Should Ellis Do More or Less?
by Evanz on Sep 8, 2010 12:04 PM PDT
Email
Print
This seems to be one of the big questions heading into next season. Does Ellis hurt the team by taking more shots? As Curry gained influence and control of the team, did Ellis back off? Did his efficiency go up? What about useage? I was really curious to look into the stats a bit deeper and see what I could see. So, I went to bball-reference.com and started crunching a few numbers. Specifically, I looked at three measures (TS vs. USG), (TS vs. Game #), and (USG vs. Game #). Game # is exactly what it sounds like (i.e. Game 1 is the first game of the season, 82 is the last). By plotting against game #, we can see temporal correlations. Here are the formulas, for reference:
(TS) True Shooting Percentage; the formula is PTS / (2 * (FGA + 0.44 * FTA)). True shooting percentage is a measure of shooting efficiency that takes into account field goals, 3-point field goals, and free throws.
(USG) Usage Percentage (available since the 1977-78 season in the NBA); the formula is 100 * ((FGA + 0.44 *FTA + TOV) * (Tm MP / 5)) / (MP * (Tm FGA + 0.44 * Tm FTA + Tm TOV)). Usage percentage is an estimate of the percentage of team plays used by a player while he was on the floor.
Here are the data (for 2009-10 obviously):
Ellis
Curry
There are many interesting trends (and non-trends). For example, there appears to be a greater trend (positive correlation) for Ellis' TS vs. USG compared to Curry. Therefore, Curry's efficiency tended to be more independent of his USG. Ellis tended to shoot better when his USG increased. This actually tells me that it is not necessarily the case that Ellis will shoot better (i.e more efficiently) if his role in the offense is diminished. Indeed, according to the data, he may shoot worse.
Some on this site have said that Ellis "backed off" (my paraphrasing) his shot-taking as Curry's influence began to grow. Well, let's take a look at the data. We can see by Curry's USG vs. Game # plot, that Curry's USG clearly increased as the season progressed. So, the question to ask (in my mind): Did Ellis' USG go down during the season? The answer is clearly NO. There doesn't appear to be any trend in Ellis' USG vs. Game #. What about Ellis' TS vs. Game #? Again, the answer is NO, there was no relation.
Some questions to think about... Why is Ellis' TS correlated with USG, whereas Curry's does not appear to be? Is it that Ellis adjusts his role in the offense when he's "feelin' it"? Does he take fewer shots when he's having an off-night? Or, perhaps, is it that Curry is getting Ellis the ball more when he sees that Ellis is shooting well? There are a lot of factors here obviously. What I've presented is only a starting point. But I think that by looking into these kinds of per-game data, it may be possible to tease out some answers, or at the very least, ask some interesting questions. Feel free to suggest some other comparisons, or provide your own below!
UPDATE: I performed a regression analysis using the function "regress()" in MATLAB (click on the link for gory details). The results are shown below. It looks like the numbers I previously put for R^2 values were actually the coefficients of the fit. It turns out the p-value for Ellis' TS vs. USG (p=0.061) is just above what would normally be considered statistically significant (p<0.05). The p-value for Curry is well above at ~0.8. Therefore, while the data for Ellis do not show statistical significance, I do think it is reasonable to call the correlation a "trend", albeit arguably weak.
This FanPost is a submission from a member of the mighty Golden State of Mind community. While we're all here to throw up that W, these words do not necessarily reflect the views of the GSoM Crew. Still, chances are the preceding post is Unstoppable Baby!
100 comments
|
5 recs |
Do you like this story?
Comments
too…many…stats.
rec’d.
http://en.wikipedia.org/wiki/Tim_Roye
by Elevation Sensation on Sep 8, 2010 12:55 PM PDT reply actions
Shoot better or worse vs whatever....
I think a lot with Ellis is mental, the old adage “let the game come to you” I think is a good cliché when talking about Ellis and when he is on his game or off.
"If God made us in his image then he must be dumb too, and a little ugly on the side."
Frank Zappa
i think
near the end of the season, ellis kind of let curry be the major ball-handler and i think ellis realized that it will benefit him and the team. i think that is why near the end of the season, the chemistry seemed to be stronger with the team understanding and allowing each other to grow.
and with the change in his attitude, im excited for the coming season.
Ellis tended to shoot better when his USG increased. This actually tells me that it is not necessarily the case that Ellis will shoot better (i.e more efficiently) if his role in the offense is diminished. Indeed, according to the data, he may shoot worse.
Ellis forced up a ton of really bad shots throughout the season. When he takes fewer shots overall, the smaller sample size leads to some abysmal shooting numbers because of his poor shot selection. However, increase his usage and the percentages tend to tread upwards slightly thanks to the extra attempts and the opportunities to make shot adjustments as the game wears on. Considering the absurdly large volume of bad shots being taken, and taking into account that Ellis is a talented, professional baller, it’s no wonder that some of them are bound to go in regardless of the lower probability.
In a way, it’s similar to the claim that Kobe is “clutch” when crunch-time shooting percentages show otherwise. I think he made between 6 or 8 game-winning buzzer beaters last season, tops in all of the NBA. But why does Kobe have this distinction? Because he takes more attempts than any other player in the league.
