I Knew Kobe Was Selfish!

Earlier this week I put up a post about how the production of certain players changes in clutch situations. The one piece of information in that analysis that didn’t initially make sense to me was the fact that Kobe Bryant averages more assists per 48 minutes in clutch situations than he does overall. This was confusing because it ran counter to my own perceptions and the bulk of the dialogue on Kobe’s clutch performance that was bouncing around the internet last week. Henry Abbot even used this point to highlight the piece when he linked it in his Tuesday Bullets:

Many players increase their field goal percentage with the game on the line. But get this: Kobe Bryant passes more in the final five minutes of close games than at other times.

Not only did my numbers show that Kobe Bryant dished out more assists in clutch situations, they showed he had the single biggest increase in that area of anyone in the data set. That single fact was frustratingly incongruent, but trusting my own calculations and the information from 82games.com I went with it. A question from a commentor caused me to look a little more closely into the matter. I discovered a very simple angle that I hadn’t considered, explaining the confusion. It turns out the number is technically correct but misleading nonetheless.

It’s absolutely true that Kobe averages more assists per 48 minutes in clutch situations. But he also attempts more shots. And more free throws. And more turnovers. In fact he averages more of just about everything in crunch time. The explanation is that he hands out more assists because he uses more possessions, essentially doing more of everything. He doesn’t actually pass more often, relative to the number of shots he takes.

The simplest way to illustrate this is by looking at the ratio of his assists to field goal attempts. On the season Kobe has an Ast/FGA ratio of 0.250. In clutch situations he has a microscopically higher Ast/FGA ratio of 0.256.

Another way to explain the increase in assists would be to look at his Usage Rate. Unfortunately we can’t calculate Usage Rate without some statistics about his team’s possessions. However, we can come to a very rough approximation. There are three ways a player can “use” a possession: a shot attempt, a free throw attempt or turnover. In addition we can look at his assists to get an idea of how many of his teammates’ used possessions he was involved in. If we combine the increase in his clutch field goal attempts, free throw attempts, assists and turnovers per 48 minutes we arrive at a total of 29.4. No one else in the data set was within 7 possessions of that total.

Kobe indeed hands out more assists in clutch situations but it still represents roughly the same percentage of his offensive distribution. He passes more because he does more of everything, not because he changes his mindset and works to involve his teammates more.

Looking at these numbers for Kobe led me to go back and calculate the Ast/FGA ratios and the change in those ratios for each player in the original dataset. Especially since I mischaracterized some players as being more willing passers in the clutch based on just those assists totals. The table below shows each players’s Ast/FGA overall, Ast/FGA in the clutch and the increase in their FGA/48, Ast/48 and Ast/FGA.

Looking at these numbers we get a slightly different view of the players who change their offensive role in late game situations. Here’s the 10 players who make the biggest transition from shooter to shot-creator:

  • Thabo Sefolosha, Ast/FGA +0.649
  • Jason Kidd, Ast/FGA +0.422
  • Mike Bibby, Ast/FGA +0.415
  • Andrei Kirilenko, Ast/FGA +0.398
  • Ryan Gomes, Ast/FGA +0.364
  • Tim Duncan, Ast/FGA +0.336
  • Nene, Ast/FGA +0.248
  • Marc Gasol, Ast/FGA +0.221
  • Dorell Wright, Ast/FGA +0.187
  • Landry Fields, Ast/FGA +0.172

Here’s the 10 players who make the biggest transition from shot-creator to shooter:

  • Luke Ridnour, Ast/FGA -0.438
  • Steve Nash, Ast/FGA -0.414
  • Hedo Turkoglu, Ast/FGA -0.339
  • Eric Bledsoe, Ast/FGA -0.335
  • Devin Harris, Ast/FGA -0.334
  • Chris Paul, Ast/FGA -0.329
  • Russell Westbrook, Ast/FGA -0.309
  • Jordan Farmar, Ast/FGA -0.285
  • David Lee, Ast/FGA -0.267
  • Mike Conley, Ast/FGA -0.264

The first list looks very much like an episode of Small Sample Size Theater. The second list makes a ton of sense. It’s populated almost exclusively by prolific shot creators. These are players who generally create a lot of opportunities for their teammates throughout the game and use the threat of that ability to create easy looks for themselves as the game is winding down.

Of course the only way to attempt to describe these numbers is by making generalizations. A player doesn’t necessarily have their Ast/FGA go down becaue he is a selfish ball hog. Perhaps opponents key in defensively on a talented teammate forcing that player to take shots he wouldn’t take earlier in the game. There are a multitude of factors which could explain the change in each player’s production from the rest of the season to clutch situations.

