How to be statistically biased: Henry Abbott on Kobe Bryant

I've often struggled with my stance on the usage of advanced statistics in the game of basketball.  On the one hand, I love the information that can be gleaned by looking at statistics in different ways.  I find statistical arguments to be fascinating and compelling.  I've done a ton of work both utilizing advanced stats, and helping to simplify them for others.  I am firmly on board with the idea that advanced stats help us understand the game in ways which were not previously possible.

On the other hand, advanced stats are often used to make arguments that I don't agree with.  A big part of that is my own bias towards the Los Angeles Lakers, a bias I freely and willingly admit to, even as I try to avoid letting it affect my judgment.  Advanced stats have rarely given the Lakers the credit I feel they deserve as a team, and stats have very often been used to knock Kobe Bryant off the pedestal that he is given by his fans, and a large portion of the league at large (GM's, players, coaches).  There is no player more statistically controversial than Kobe.  People love to make statistical arguments about why Kobe Bryant is not as good as we think he is.  I don't agree with many of these arguments.

Disagreeing with a statistical argument is hard to do objectively.  After all, statistics are cold, hard facts.  Derek Fisher is shooting 38% on the year, and no matter how many ways you look into justifying that number, it is set in stone.  And the people who make statistical arguments are quick to point out the factual basis of their message.  In this, they are absolutely correct.  The problem is that a statistical argument can be presented with every bit as much bias as a non-statistical argument.  A prime example is the latest opus from Henry Abbott. 

Just so we're clear, this is not an anti-stats piece.  It's an anti-Abbott piece.

Abbott today made the case that Kobe Bryant is not the king of crunch time, as so many have labeled him.  His argument is not wrong.  It's not right, either.  It is simply flawed, in many ways both intentional and (hopefully) unintentional.  Let's go through his argument piece by piece.

The out-dated data set

Defining "clutch" has always been a sticking point in the argument of statistically defining end game greatness, so Abbott makes sure to cover all the bases ... sort of.  He starts with Kobe's ability to hit the big shot.  He brings up a piece on 82games.com which shows, over a five year period, that the only reason Kobe is even in the conversation for "Best closer in the game" is because of the sheer volume of attempts that Kobe has taken in these types of situations.  He throws out some grim numbers.  In these situations, Kobe has made 14 "big" shots in 56 opportunities, with only 1 assist to his name over that time period.  Names like LeBron James and Vince Carter are higher on the list, both in terms of volume of big shots made and consistency.  There are many, many names, both stars and role players, that shoot for a higher percentage than Kobe in these situations.  Pretty convincing stuff.  There's just one problem ... take a look at the data set.  All data for that piece comes from 2003-2009, and includes only 1/2 the 2009 season.

Let's take a quick (and painful) trip down memory lane for those seasons.  03-04 was the last year of the Shaq-Kobe union, the year Kobe was dealing with rape charges in Colorado, and the season in which his identity as a selfish player was coming to a head.  The next season, Shaq left, and Kobe was left with a roster of bit pieces, until 2008 when Andrew Bynum started developing and then Pau Gasol was acquired.  So, the first sample size Abbott uses to point the accusing finger at Kobe is one in which Kobe was at his most selfish, and his team had the least amount of quality.  Interesting.  He then uses stats provided to him by somebody from ESPN's stats department that show Kobe isn't much better over the course of his entire 15 year career.  OK, point taken.

Quick clarification: I'm not making any excuses for Kobe Bryant here.  Yes, in 2004 he played selfishly, and it's quite possible he cost that team a chance at a championship.  He brought the crap roster that followed the next few years on himself, and he continued to play a me-first brand of ball through many of those years.  If Abbott were making this argument about Kobe 2 or 3 seasons ago, I'd have no more ammunition to refute it with than "Abbotz a HATERZ". 

Instead, the data set used for his argument is just a wee bit out of date.  Let me fill in the blanks.  Since the beginning of 2009, Kobe Bryant is 11 for 22 in those same "big" shot situations.  He's also got three assists in that time period, which is a higher ratio of assist to FGA ratio than LeBron James had in the initial study.  Now its time for a shocking admission; over that same time period, Kobe is 0-4 in the playoffs.  Notice, that little tid-bit of information could be used to refute my overall argument, and yet I still include it.  Why?  Because I'm trying to make this argument as un-biased a way as possible.  Besides, even with those four misses, Kobe is shooting 50% in do or die situations over the past 2+ seasons.  Gee, that paints a little different portrait, doesn't it?