Ultimately, I personally think the stats you provided are a little misleading. The issue really isn’t whether or not we should increase/decrease Monta’s usage, but it’s his shot selection that needs work. You do have a point though; all other factors left the same (same poor shot-selection, inability to hit the 3 consistently, etc.) Ellis could likely shoot even worse if his usage was decreased. Improving the types of shots he takes, however, will make him more effective regardless of how often he touches the ball.
by WYK on Sep 8, 2010 1:17 PM PDT reply actions 1 recs
Oh, and rec’d for posting up a well-written and well-researched topic.
by WYK on Sep 8, 2010 1:21 PM PDT up reply actions
When he takes fewer shots overall, the smaller sample size leads to some abysmal shooting numbers because of his poor shot selection. However, increase his usage and the percentages tend to tread upwards slightly thanks to the extra attempts and the opportunities to make shot adjustments as the game wears on.
I think you hit the nail on the head here. Rec’age! Also, with regard to his shot selection last season, I don’t think one can stress enough the total disarray that he was playing in. If I were a player that hung my hat on scoring in 1-on-1 situations, and half to most of my team was injured on a nightly basis, I would probably take some bad shots too. I’ve certainly played in my fair share of pick-up games where I felt I was, unfortunately, the best scoring option. If I were Monta and saw a Tolliver or Hunter with his hands up, I might not pass either.
Barring continued improbable injuries or a major trade, I am convinced Monta’s shot selection will be much improved this season. Convinced!
Amnesty for all immigrants! Boycott Arizona!
by The Bimbo Coles Experience on Sep 8, 2010 5:13 PM PDT up reply actions
If I were a player that hung my hat on scoring in 1-on-1 situations, and half to most of my team was injured on a nightly basis, I would probably take some bad shots too.
Even when two of the guys being left open were two of the best 3-pt shooters in the league?
What I was eluding to is the effect of frustration on his decision-making. It’s easy to forget how good of a spot-up shooter a guy is when you’re down big, and considered the go-to guy. Ultimately, not dishing to an open Curry or Morrow on the arc is indefensible, but given the piss-poor conditions he was playing in, I’m willing to give him a pass, if you’ll pardon the pun, on this one. In other words, it will be much easier to make the right decision this season with a consistent lineup featuring proven shooters in Curry, Williams, and Lee in the high post.
Amnesty for all immigrants! Boycott Arizona!
by The Bimbo Coles Experience on Sep 8, 2010 7:03 PM PDT up reply actions
I'll give you Lee
But he had “proven shooters” Curry and Williams last year, along with the unproven non-shooters Morrow (.460 career 3%) and Watson (.351 career 3%) last year and he didn’t pass the ball.
Monta Ellis's #1 Fan!!!
My point is that he is not going to get as many touches. He will be forced to either become more efficient or prove himself ineffective.
Amnesty for all immigrants! Boycott Arizona!
by The Bimbo Coles Experience on Sep 8, 2010 11:15 PM PDT up reply actions
I honestly didn’t see Ellis “backing off” at all last season. If he did, it was probably toward the end, where he had several DNP’s, so I didn’t notice. With Curry, no one expected him to get better as the season progressed, but he did and we’re all happy about that.
I think it would have been hard for Ellis to adapt to Curry in the matter of one season, let alone the 2nd half of the season (when Curry’s production really took off) so I don’t think we can really draw too many conclusions from the above data.
It would be interesting to see this data side-by-side with the upcoming season to see if any of the trends continue the way they’re suggesting. I would think that Ellis’ USG and TS would go up even more since his role will hopefully be more defined and simpler than last year.
"the world is not your trade machine" - wallywagon11
He'll do more by doing less.
As in less ball handling leads to more off-ball play which hopefully leads to more efficiency.
"Monta is the MAN." -Bob Fitzgerald
How come people never take into account that last year was an injury plagued season for the Warriors and that we only had 7 healthy guys at the most? Ellis was able to create his own shot, which is needed when a team is exhausted during the course of a game.
In my humble opinion, people are being way too harsh on the way Monta performed last year. Sure his shooting percentages were down and there were many ill-advised heaves, but at the end of the day, someone needed to take those shots. Let’s remember Monta averaged over 40 minutes a game last year. I play ball often and after running up and down the court for that long…I’m pretty sure my shots would start to fall short too. Fatigue matters.
by LetsGoDubs on Sep 8, 2010 1:31 PM PDT reply actions 3 recs
Also, wanted to add that teams were able to focus their attention on Monta last year because quite frankly, he’s all we had for the most part until Curry started to perform in the last 2 months of the season. I believe the addition of David Lee will help Monta the most because he won’t be the only one opposing teams need to concentrate on.
by LetsGoDubs on Sep 8, 2010 1:34 PM PDT up reply actions 2 recs
Sure his shooting percentages were down and there were many ill-advised heaves, but at the end of the day, someone needed to take those shots.
Yes, someone needed to take those shots. That someone should not have been Monta. Even if he did “need” to take them he shouldn’t have wasted them chucking 3’s or forcing a shot on a 3-on-1 isolation play.
Also, wanted to add that teams were able to focus their attention on Monta last year because quite frankly, he’s all we had for the most part until Curry started to perform in the last 2 months of the season.
Did you forget that we had Maggette? How about Morrow? Or Reggie? Or even CJ at times. There were a ton of better scoring options on the floor, but Monta only looked for his own shot. Also, Curry only struggled during the first 2 months of the season. By January he had already picked up his game, so there’s another option.
by WYK on Sep 8, 2010 1:49 PM PDT up reply actions
That someone should not have been Monta
Than who would?
Did you forget that we had Maggette? How about Morrow? Or Reggie? Or even CJ at times
Maggs cannot be the focus of an offense for more than 2 minutes without making ball movement go extinct, CJ is not the best at creating for himself , Morrow couldn’t create for himself to save his life and Reggie was only here at the very end of the season.
