11-2016, 02:19 PM
PPS is a lesser attempt at TS% really

Hawkeye15
02-11-2016, 02:19 PM
It doesn't include ft/a tho. Thats the reason why its the worst of all the barometers for efficiency, it only pretends to measure efficiency. I've been telling JJH this for years but he insists on ignoring the best tools available, just as he insists offensive rebounding is a part of defense even tho coaches have all but abandoned chasing them to actually improve defensively. Weird old school logic.

basically, yep

aman_13
02-11-2016, 02:32 PM
Yeah it's much better to use TS%.

Vee-Rex
02-11-2016, 02:38 PM
Yeah it's much better to use TS%.

TS% has its flaws too, though. It weighs FTs more heavily than it should at 0.44.

IndyRealist
02-11-2016, 02:43 PM
TS% has its flaws too, though. It weighs FTs more heavily than it should at 0.44.
But why is 0.44 wrong? If you bring it up you gotta explain your reasoning.

Chronz
02-11-2016, 04:01 PM
But why is 0.44 wrong? If you bring it up you gotta explain your reasoning.

My guess is he wants to abandon the estimate in favor of the accurate stats we have now but things don't change that much, really only effects go to technical guys

aman_13
02-11-2016, 04:09 PM
TS% has its flaws too, though. It weighs FTs more heavily than it should at 0.44.

I cant explain how they came up with that number. It's an estimate based on research and it accounts for the fact that not all free throws equal a possession. For example: and 1 plays and technical free throws.

Scoots
02-12-2016, 11:25 AM
.44 definitely seems high to me.

LOb0
02-12-2016, 11:29 AM
Points per shot is flawed because it doesn't take into account possessions used on free throws but still awards the points, so in essence it overly rewards people who go to the line.

However it is very easy to calculate and doesn't rely on guesstimation of posessions or physically counting from video. You just have to understand what you're looking at, like all stats.

I don't really see how that makes it flawed. Its just rewarding free throws. I suppose if it doesn't count missed FTs that's a flaw.

mike44
02-12-2016, 01:40 PM
Two years ago durant was at about 1.54 while scoring 32 points per game. That might be the highest for a high volume scorer.

jerellh528
02-12-2016, 02:22 PM
Two years ago durant was at about 1.54 while scoring 32 points per game. That might be the highest for a high volume scorer.

KD is having another great year under the radar too.
But yeah that year was KD was wild. It would be pretty hard for me to choose between that season from durant and this one from curry. I think I'd lean toward KD for being the more constantly dominating scorer that season, where this season I feel like curry has had some games where he seems mortal a little bit.

tredigs
02-12-2016, 03:06 PM
KD is having another great year under the radar too.
But yeah that year was KD was wild. It would be pretty hard for me to choose between that season from durant and this one from curry. I think I'd lean toward KD for being the more constantly dominating scorer that season, where this season I feel like curry has had some games where he seems mortal a little bit.

Pretty close to be honest. That year was insane by Durant. Pretty sure that's when Westbrook went down for a bit and he went full bore for over a month and it was a sight. I think I'd have to give Curry the edge for both having better per-36 volume (he's had to sit 15 full 4th quarters, I think we have to use per-36 to make a legit comparison), and the better efficiency (64% TS to 68% TS). Consistency wise, I'd agree it was probably KD. With Curry having the higher peaks. I think that's the nature of one guy who when it's not going for him, is still going to get to the line 13-15 times and keep his scoring level. I'm looking at KD's first game of the season that year, and he had a lackluster 9-24 with 2-8 on three. But, he put up 42 points... (22-24 from the line). That's something Curry's never going to be able to do.

And yeah, KD's being a little overshadowed by Curry this year, and Westbrook on his own team, but he's doing his thing at an extremely high level himself.

IndyRealist
02-12-2016, 09:12 PM
I don't really see how that makes it flawed. Its just rewarding free throws. I suppose if it doesn't count missed FTs that's a flaw.