There are many different views on the best player in the clutch. If you ask me Dirk Nowitzki should be near the top of just about anyone’s list. My first contribution to The Two Man Game goes up tomorrow and will look at how Dirk’s scoring efficiency in the clutch is helping push the Mavericks over the top in close games. While we’re here looking at sharing the ball let me point out that other than Kobe, Dirk is the only player in the top 17 crunch time scorers who also increases his Ast/FGA ratio in these situations. In fact, you have to go all the way down the list to Paul Pierce at 24th to find a crunch time scorer who increases his Ast/FGA ratio by as much as Dirk.

I apologize if the information contained in my earlier post was misleading. I’m glad I was able to clear this up and demonstrate that Kobe is the same self-serving, puppy-eating, unicorn hating meglomaniac we always thought he was.



Filed under Dallas Mavericks, Los Angeles Lakers, NBA, Statistical Analysis

7 responses to “I Knew Kobe Was Selfish!

  1. Pingback: Dirk Nowitzki and Jason Terry help the Mavericks steal wins with the their clutch performance. | The Two Man Game

  2. boyer

    Wow, this is garbage. So, kobe increases his ast/fga in crunch time, and you still label him a ballhog because other players of your liking actually decrease their assists. That doesn’t even make sense. Your contradicting yourself, and wanting to always come to your own conclusion. Just like abbott. Anybody can come up with any sort of parameters and make these data support their already thought of conclusion.

    Also, you like most people equate assists with passing. Doesn’t work that way. You can pass the ball all game long and end up with low assist totals, especially with the triangle offense that the lakers use. But, if thinking vice versa, you do need to pass to get assists. So, people only look at assists, and think that whoever has higher assists passes more. Couldn’t be further from the truth.

    Also, if Kobe is a ‘ballhog’, why is that a bad thing? Him and jordan are ballhogs according to most people, and they have 11 rings and 13 finals appearances between them. Seems to me being a ballhog is a good thing.

    • I’m sorry I didn’t make my use of sarcasm more obvious. I don’t actually think Kobe eats puppies or hates unicorns. Other than the tongue in cheek title and those two jokes I think I went out of my way to not criticize Kobe. I double checked and at no point did I label him a ball hog. In fact I even made this qualification:

      A player doesn’t necessarily have their Ast/FGA go down becaue he is a selfish ball hog

      I would think it’s especially clear I am not using ball hog to characterize Kobe since I point out that his Ast/FGA ratio actually GOES UP. You’re absolutely right that assists are not a complete way to judge a player’s ability or willingness to pass. However, I’m aware of no statistic which does a better job of capturing this. If there is a better way to statistically measure a player’s willingness and ability to pass please let me know, I’m always looking for ways to improve my analysis.

      The entire point of this piece was to clarify a piece of data that was misleading. Regardless of what you think of Kobe, I believe anybody would be surprised to hear he passes much more often in crunch time situations. My intial post showed this to be the case. This post was merely meant to clarify that he does pass more often, but not relative to how much he shoots.

      I also think I was very careful to not draw any judgements on the value of his passing or shot selection at all. I don’t believe there is any spot in my post where I state that his passing or shooting is good or bad for his team. While the way he handles the offense in crunch time situations may not be my personal preference I completely understand how effective it has been for him historically. Kobe’s won 5 rings and he earned every one of them with his play on the court, you’ll get no arguements from me on that point.

      This entire post was meant to show that in crunch time he passes roughly the same amount relative to the number of shots he takes as he does overall. I’m not saying that’s good or bad, just clarifying the point because it wasn’t clear in something else I’d written. I’m sorry if my saracastic jokes confused the central point for you. I appreciate the comments and hope you’ll come back and read some other posts here. I do have plenty of analysis that doesn’t have anything to do with Kobe Bryant.

      • boyer

        After re-reading your post, yes, it doesn’t seem as if you’re ripping Kobe as much, though a little, regardless of your final paragraph.

        I guess you don’t make any real thorough conclusions, which I think you should’ve, regardless of what your opinion is. It seemed weird when you mentioned only kobe, dirk, and pierce as prime time scorers increasing their assists in crunch time. This seemed a little confusing, and you didn’t go too far into it.

        Most of these stats’ articles about ‘clutchness,’ are designed to specifically find fault into Kobe. I’ve read abbott’s article and articles about his article, about how he already had his conclusion, and he just picked specific times and data to support his conclusion the best. Not to mention, only looking at ‘game tying/winning shots’ without any other game analysis/situations is rather arbitrary. So, your initial article didn’t quite do that, and the fact that this article’s headline is: I knew Kobe was selfish!’ I would think you would understand if most people were misled by this article.

        Now, your initial article seemed to have some loopholes, or you didnt’ really tell much about it. Also, Wade isn’t even included, and less than half of nba players are involved.

        As far as assists and passing. Sure, assists is the best stat to use to somewhat correlate with passing. But, that doesn’t mean assists correlate with passing. There’s just no other current stat that can be used to do this. Kobe had his best assist totals/assist ratios, I believe, in 05, by far his worst laker team. Also, I’ve seen most of his games this year, and he’s passing a lot more than most years. Though, I think he’s always passed a lot, been pretty much the top assist SG in the league for many years now. But, his assist totals are only at 5.0 this season. The triangle offense is great for the ‘team’ as a whole to have a lot of assists, but it isn’t a great offense for one individual to average a lot of assists. Fisher might be the worst starting PG in the league, but he would still average at least 4 assists with any other team, whereas, I think he averages below 3.