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
Than who would?
Those other guys he said.
I must look like a dork.
by Reverend_Randy on Sep 8, 2010 2:02 PM PDT up reply actions
If your man isn't working to get open/isn't open, I don't expect them to get the ball
Morrow nor CJ move well without the ball and Maggs is an iso wrecking ball. I don’t know about you but I don’t like giving the ball to people if there not open, it would likely end up a turnover.
Going 2 on 1 / 3 on 1 and maybe making the shot > throwing a turnover.
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
Maggs cannot be the focus of an offense for more than 2 minutes without making ball movement go extinct
Oh, and Monta hogging the ball to himself trying to play “hero” doesn’t make ball movement go extinct either? Nice double-standard you got going there.
CJ is not the best at creating for himself , Morrow couldn’t create for himself to save his life
Right, because the ability to create shots for oneself is the ultimate indicator of whether or not a person should be allowed to shoot even while playing off-ball. Even if Morrow and CJ can’t exactly create for themselves, you honestly believe Monta going 3-on-1 is a better option than passing to an open Morrow or CJ?
Reggie was only here at the very end of the season.
And Monta was still jacking up stupid shots when yet another option was added towards the end of the season.
by WYK on Sep 8, 2010 2:26 PM PDT up reply actions
Oh, and Monta hogging the ball to himself trying to play "hero" doesn’t make ball movement go extinct either?
No cause Monta can actually find people unlike Maggs
Right, because the ability to create shots for oneself is the ultimate indicator of whether or not a person should be allowed to shoot even while playing off-ball
If your not moving and trying to get open off the ball how will you get the ball? And if you struggle on the ball and don’t move off the ball chances are you won’t be taking to many shots. Think about it, if you were 1 on 1 with someone and you had 2 bigs in the paint waiting and you looked to your sides and your own guys were just watching you while covered would you throw the turnover or take the shot? If you don’t get open you don’t get the rock.
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
No cause Monta can actually find people unlike Maggs
If he can, he seems to have no intention of displaying that skill. It’s either that, or he doesn’t have that ability. It appears to me that you’re simply looking at the “5.3 assists per game” value and trying to misconstrue it into being an indication that he’s a “good passer” or that he “sees the floor well.” He isn’t and he doesn’t.
If your [sic] not moving and trying to get open off the ball how will you get the ball?
Morrow or CJ’s guys sag off their man to double or even triple team Monta, leaving one or both of them wide open and they didn’t even have to do anything but wait for the ball to come their way. Yet Monta almost always charged straight into the triple team to force a shot or lose the ball on a turnover.
And if you struggle on the ball and don’t move off the ball chances are you won’t be taking to many shots. Think about it, if you were 1 on 1 with someone and you had 2 bigs in the paint waiting and you looked to your sides and your own guys were just watching you while covered would you throw the turnover or take the shot? If you don’t get open you don’t get the rock.
You don’t seem to understand the concept of a double-team. The situation you presented here comes off more as an isolation play used to attack a zone defensive scheme, which is not what Ellis faced considering very few teams use the zone.
In a double-team, when more than one guy comes to guard you, it’s almost certain that there’s a wide open teammate with a better chance at making a shot than you do. So you, as the ball handler, must find the open man and pass to him. Monta seems to have a problem finding this open teammate, perhaps Monta might even refuse to find him under the assumption that he himself must be “the man.” Defenses thus collapse on him, leaving open teammates (remember: they don’t even have to move because they’re wide open already) who don’t get the ball because Monta is too busy trying to pull off half-a-dozen spin moves to open up space and force a shot.
by WYK on Sep 8, 2010 8:45 PM PDT up reply actions
Monta was hardly ever double teamed up top
He was doubled/tripled in the paint when he drove. Morrow was always guarded and never moved, same with CJ. Also I never said Monta has great court vision or is a great passer but just said that he can find people. Which he can and does, maybe not like a PG but he does find people.
When he got double/tripled in the paint he was usually either on his way up or trying to get contact. He doesn’t have the court vision to in mid air throw to the 3 point line and find someone. That’s what Steph does. Defenses collapsed on him on the inside, it’s very hard to to pass once going up in the paint with 2 – 3 guys around you. Only really good PG’s can do that well, he’s not one.
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
Monta was hardly ever double teamed up top He was doubled/tripled in the paint when he drove. Morrow was always guarded and never moved, same with CJ.
The list of “open teammates” isn’t limited to just CJ and Morrow. Anybody who’s left open will have an infinitely better chance at making a bucket than Monta does in a double/triple-team situation. Even with Chris Hunter’s stone hands, Monta should have passed to him if he was the guy who’s open.
Whether or not Hunter (or any other Warrior for that matter) finishes from the pass would be on him, not Monta, and we wouldn’t be here arguing about Ellis making boneheaded 3-on-1 plays. However that clearly isn’t the case, so here I am arguing that there’s no reason for him to be so horridly ineffective while you and LetsGoDubs up there are trying to justify the stupid things he does on the court.
Also I never said Monta has great court vision or is a great passer but just said that he can find people. Which he can and does, maybe not like a PG but he does find people.
Yeah, and Biedrins can make his free throws occasionally too. Only, Biedrins gets constant flak for hardly hitting his free throws. Why should we treat Monta any different when he hardly “finds people” to pass to when he definitely “can?”
When he got double/tripled in the paint he was usually either on his way up or trying to get contact. He doesn’t have the court vision to in mid air throw to the 3 point line and find someone.