The sin of inconsistency

There is nothing inherently biased in using an outdated data set.  That he uses it without investigating if more recent events change the picture is sloppy, and potentially lazy, but its not done out of agenda.  His next trick is much more underhanded.  Observe the following transition from his piece:

One shot for all the cookies. And the NBA is nearly unanimous that this is the guy to take it, even though he has more than twice as many misses as makes?

His crunch time production is slightly higher in the first half of this season, but still certainly not the best in the league. And analyzing any large number of games, one year, five years or fifteen years, and defining crunch time a number of different ways, shows the same pattern. (There are a many ways this has been sliced.)

In case you missed it, that paragraph change over indicates when he stopped talking about "big" shots and moved on to the other definition of "clutch", that being the last 5 minutes of any game in which the score is within 5 points.  I can't blame you if you did miss it, since THERE IS NO INDICATION THAT HE HAS CHANGED THE CRITERIA OF HIS ARGUMENT MIDSTREAM.  If we are sticking with the whole "big" shot criteria, the Lakers have only been down 2 or less with 24 seconds left to go twice this season.  In those two situations, Kobe is 1 for 1 (ironically, this shot did not win the game, as Derek Fisher had to hit a layup as time expired to beat the Clippers) and has an assist to Steve Blake for a game winner (in the first game of the season no less.)

I'm sure, if asked point blank, Abbott will state that he was simply trying to cover all the bases, but you might want to let people know which bases are being covered, especially when one base and another base are making different arguments.  Oh wait, they aren't making different arguments.  Notice the phrasing of the first sentence of the 2nd paragraph.  "His crunch time production is slightly higher in the first half of this season, but still certainly not the best in the league".  If you actually click that link, you'll notice that this is accurate ... because the first six names are all point guards.  Through the first half of this season, Kobe Bryant is apparently the most clutch non point guard in the NBA, something the post's original author admits painfully (because, you know, that's what unbiased pieces do when they find out something counter to their own argument).  But Abbott phrases the link as if this is just another piece that confirms his message all along.

And then there's the little issue of whether this method of reviewing clutchness gives point guards an unfair advantage.  With 17 of the top 30 being point guards, including such notables as Jordan Farmar and Ty Lawson, you have to at least ask that question right?  Point guards all "generate" a ton of points per possession in the clutch.  What do all point guards have in common?  Assists.  What happens when a point guard passes the ball to a teammate and the teammate misses the shot? Bingo.  Point guards get credit for their assists and don't get punished when their assist doesn't pan out.  Meanwhile, scorers like Kobe have just about everything they do count against them.  This does NOT refute the concept that all the players ahead of Kobe are more clutch or more valuable to their teams in the clutch.  It does NOT refute the concept that, if Kobe passed the ball more in crunch time, maybe his numbers would be better.  It DOES question the validity of a data set that concludes backup point guards are more clutch than some of the game's biggest prime time performers.  To not ask that question, and accept the data set completely is irresponsible at best, and dishonest at worst.

The un-clutch Lakers ... of the past 15 years

Ah yes, the team argument.  Here we once again see Abbott delve into Kobe's entire career to make a point about how he is not the king of crunch time right now.  Somehow, I don't think all those GM's and players had 1998 in mind when labeling Kobe the guy they want taking the last shot.  Then we get this statement ...

Over Bryant's career, 11 teams have had better crunch time offenses, led by the Hornets with a shocking 107 points per 100 possessions in crunch time, a huge credit to Chris Paul. 

Kobe Bryant is in the middle of his 15th NBA season.  Chris Paul is in his 6th. And yet, over Bryant's career, the Hornets lead the league in points per possession thanks to Chris Paul.  Some part of that is wrong, and it doesn't even matter which part.

That point aside, guess what?  His argument regarding the Lakers' team performance in crunch time is well done.  I don't know exactly which time frame the numbers come from, but the links are at least all rather timely.  I can admit that, because I'm trying my hardest to make an unbiased argument.  If the numbers are legit, and I have no reason to think that they aren't, it's a compelling point that Kobe may not be the best leader in the clutch.

But a compelling point isn't enough.  Abbott concludes this section thusly:

A great offensive team performing at average levels, with a star setting records for number of shots attempted. Teammates left wide open. Evidence, even, that Bryant's play puts his team into nailbiters that needn't be so close.

That, my friends, is a ballhog.

I find the link particularly humorous.  It's a link to Ron Artest's game winning putback against the Phoenix Suns, and the teammate left wide open is presumably Derek Fisher, who's man leaves him to attempt a block on Kobe's shot.  Never mind that there was less than 3 seconds left on the clock (how many two pass shots have you seen with that little time left?) or that Fisher being "wide open" invovled his man being about as far away from Fisher as he was intially from Bryant when Bryant went up for the shot.  Clearly, this play shows that Bryant is a ballhog, and illustrates the overall point provided by all that evidence.

