28
Jan

Data supports claim that if Kobe stops ball hogging the Lakers will win more

Tweet about this on Twitter54Share on Facebook316Share on Google+20Share on LinkedIn2Email this to someone

The Lakers recently snapped a four game losing streak. In that game Kobe, the league leader in field goal attempts and missed shots, had a season low of 14 points but a season high of 14 assists. This makes sense to me since Kobe shooting less means more efficient players are shooting more. Kobe has a lower career true shooting % than Gasol, Howard and Nash (ranked 17,3 and 2 respectively). Despite this he takes more than 1/4 of the shots. Commentators usually praise top scorers no matter what, but recently they have started looking at data and noticed that the Lakers are 6-22 when Kobe has more than 19 field goal attempts and 12-3 in the rest of the games.

kobelakers

This graph shows score differential versus % of shots taken by Kobe* . Linear regression suggests that an increase of 1% in % of shots taken by Kobe results in a drop of 1.16 points (+/- 0.22)  in score differential. It also suggests that when Kobe takes 15% of the shots, the Lakers win by an average of about 10 points, when he takes 30% (not a rare occurrence) they lose by an average of about 5. Of course we should not take this regression analysis to seriously but it's hard to ignore the fact that when Kobe takes less than 23 23.25% of the shots the Lakers are 13-1.

I suspect that this relationship is not unique to Kobe and the Lakers. In general, teams with  a more balanced attack probably do better. Testing this could be a good project for Jeff's class.

* I approximated shots taken as field goal attempts + floor(0.5 x Free Throw Attempts).

Data is here.

Update: Commentator Sidney fixed some entires in the  data file. Data and plot updated.

  • Jay

    There are certainly limitations to this data. Just a couple off the top of my head.

    1) There is a law of diminishing returns. If Kobe was to start taking 0 shots, the Lakers won-loss record would obviously not continue to improve.

    2) This also doesn’t factor in PPS (points per shot). If Kobe is getting fouled (or foul calls) and making a lot of FTs then it would appear that Kobe has taken less shots and yet the
    Lakers team would have more points (as a direct result of Kobe’s increased shooting).

    • Rafael Irizarry

      1) I agree. The linear regression is just a crass summary. Another limitation is that stars are rested when winning by a lot. But I do think the general pattern is a real thing.
      2) The X axis is shots taken (including when he gets fouled) not field goal attempts (well, an approximation). Note the footnote.

  • Adam Calhoun

    The traditional rebuttal is that Kobe just has to take more shots when the Lakers are doing poorly, because no one else is able to get anything in either; you'd then be getting the causation backwards.

    It would be interesting to see some kind of Granger causality: how does the +/- of the team affect his shooting % and vice versa.

  • August Blackburn

    I would like to point out that there is really nothing magical about 23%. That is just a fluke. Or is it? Jordan wore 23.

  • http://www.facebook.com/hauser.quaid.3 Hauser Quaid

    I've been watching almost all Lakers games this season. And I have to agree 100%. I would especially mention the last quarter shots when he shoots lot of low percentage shots (1m from 3pt line over hand of a defender).

    There is also another factor, when he shoots so much other teammates fall out of rhythm, everything halts, no ball movement, offence is static, and after a bad shot he can't get back into defense.

    One of his nicknames "Mr. Clutch" also does more harm than merit, as it's easier to defend against one than 5 players.

    On the other hand, last 2 games where another story, 14 assists, ball was moving fast, lot of open shots. If he took about 15 shots per game on average, they would play much better.

    I'm afraid that in the long run it's hard to expect this to continue, he's been shooting 20+ shot attempts his whole career. If however he does manage to change that, he'd be one of top 3 contenders for MVP.

  • RedDot

    I don't buy this at all. There are too many factors in a game to try to oversimplify it in these terms. The data can support damn near whatever you want it to support.

    How about the rest of the Lakers' FG% or overall "negative" plays, such as turnovers, free throws missed, fg missed, etc.

    I have seen the Kobe Bryant fga vs wins analysis done ad-nauseum, but never really an analysis of the rest of the Lakers in these situations.

  • Sidney

    After getting data from the NBA website, I have found that dataset originally used contains some wrong values:

    - 28 Dec 2012 game: Kobe FGA is 18 (not 17).
    - 05 Dec 2012 game: Lakers FTA is 22 (not 28).
    - 18 Nov 2012 game: Kobe FGM is 9 (not 7); Lakers FGA is 85 (not 90); Lakers FTA is 28 (not 18).
    - 16 Nov 2012 game: Lakers FTA is 28 (not 21).
    - 13 Nov 2012 game: Lakers FTA is 22 (not 20).

    These wrong points don't change significantly the overall conclusions stated in the post, except for some details like this:

    - When Kobe takes less than 23% of the shots the Lakers are 11-1 (not 13-1).