More reason for him not to be charging into a crowded area like he usually does.
I’m going to reiterate again: Monta was not the only offensive option on that team. This wasn’t the ‘00-’01 defensive-minded Sixers squad where only Allen Iverson had any kind of offensive production on the roster. This was a team that scored 108 points a game. Claiming Ellis “had no help” or that he was the “only offensive option” isn’t justification for any of the dumb plays he pulled.
by WYK on Sep 9, 2010 12:43 AM PDT up reply actions 1 recs
this entire thread cracked me the heck up
There were a ton of better scoring options on the floor, but Monta only looked for his own shot … Did you forget that we had Maggette?
Maggs cannot be the focus of an offense for more than 2 minutes without making ball movement go extinct
Oh, and Monta hogging the ball to himself trying to play "hero" doesn’t make ball movement go extinct either? Nice double-standard you got going there.
No cause Monta can actually find people unlike Maggs
Seriously you can just repeat the cycle at this point
There were a ton of better scoring options on the floor, but Monta only looked for his own shot … Did you forget that we had Maggette?
Maggs cannot be the focus of an offense for more than 2 minutes without making ball movement go extinct
Oh, and Monta hogging the ball to himself trying to play "hero" doesn’t make ball movement go extinct either? Nice double-standard you got going there.
And it goes on and on and on and on …
Because talent? You want to add a guy like boozer because hes talented? I’m sorry i just don’t see that as a valid reason. -- fffindeed
by wallywagon11 on Sep 9, 2010 11:16 AM PDT up reply actions
Yeah, and Biedrins can make his free throws occasionally too.
The differences is Biedrins is one of the worst FT shooters in the NBA, I wouldn’t call Monta the worst passer in the NBA by any stretch.
I never once said he was the only option, never. I said people would just stand and watch, which basically gave him little options. And driving in and going up then all of a sudden passing to Hunter or Moore or whomever is there is much easier said than done. Again, he doesn’t have the vision for that and with no one trying to get open he has little options when he had the ball.
Whether or not Hunter (or any other Warrior for that matter) finishes from the pass would be on him,
Actually that did happen a lot, mainly from Hunter, Moore, Carl, Martin and TOlliver (when he was down low). Our options this year were Curry, Ellis and Maggs. Maggs is a black hole of epic proportions, and Curry didn’t come on until the 2nd half of the season. Morrow would disappear often and AR was injured . The lack of players due to injury really limited the options
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
Maggs is a black hole of epic proportions
That’s a reason to not expect to get the ball back from Maggette, not a reason to avoid finding him when he’s open. “Black hole” or not, (and “not” has a lot going for it—there are a large number of players who were far less likely to get assists than he was) when he set himself on a path to the basket, there were few more efficient in getting points out of the attempt than Maggs.
by jae on Sep 9, 2010 2:33 PM PDT up reply actions
Also, Curry only struggled during the first 2 months of the season.
Did he “struggle”? The data doesn’t show a trend in his TS vs. Game #.
http://www.basketball-reference.com/players/c/curryst01/splits/2010/
Scroll down a bit until you see the monthly splits. Curry had a pretty bad November and a decent December. It wasn’t until January when his production really took off. So yeah… compared to the rest of his season, the first two months was a bit of a “struggle” for him.
by WYK on Sep 8, 2010 2:13 PM PDT up reply actions
Just based on some per-minute production numbers and TS%:
November (per 36): 12.1 points, 3.9 rebs, 6.6 asts, 0.518 TS%
December (per 36): 14.3 points, 4.7 rebs, 3.9 asts, 0.546 TS%
January (per 36): 17.4 points, 4.0 rebs, 4.6 asts, 0.589 TS%
February (per 36): 19.6 points, 4.8 rebs, 6.7 asts, 0.566 TS%
March (per 36): 19.0 points, 4.1 rebs, 7.0 asts, 0.573 TS%
April (per 36): 24.0 points, 5.8 rebs, 7.3 asts, 0.590 TS%
Take what you want from it. Personally, I do think he “struggled” (to use the term as liberally, I guess) getting used to playing in a pro atmosphere during his first two months.
by WYK on Sep 8, 2010 2:55 PM PDT up reply actions
One month numbers can be deceiving...
February 2008 (Per 36): 23.3 points, 4.2 rebs, 4.3 asts, 0.634 TS%
Has he ever even come close to a line like that again?
Monta Ellis's #1 Fan!!!
Where'd you find that?
"Everybody loves Basketball-Reference.com. Except the Kobe fans".- DubsFan408
by GovernorStephCurry on Sep 8, 2010 8:16 PM PDT up reply actions
No, i meant the actual raw data.
"Everybody loves Basketball-Reference.com. Except the Kobe fans".- DubsFan408
by GovernorStephCurry on Sep 8, 2010 9:24 PM PDT up reply actions
Can you send a link of how to get his TS% and per36 numbers per month. I couldn’t do it.
"Everybody loves Basketball-Reference.com. Except the Kobe fans".- DubsFan408
by GovernorStephCurry on Sep 8, 2010 8:16 PM PDT up reply actions
http://www.basketball-reference.com/players/c/curryst01/splits/2010/
It doesn’t have TS directly, but you can calculate it. I assume that’s what WYK did to get those numbers.