The evidence that is once again taken from Bryant's entire career.  Ready for another shocking admission?  There have been times in Kobe's career in which he has absolutely been a ball hog.  The beginning when he was trying to establish himself as a star, the tail end of the Shaq-Kobe years when he was trying to wrest control of the team away from Shaq, the post-Shaq years when he had no choice but to be a ball hog.  Even now, there will be games or entire stretches in which Kobe does not play within the confines of the offense.  If you choose to focus on these stretches, you could say that Kobe is still a ball hog.  Of course, there are other periods in recent memory (a significant majority of them, by the way) in which Kobe plays within the offense, takes what is there, and distributes for his teammates.  And it happens more and more with every passing season.  One wonders if Abbott knows that Kobe is posting his highest assist percentage since 2005 this season, or his highest PER since the Lakers acquired Pau Gasol.

Seeing what you want to see

Abbott spends the last two sections trying to figure out how so many people have come to their (incorrect) opinion regarding Kobe's value in crunch time.  He theorizes that, by sheer volume, Kobe has imprinted so many magical moments in our memories that he comes to mind automatically.  We remember his makes and ignore his misses and that's why we are fooled.  That's not just us silly fans by the way.  This is why the players, GMs, and coaches also think Kobe is the best man for the job of clutch shooting.  We are all misled by our own minds, because we see what we want to see.

In all honesty, he's probably right.  But once again, the way he makes his point does more damage than good to his argument.  We are once again treated to career statistics, and then he goes to Phil Jackson's book to find statements about Kobe, unearthing these gems:

However, don't confuse Bryant's domination of the ball with Jackson's endorsement of the plan. In the same book, Jackson tells of his annoyance at Bryant's ballhogging in crunch time. In one instance, he describes drawing up a play with multiple options, in crunch time of a 2004 playoff series against Houston. Bryant destroyed all the options; instead of setting a baseline screen for Shaquille O'Neal he ran straight to the ball. "With the twenty-four-second clock winding down," writes Jackson, "Kobe forced a long jumper, a horrible shot in the game's most critical possession. The ball did not reach the rim..."

Jackson also tells of marching, more than once, into Mitch Kupchak's office to demand that the Lakers trade Bryant. He writes things like:

• "Kobe tends to hold on to the ball longer than necessary causing the offense to stagnate."
• "He won't listen to anyone. I've had it with this kid."
• "As usual, Kobe seemed intent on taking over."

Never mind that Jackson decided to give it another go with Kobe after the book was written.  Never mind that Jackson and Kobe have combined to win another two championships, with a decent possibility of at least one more.  Never mind that Phil has been close to retirement the past two seasons, and Kobe has used every ounce of his persuasion to ask Phil to come back just one more year.  Nope, Kobe was a ballhog once, Phil Jackson said so, and now it is writ in stone.

This is a unifying theme throughout the piece.  He picks and chooses data from throughout Kobe's entire career that suit his argument, and is ignorant of, or chooses to disregard, any piece of evidence counter to his theory.  He accuses fans, players and GMs of remembering Kobe's greatness in our hearts and letting it cloud our judgment, but how is that different from the system he has displayed to make his entire case?  Actually, there is one rather large difference.  We make that mistake because we forget.  He makes it by choice.

That, my friends, is a hypocrite.

The real truth

The truth about Kobe Bryant in crunch time is that there is no truth.  Kobe Bryant is not perfect, far from it.  Even as I'm willing to sing his praises, he still goes through periods that leave me frustrated with him, for all the same reasons that Abbott uses to attack him.  He spends over 2500 words trying to convince you that Kobe Bryant is a selfish ballhog who is not very successful in crunch time situations.  Three years ago, he might have been right.  He uses numbers that appear convincing, because he ignores any numbers that don't fit his argument.  And he has the audacity at the end to say that any open mind is forced to acknowledge his argument based on the numbers.  In the end, the piece is a much more accurate reflection of Henry Abbott than it is of Kobe Bryant.

Is Kobe the best clutch player in the league?  It seems that he's certainly in the conversation, but that's all it is, a conversation.  That conversation will continue to evolve.  Based on the numbers that I've produced, I think I've shown that Kobe Bryant continues to evolve.  The only thing not evolving is Abbott's opinion of Bryant.

As long as your mind is open to all that, it has to be closed to the idea that Henry Abbott is capable of objectivity towards Kobe Bryant.