    I'm thinking about using data from NBA games for my class with Jeff Leek... Maybe I try to look at Michael Jordan data…

  • Sidney

    Looking at the data on the NBA website, I have found that some values of the dataset
    used in the post don’t match the NBA data:

    - 28.Dec.2012 game: Kobe FGA is 18 (not 17).
    - 05.Dec.2012 game: Lakers FTA is 22 (not 28).
    - 18.Nov.2012 game: Kobe FGM is 9 (not 7); Lakers FGA is 85 (not 90); Lakers FTA is 28 (not 18).
    - 16.Nov.2012 game: Lakers FTA is 28 (not 21).
    - 13.Nov.2012 game: Lakers FTA is 22 (not 20).

    These wrong points don’t change the overall conclusions from the post, except for some details like this:

    - When Kobe takes less than 23% of the shots the Lakers are 11-1 (not 13-1).

    This is an interesting example on how to use data analysis to explore everyday things.

    Here’s the graph I have obtained. Maybe I will try to extend this example for my class assignments.

    • Rafael Irizarry

      Do you mind making the dataset available with your corrections?

      • Sidney

        Sure!
        Here is the file content (I don't yet know how to attach a file here, so let's copy+paste):

        LA Opponente FGM FGA FTM FTA TeamFGA TeamFTA
        91 99 11 14 0 0 77 31
        106 116 10 20 6 7 72 32
        95 105 14 23 10 10 68 28
        108 79 5 10 2 2 77 26
        86 95 7 17 15 17 74 46
        101 77 10 18 5 6 90 28
        103 90 6 15 6 6 81 30
        82 84 12 19 2 2 74 22
        114 102 10 24 10 11 89 28
        119 108 9 18 3 5 85 28
        95 90 8 15 8 10 73 37
        97 113 11 20 11 13 65 39
        98 106 7 23 13 14 74 24
        115 89 6 11 6 8 82 34
        77 79 12 28 11 13 76 43
        122 103 5 15 4 4 87 17
        103 113 12 27 10 11 86 39
        105 107 14 31 9 12 83 34
        103 87 10 17 8 9 80 22
        108 114 11 24 9 10 84 31
        110 117 9 24 12 14 85 18
        94 100 16 28 7 10 75 40
        107 116 10 24 6 6 84 29
        102 96 9 29 11 13 82 28
        111 98 12 21 8 9 80 24
        101 100 11 24 6 7 88 27
        118 115 16 41 0 1 107 14
        100 94 14 24 5 7 77 31
        114 126 13 24 9 11 83 30
        104 87 9 18 8 11 86 28
        99 103 14 29 7 10 94 33
        102 107 15 25 7 8 76 36
        105 112 11 26 4 4 82 27
        112 125 8 22 2 2 91 13
        105 108 10 24 4 7 95 15
        101 116 8 23 10 13 98 22
        113 93 9 14 2 4 69 27
        104 88 12 19 4 4 84 18
        90 99 8 25 2 2 72 28
        103 108 10 32 3 3 88 21
        83 95 7 22 2 3 81 23
        93 106 11 23 7 8 73 32
        102 84 7 10 0 0 80 13
        105 95 8 12 5 6 74 29

        • Rafael Irizarry

          Thanks. Updated.

          • Sidney

            Rafael,

            The plot I obtained is a little different from yours. The three points that appear close one to each other in your plot, at the aproximate position (x, y) = (23%, 17 points) appear a bit spreaded in my plot. Since we are using the same dataset, I think that's may be any subtle difference between our codes. Because of this, your results indicates that the Lakers are more likely 13-1 when Kobe is 23,25%, whareas my result points out that the Lakers are 11-1.

            Could you show the exact formula you are using to compute Kobe's FGA% ? I'm using this:

            FGA% = 100 * (FGA + 0.5*floor(FTA)) / (FGA.Team + 0.5*floor(FTA.Team))

            Sidney

          • Guest

            in mine, the 0.5 goes inside the floor...

          • Rafael Irizarry

            floor(0.5*FTA) as opposed to 0.5*floor(FTA)

          • Sidney

            Ok, I changed the equation and got the same result as you.

  • OmarShaik10

    Hi everyone. I'm here to learn from all of you geniuses. First and foremost, how do you access the data sets directly from NBA.com?

  • BillP

    An alternate explanation is that if the opposing team is guarding Kobe more, he will take fewer shots and his team mates will be open more often. The plot may be less about Kobe and more about defense. One could imagine that the same correlation was present for all players - because it reflected asymmetry, not player productivity.

  • http://twitter.com/Scrilla100 Scott Messer

    NBA just updated stats to allow everyone access... http://stats.nba.com let's revisit the hypothesis!!

  • Roshan

    Hello Guys ,

    Thanks guys for the wonderful dataset and amazing insights about it . I have some basics of Data analysis and looking to better my skills . When looking at the graph I wonder how did you calculate Score differential for the datatset. This will help me in analysing the dataset

    Also has Jeff covered this in his coursera videos ?

    Thanks in advance guys

  • Sam

    So how'd it go with Kobe not being in the game at all?