Indeed. Most of the stats shown in these splits are in their raw numbers. All that was needed to calculate the per-minute numbers and TS% was some simple math.
by WYK on Sep 8, 2010 8:47 PM PDT up reply actions
people are being way too harsh on the way Monta performed last year
My post really isn’t about that, though. It’s really about what the title says, should he shoot more or less? I’m not calling him out. If anything, I was a bit surprised not to find a trend that I thought would be there. I thought that as Curry played more, Ellis’ USG would go down and his TS would go up, but that doesn’t seem to be the case. I was also (mildly) surprised to find a trend that Ellis’ TS is positively correlated with USG, not negatively, which is what you seem to be suggesting by discussing fatigue, ill-advised shots, etc.
I’m just trying to objectively look into the numbers, man. If you can come up with some other data that take into account the “injury hypothesis”, by all means, do so. Maybe you can look through the game logs and calculate some kind “injury factor” and do a regression for Monta’s TS vs. IF. :)
No, no. I apologize if you took it the wrong way, I definitely appreciate your research with the percentages and every thing and by no means, was I belittling your post. It was more of a general thought about everyone vs Monta.
In regards to your post, he shot a ridiculous amount of shots last year because, well, he had to. I just think that all the hatred and even analysis from last year is really too hard to guage and it’s much ado about nothing.
I’m really excited to see how the Curry, Monta, and Lee trio works this year. Not saying they are a playoff contendor this year, but if chemistry clicks, we may surprise some teams. I’m thinking 36-40 wins
How come people never take into account that last year
Uh, you and every other Monta apologist does it constantly.
I must look like a dork.
by Reverend_Randy on Sep 8, 2010 2:02 PM PDT up reply actions
Time flies like an arrow. Fruit flies like a banana.
WD-40 will keep time flies off your arrows.
Lights please, lights please, turn off the lights.
by Skeptic con Urquell on Sep 8, 2010 3:07 PM PDT up reply actions
I can tell you just can wait
for the season to start. Neither can I. If dubs go pick n roll offense, I think you’ll see fewer isolation plays and fewer jacked up shots. Monta, Curry, Lee & Biedrins can pick n roll all game. We didnt have that last year, leading to iso’s
keep on truckin...
Two points:
First, don’t confuse correlation with causality. For example, when you write:
“This actually tells me that it is not necessarily the case that Ellis will shoot better (i.e more efficiently) if his role in the offense is diminished. Indeed, according to the data, he may shoot worse.” this shows me that you are in danger of making this mistake.
You seem to be implying that Monta shoots well when he shoots a lot. I’d like to suggest that the opposite is more likely to be the case: when Monta is shooting well, he’s more likely to take more shots. Therefore his higher usage is, in fact, a function of his higher TS … NOT vice-versa.
This strikes me as far more logical than the opposite possibility. Monta shot so much that it’s absurd to suggest that he “warmed up more” by shooting a couple of more shots a game.
My second concern with this data is that it’s obviously a really really noisy data set. Can you share an R^2 on your regression line? It seems self-evident from looking at the graphs that your R^2 is really low, and, in fact, it may not be true that your regression coefficients even stand the test of statistical significance.
Without those two numbers to provide vital context, you are practicing bad statistics. The incredibly noisiness of your data set suggests, at the very least, that your two-dimensional regression is hideously inadequate to the task you’re trying to measure.
Therefore I would encourage you not to try to draw any meaningfully conclusions from this analysis.
I agree with a lot of what you're saying
Did you read the last paragraph of the post:
Some questions to think about… Why is Ellis’ TS correlated with USG, whereas Curry’s does not appear to be? Is it that Ellis adjusts his role in the offense when he’s “feelin’ it”? Does he take fewer shots when he’s having an off-night? Or, perhaps, is it that Curry is getting Ellis the ball more when he sees that Ellis is shooting well? There are a lot of factors here obviously. What I’ve presented is only a starting point.
by Evanz on Sep 8, 2010 2:08 PM PDT up reply actions 1 recs
about the R^2
I agree completely. Are these data “statistically significant”? I need to get a p-value. However, given the data that I have shown, I think it’s safe to say that we could rule out the hypothesis, “Monta shoots better when he takes fewer shots.” Sometimes, negative data are as interesting as positive data.
Of all the regressions above, I think the only obvious statistically significant results will be the one that shows Curry’s USG went up during the season. But I don’t think we need to crunch the numbers to know that.
I was fairly careful to use the word “trend” because I knew that you and some others would bring up these issues. I’ve published several scientific publications, and I know when a claim is strong, and when it is not.
I agree completely. Are these data "statistically significant"? I need to get a p-value. However, given the data that I have shown, I think it’s safe to say that we could rule out the hypothesis, "Monta shoots better when he takes fewer shots."
No, you can’t. This is what statistical significance means. If the regression coefficient is not statistically significant, then you can not draw any conclusions from your analysis, positive or negative.
Monta may well shoot better when he shoots fewer shots, and this dataset can’t show it because it’s too noisy and the effect is small compared to larger variables that you weren’t controlling for. We just don’t know.
true enough
but the effect, even if statistically significant, does not appear to be meaningful in either direction. I’ll run the test tomorrow to find the p-value. I have a statistics toolbox in Matlab that I can use.
We just don’t know
Do you have any suggestions for how we can look at this more rigorously? I’m all ears.
Well ...
In theory, the way you’d do this would be to come up with a list of factors that you thought affected Monta’s TS%, usage being one of them, and regress that. See which ones ended up being statistically significant, remove the others, and rerun the regression.
However, when you run multivariable regressions you need a lot of data points to get meaningful data, and I suspect that we just don’t have them here.