Because it doesn't take into account the cost of taking the free throws (using up a possession), but does count the benefit (the points scored). Essentially it says that no matter how badly you shoot free throws it's always good to go to the line. If that were the case we wouldn't be having threads dedicated to discussing hack-a-shaq. PPS doesn't come close to representing reality.

nastynice
02-13-2016, 01:17 AM
I don't really see how that makes it flawed. Its just rewarding free throws. I suppose if it doesn't count missed FTs that's a flaw.

but it rewards free throws to a much higher degree. If a player gets fouled 5 times and makes 10/10 free throws, they get 10 points at the expense of ZERO shots. When the reality is, they used 5 possessions for those 10 points. Where as if a player goes 5/5 from the field, they get 10 points at the expense of 5 shots. Both players used the same amount of possessions to get the same amount of points, but the "points per shot" will be heavily swayed in the favor of one player over the other.

Think about it, lets say a player goes 1/1 from the field, and 8/8 from the free throw line, they will have a point per shot of TEN points PER shot. That doesn't even make sense, there's no such thing as a 10 pointer. The most points a player can get per shot is 4, if they shoot and make a three while being fouled, and also make the free throw. That is a legit 4 points per possession used.

**another example could be let's say deandre jordan is getting hacked, and he's hitting free throws at a rate of 1/4. That means every 2 offensive trips, he's getting only one point. But point per shot doesn't take that into account, instead it rewards him as tho he's just getting and 1's, rather than shooting horribly from the line and scoring very low amount of points per offensive possession used. Its extremely inefficient offense, that's why teams use this strategy, it allows them to climb back into games.

It doesn't make this a bad or ****** stat, but it is misleading in a way. But ALL stats can be misleading in the wrong context. Just pointing out the particular flaw of this particular stat.

In a nutshell, I guess what I'm saying is, points per possession used would be a much more accurate stat in order to gauge and compare various player's offensive efficiency. That could take into account missed free throws and turnovers.

tredigs
02-13-2016, 02:30 AM
It actually definitely does make it a bad stat. It requires WAY too much context. Originally I had no idea that FT's only counted for you, not against you what so ever. It explains Deandre's 1.95 point per shot. I was trying to figure out how he got nearly a 2 point per shot, when obviously he never takes a 3 and is terrible from the line. But, being that he takes so many FT's, it gets up to nearly 2.

Overall, bad stat. Should never be used.

IndyRealist
02-13-2016, 12:17 PM
It actually definitely does make it a bad stat. It requires WAY too much context. Originally I had no idea that FT's only counted for you, not against you what so ever. It explains Deandre's 1.95 point per shot. I was trying to figure out how he got nearly a 2 point per shot, when obviously he never takes a 3 and is terrible from the line. But, being that he takes so many FT's, it gets up to nearly 2.

Overall, bad stat. Should never be used.

I see it as more of a quick and dirty calculation, like Win Score, Game Score, and TENDEX. They all give up accuracy for ease of calculation.

Scoots
02-13-2016, 09:02 PM
I see it as more of a quick and dirty calculation, like Win Score, Game Score, and TENDEX. They all give up accuracy for ease of calculation.

But why not include 50% of FT attempts as shots at least?

IndyRealist
02-13-2016, 09:31 PM
But why not include 50% of FT attempts as shots at least?

I didn't design it, but if I had to guess I'd say it's so PPS is easily understood. Points per shot plus free throws is kind of ambiguous.

Scoots
02-13-2016, 09:37 PM
I didn't design it, but if I had to guess I'd say it's so PPS is easily understood. Points per shot plus free throws is kind of ambiguous.
I know you didn't design it ... Or at least i assumed. :)

PPP isn't perfect but it's far better. PPS is dead to me.

IndyRealist
02-13-2016, 10:18 PM
I know you didn't design it ... Or at least i assumed. :)

PPP isn't perfect but it's far better. PPS is dead to me.

In an age where the results of complicated calculations are at your fingertips with virtually no work on your part, there is really no reason to use something like PPS over PPP. PPS is pretty obscure anyway, the only reason people use it is because it skews toward their favorite player.

Scoots
02-13-2016, 11:07 PM
In an age where the results of complicated calculations are at your fingertips with virtually no work on your part, there is really no reason to use something like PPS over PPP. PPS is pretty obscure anyway, the only reason people use it is because it skews toward their favorite player.

I'm waiting for all the SportVu based advanced stats. "Dribbles/ball time per shot/assist" for PGs will be interesting. Total distance traveled per game/player. I can't even imagine what all is coming.