        I think you seem like a stat guy, so stats rule above all else. I think a lot of the stats out there are good and relevant to an extent, but they have to be taken into context. Bruce Bowen is a great example: phenomenal defensive players for many years, but always weak in steals/blocks/PER. Which brings us to Kobe’s clutchness. I’m not sure what you think about it. But, do people honestly think Kobe isn’t clutch? That’s just absolutely absurd to me. I can handle if you don’t think he’s the most clutch, maybe, but to say he’s just slightly above average or average. People love hating on him and finding anything to point to him not being as nearly great as some make him out to be.

        As far as my most clutch players, not in any specific order: kobe, pierce, dirk, ginobili, wade, melo.

        Also, the with the ast/fga going down. I agree if that goes down, it doesn’t necessarily mean that player is a ballhog. But, if it went down for Kobe, just imagine the uproar.

        Personally, I really don’t think many players are ballhogs, or if by saying someone is a ballhog, then it really should be a complement. Iverson is the exception. The top players on their respective teams should have the ball more. Nash dominates the ball more than anyone in the league. But, the only thing people associate with ballhoggedness, is shooting. And Kobe is a SG, and he’s supposed to shoot the most, because that’s his position and he’s the best.

      • Thanks for coming back Boyer. Just to clear up a few points:

        – I only mentioned Pierce, Dirk and Kobe because they were the only ones in the top 25 crunch scorers who actually increased their Ast/FGA ratio. This means they are the only ones in the top 25 who totalled more assists relative to how many shots they took then they did during the season as a whole. Since I was focusing on that angle in particular they were the only ones I mentioned.

        – I can’t speak to Henry Abbott’s piece but I didn’t intend for this piece to find fault with Kobe. I was just very confused about how his assists totals looked in my first post on clutchness and went back to find an explanation. I found the explanation, thought it was interesting and wanted to share it. I picked the title and inserted a few sarcastic jokes about Kobe because I knew where most of the discussion was going and thought they would get a reaction. I’m actually surprised you’re the only one who has left a negative comment about that.

        – I agree with you that statistics don’t explain everything and that context is important. I also think there are way too many people who rely on context alone and don’t even bother with statistics (see most sports radio hosts). As always the real truth lies somewhere in between. I often don’t have enough knowledge of the context to make a definitive argument so I present the stats and try to let the reader draw their own conclusions. That is what is was attempting do here. I presented the information on Kobe’s Ast/FGA ratio, you get to decide whether it’s a good thing or a bad thing.

        – The reason Wade and many, many other NBA players aren’t included is that I didn’t have access to their data. I don’t have the time to compile all these clutch statistics on my own so I rely on even more devoted stats-philes than myself. 82games.com was the source of all my clutch statistics and I’m not entirely sure why some players are included and some aren’t. I believe a player has to have played in 10 games of crunch time with at least 40 minutes played to be included, but I’m not sure that’s the only factor. In my first clutch post I tried to say very clearly that it was difficult to reach concrete conclusions without the data for every player.

        I appreciate you coming back and having an honest discussion. Nothing troubles me more than when name-calling is substituted for honest dialogue. I hope you continue to read and offer your opinions here.

  3. Pingback: The Outside World – 2/11/11 | Pickin' Splinters

  4. boyer

    Ok, I see what you’re doing here, for the most part. And you’ll get much more recognition if it deals with Kobe, even in the slightest. There’s a reason why these articles are aimed primarily at Kobe. But, I see you’re not really supporting one way or the other.

    As far as statistics go, I understand nobody can watch every game; therefore, in the end, all we have are statistics. And people can make their arguments seem sound with any statistics they want, even if these statistics are quite misleading. I work in the scientific world, and I know that if you want to make your data look good, you can, even though it might be a pile of garbage. Same thing with sports’s statistics nowadays.

    I think clutch situations are extremely unique, especially if only a few seconds left on the clock. Each situation has to be taken into account. Here’s a great article chronicling all of Kobe’s playoff end-game shots, and basically showing that for the most part he was great in almost every game:

    As I mentioned before, bruce bowen looks like a terrible player with stats, but he wasn’t, at least defensively. And I think iverson/lebron/nash/probably most PGs a lot of times, if not most of the time, only pass to a teammate with a high probability of that teammate getting the shot, and getting the assist.

    Not sure if you saw lakers/celtics last night, but Kobe was passing great for most of the game, but only ended up with 4 assists. But, he passed out of the double well, and then that guy passed another time to a wide open player, but no hockey assists in b-ball. Also, when he passed, a lot of times that player gets fouled, but no assists there, even though that player goes to the line.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s