Instead, you could create a larger data set by looking at individual shots. Take the list of factors that affect his shot (usage for the game this far, distance from basket, is he double teamed, type of shot, etc) and regress them based on whether or not an individual shot went in, using a binary variable as your dependent variable.
Obviously that’s a monster task. But I think it’s the only way you’ll get meaningful information along these lines.
when Monta is shooting well, he’s more likely to take more shots
That very well may be true, as I noted in my post. And if it is true, well, good! It would refute the idea that Monta just throws it up there willy-nilly, whether he’s having an off-night or not. Either way, I think we can learn something from looking at these data. At the very least, we can reject some hypotheses that are clearly not true (e.g. Monta’s efficiency goes down as he shoots more).
That very well may be true, as I noted in my post. And if it is true, well, good! It would refute the idea that Monta just throws it up there willy-nilly, whether he’s having an off-night or not.
Look at the slope of that regression line. If there’s an effect there, it’s tiny.
Don’t take this the wrong way, but you seem really eager to find something in this data which vindicates Monta.
you seem really eager to find something in this data which vindicates Monta
No, that’s not at all what I’m trying to do. I was simply trying to address the question of whether Monta is more efficient when shooting less. As you pointed out, the data don’t support that argument. At least, not the data that I’ve shown. Are you reading into it more than I am?
As you pointed out, the data don’t support that argument. At least, not the data that I’ve shown.
I believe the data neither supports nor refutes that argument, although of course I reserve the right to change my mind once I see an R-squared and some standard errors.
Saying “the data doesn’t support that argument” may be true, but the data also doesn’t support the argument that the sky is blue. That doesn’t make in an argument that the sky is green. :)
Agreed!
So when you hear someone make the argument, “Monta should shoot less, then his efficiency will go up.” I’m showing that it’s not obviously true. That’s all. Whoever is arguing that either a) is not using any stats whatsoever or b) is using some metric that you yourself said would be a “monster task” to develop. I suspect, and I’m sure you will agree, that most people belong in group (a).
I’m really not disagreeing with you. Not sure why you seem determined to make me.
I've calculated the p-values
and updated the post. You were right that neither set of data was statistically significant. However, the p-value for ellis was 0.06, which is almost there. The regression I performed does not discount outliers. Perhaps, the p-value would fall below 0.05, if that were done.
Thanks.
I have to say, however, that I hate discounting outliers to find statistical significance.
I will admit that I’m in the category who thinks Monta will score better if he shoots less simply by watching him – that is, so long as he stops taking the worst shots.
With a lot of players, I don’t think it’s reasonable to imagine that they’ll stop taking their worst shots. With Monta, however, his worst shows are so bad, and the alternative decisions on those plays are so clear, that it seems possible that he could do it.
I recognize that this isn’t a statistical argument, but I’ve never claimed to ignore what my eyes show me, either. :)
Interestingly enough, if we had a more sophisticated usage stat which included assists, I’m not sure that stat would have to go down for Monta to do better. It’s worth remembering that Monta is really really good at getting to the rim, and drawing a double team while doing so.
If he just consistently passed the ball in those situations, he’d be handlign the ball just as much, but the team would be doing a lot better. He could still be “the man” I wouldn’t even mind him running the point on occasion.
But I’m not sure he has the desire, and even if he did, I’m not sure he sees the passes as they develop the way a real creator like Nash does.
Sure would be nice to see.
one possible hypothesis
I’m trying to think of another variable that would influence Monta’s shot selection (or shot quantity really). One idea I have is to add the opposing team’s winning % as another variable and do a multivariate regression. It could be that against better teams, he takes more shots because he feels “pressure”. Or maybe not. What do you think about this idea?
I would include opponents FG% against, as well as some sort of variable for the individual defender Monta was facing (his Synergy rank?). I like those variables better than opponents winning , but the problem is that opponents FG against probably correlates reasonably strongly with win%, and when you regress with two highly-correlated independent variables it tends to make a mess of things.
I honestly really don’t think Monta is thinking about it that much. He goes out there and does his thing. I suspect that players “feel the spotlight” more in some games, but it’s not opponents win percentage, it’s things like “is the game on national TV?” “are we playing the Lakers?” etc.
Another thing worth looking at: what is the average NBA swingman’s correlation between usage and TS%. Before we suppose that Monta is doing anything unusual, we need to compare him to the norms for his position.
Players, rightly or wrong, have a tendency to want to feed a hot hand, and most feel more confident shooting the ball if their last shot went it. It wouldn’t surprise me if there was a slight positive slope for all players who play somewhat like Monta. (Heck, I can’t remember the source, but I think I remember reading something that just showed that a player was more likely to shoot if he scored on the team’s last play. I might be remembering that wrong, though, it was a long time ago, so don’t hold me to it).
So the slope we’re (not quite) seeing here could just be the NBA norm. Or, heck, Monta’s slope could be LOWER than average for guards of his type.
This is quite a little rabbit hole.
(Heck, I can’t remember the source, but I think I remember reading something that just showed that a player was more likely to shoot if he scored on the team’s last play. I might be remembering that wrong, though, it was a long time ago, so don’t hold me to it).
There are data indicating that a player who scored on the last possession is more likely to shoot on the next possession than chance would indicate. The flip side of that is that the player also more likely to miss than his overall FG% would indicate. I would seem that a player who “feels hot” because he’s made several shots in a row starts to take shots he’d otherwise not take. If the “hot hand” actually exists, it’s not enough to compensate for the more difficult shots that the guy who is “feeling it” eventually starts to take.
player was more likely to shoot if he scored on the team’s last play
I just read Stumbling on Wins recently, and I’m pretty sure they mention this.
Less.
He’ll be more efficient and there will be less pressure on him. Stephen Curry and David Lee will be a big help for him next season.
by Precise Films Productions on Sep 8, 2010 3:47 PM PDT reply actions
Usage Vs. Efficiency
There’s a couple problems here, one is that on a per game level usage and efficiency have a positive relationship. The Wage of Wins guys have shown us this. The thing is, neither you or Martin Schmidt have shown that the rule of usage vs. efficiency (the more shots you take, the less efficient you will be), while seemingly true on a per game level, is not true on a seasonal level. This is because of the hot hand phenomenon. If a player is making shots, he is going to take more shots. That’s how Anthony Tolliver is allowed to take twice his normal number of shots in a game, because he was shooting over 20% better than he normally does.
However, Monta’s usage was up across the board, so while the amount of games where he took, and made, a ton of shots was about the same as when he had a lower usage%, the number of games where he took and missed a lot of shots was much more than his lower usage year.
Monta Ellis's #1 Fan!!!
This is because of the hot hand phenomenon
Are you saying that I am ignoring this phenomenon? If so, perhaps I can refer you to the last paragraph of my post:
Some questions to think about… Why is Ellis’ TS correlated with USG, whereas Curry’s does not appear to be? Is it that Ellis adjusts his role in the offense when he’s “feelin’ it”? Does he take fewer shots when he’s having an off-night? Or, perhaps, is it that Curry is getting Ellis the ball more when he sees that Ellis is shooting well? There are a lot of factors here obviously. What I’ve presented is only a starting point. But I think that by looking into these kinds of per-game data, it may be possible to tease out some answers, or at the very least, ask some interesting questions. Feel free to suggest some other comparisons, or provide your own below!
I thought I was being fairly explicit in acknowledging that hypothesis. Was I not clear enough?
Not saying that didn’t acknowledge it, just saying that your results are pretty much exactly in line with what we’d expect, but it doesn’t suggest that Monta should shoot the ball more.
Monta Ellis's #1 Fan!!!
again, I'll refer you to my post
Ellis tended to shoot better when his USG increased. This actually tells me that it is not necessarily the case that Ellis will shoot better (i.e more efficiently) if his role in the offense is diminished. Indeed, according to the data, he may shoot worse.
Tell me which part of that is me suggesting that Monta should shoot more? I think you’re reading between the lines something that isn’t my intention.
results are pretty much exactly in line with what we’d expect
that’s not exactly true either, though. The regression fit for Ellis TS vs. USG is weak at best (as ronaldinho points out). For Curry, there hardly seems to be a relationship at all. Do you see a “hot hand” in the data I’ve presented?
I guess I'm not sure what you're trying to show here...
Basically, your graphs show me nothing. There basically scatter-plots with weak fitting lines. There’s not real conclusion to draw from them.
What I was trying to say is, if you do see a trend that games where a player has a high usage, his efficiency goes up, then that is actually normal, but is no way implies that said player should shoot more his next game (Basically, I was looking at the data you presented and answering the question in your title.)
You can easily see that on a macro level, the more possessions Monta uses the less efficient he is. In fact, I’ve personally written a lot on the subject. I even had a pretty picture:
The regression fit for Ellis TS vs. USG is weak at best (as ronaldinho points out). For Curry, there hardly seems to be a relationship at all.
A very weak positive correlation or no correlation at all IS what we expect here. From the article I keep linking to:
Specifically, if a player took 1 extra shot during the game, his field goal percentage would increase by a whopping 0.0017. 10 more shots during the game increased field goal percentage by 0.017. Not much of an effect. As we say in the book, not much oomph!
If you look on usage and efficiency on a per game basis, you are not going to find much.
Monta Ellis's #1 Fan!!!
You can easily see that on a macro level, the more possessions Monta uses the less efficient he is
Your plot has 5 data points. What do those represent exactly?
I think they're years
I must look like a dork.
by Reverend_Randy on Sep 9, 2010 1:20 PM PDT up reply actions
This is all academic
Ellis will not be a Warrior beyond December. This whole love in is a charade so Ellis’ trade value doesn’t plummet and he becomes impossible to trade.
Warriors and Ellis want the WORLD to this they’re happy together. Then the asking price will be higher.
Here's a proposal I haven't heard yet
How about Ellis to Memphis for Xavier Henry and Hasheem Thabeet?
No thanks
we can get Henry for a lot less. Like a future 1st rdr/expirings/and/or trade exception.
"It's like Will Smith, remember the Fresh Prince? Get the ball don't let nobody else shoot? That's kinda what the offense can be sometimes, and they're just standing around waiting for Monta to make a play"
-MT2
by golden_solitude on Sep 9, 2010 2:11 AM PDT up reply actions
Wait until next off-season to trade Ellis
His stock will go up. Since we all agree that last season was by far his worst, it is safe to say that he can only get better, thus increasing his value. We’re not going to make a run at anything substantial this post season, so we should just keep him around. We can then decide what to do with our cap space, what our team truly needs after observing a season with a complete roster, and we will learn the absolute value for Monta Ellis, as we look to add that final piece to the puzzle who will carry our momentum towards a big 10-11 season.
We’re not going to make a run at anything substantial this post season
unless Marc Gasol is a possibility…
Since we all agree that last season was by far his worst, it is safe to say that he can only get better, thus increasing his value.
In terms of his value around the league, I’m not sure this is true. Monta could score more efficiently and score less and be perceived as being worse than when he was among the top scorers in the league. In real terms, this forum contains a higher percentage of individuals who appreciate efficiency and many other factors related to improving a team’s odds of winning than exists elsewhere. A large number of GMs still value ppg primarily and would see a decline in point output with a higher efficiency as Monta ‘regressing’ from his 25.5ppg ‘explosion’.
No
Neither have proven anything
" Only build on positives , don't stack the negatives...Instead of criticizing , what was a positive?" - Donavin Darius
"There is nothing like a wise phrase or quote to help convince others that your decision makes sense." - Anon
"If you're ever in a fair fight, then your tactics suck." ಠ_ಠ
After further analysis...
After further analysis it has been concluded that Monta should indeed do more… with less
by WestCoastWarrior on Sep 16, 2010 3:08 PM PDT reply actions
Ellis needs to do more.
He could be our Tyreke Evans if he can get his game together. Ellis needs to do more to become a go to guy.in the context of what the team is trying to do. If he can become a consistent and efficient scorer, pass the ball to open teammates and play some defense once in a while, then Monta should be a keeper.
"Go ahead. Make my day."
Ellis probably did too much last season
it’s why his efficiency was so low. I mean, he probably did more than Tyreke Evans did last season.
I must look like a dork.
by Reverend_Randy on Sep 20, 2010 2:07 PM PDT up reply actions
Not sure he could possibly do more
He’s a great scorer don’t get me wrong, but I can’t see him becoming a 26-30 ppg scorer without hurting the team significantly. He doesn’t need to be our Tyreke Evans as long as we have Steph who can be our #2 or #3 scorer but our primary facilitator. Also Curry/Ellis/Lee can now share the scoring burden, making Ellis’ job easier this year. I look forward to seeing how the 3 play together.
COME. ON. YOU. SPURS.
by Tom Huddlestone on Sep 23, 2010 12:57 PM PDT up reply actions
Even tho I think Derrick Rose and Ellis have the same skill set I dont want Monta running the 1. I think Monta should be averaging 20-22 points at the most. Anything more means he is taking too many shots.
The reason it's okay for Derrick Rose to run the one is because he can control his dribble
slightly more willing passer too… but only slightly
I must look like a dork.
by Reverend_Randy on Sep 20, 2010 2:07 PM PDT up reply actions
DRose is a better decision maker
Meanwhile, Monta is on the short-list for lowest basketball IQ in the NBA
COME. ON. YOU. SPURS.
by Tom Huddlestone on Sep 23, 2010 12:53 PM PDT up reply actions
How about his defense?
On Monta’s shooting habits, I believe he recognized Curry’s 2nd half growth as a point guard and backed off quite a bit. This upcoming season will be interesting to see how Monta adjusts his game being a 2nd option in a pick’n’roll offense featuring Curry/Lee Curry/Beans. With Reggie Williams spotting up and Monta playing off the ball, the Warriors are going to be a handful to guard if they can stay healthy.
But I wanted to bring up his defense capabilities if nobody else will. His defense vastly improved last year as he was guarding the other team’s best option for most games. I believe he held Kevin Durant to 20 points for a game when he was making 30 ppg look simple. He’s gritty and he gets a lot of deflections and steals.
Monta is a valuable piece to the future of the Warriors!
Monta will not be the second option as long as he is a Warrior. He wont take a back seat to Curry. He’ll get along with him tho but he wont play second fiddle.
He should take a back seat to David Lee
I must look like a dork.
by Reverend_Randy on Sep 20, 2010 3:48 PM PDT up reply actions
Next year's offense
Will surely feature a lot of DLee in the pick and roll with either Curry or Ellis. I would like to say now that I want to see this split about 70/30 in favor of Curry, due to his better outside shooting and better decision making. Monta needs to get his points from the free throw line/slashing to the hoop/getting buckets in transition. Of course, if this were the case Monta would not be a happy camper.
COME. ON. YOU. SPURS.
by Tom Huddlestone on Sep 23, 2010 1:08 PM PDT up reply actions
This is easy
Less. Needs to realize that when Steph Curry plays, the team plays. Two main offensive options last year: Steph Curry creates good shots for either himself or a teammate, or Ellis isolation which many teams led to a questionable shot, to say the least.
Of course this is easier said than done, as Ellis will never take a back seat to Curry, a player who is younger than him, who the Warriors want to promote as the face of the franchise (instead of Ellis), and a player who was drafted significantly higher than him.
I wish Ellis would go back to being the player he was in the We Believe era…ridiculously efficient scorer, a catch on the wing and slash to the rim guy. Remember when Monta used to average 2.1 TO/g? He’s best as an off the ball scorer, letting Steph be the primary ball handler. Letting him dominate the ball is the worst thing the Warriors could do this season.
COME. ON. YOU. SPURS.
by Tom Huddlestone on Sep 23, 2010 12:51 PM PDT reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
"UNSTOPPABLE BABY!"
Golden State Warriors rookie Marc Jackson to the Mavericks' bench, after hitting a lay-up during a 29-point loss (2000)
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
Featured Stories
Most Commented
In Association With
Ball Don't Lie - NBA - Yahoo! Sport
- Days of NBA Lives: Wherein Roy Hibbert tap-dances to Maine
- Grizzlies Valentine’s Day cards are the key to love
- Kevin Love would like the NBA to take the MLB’s approach on All-Star games
- Kobe Bryant fires a few hilarious salvos at Jeremy Lin
- Brandon Jennings states the obvious regarding his 2014 free agency, everyone goes nuts
Certain photos copyright © 2012 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
