Help With Kobe-LeBron Article

@trodgers This is incredible! Thank you so much for taking the time to put this together. I am definitely going to take this information and use it in the article.

I think the big holes I have left at this point are around the efficiency arguments. I have been studying some of the schemes you mentioned designed to stop Kobe, and I've never seen anything like it used. The way the 2004 Pistons and 2008 C Bags focused on aggressively containing Kobe even on the perimeter is something no other player has faced. I know teams like the 2011 Mavs and Warriors as of late have centered on slowing down LeBron, but it's not the same thing. I am going to use that to explain away some of the discrepancies in efficiency (along with discrediting the notion that efficiency is the end all be all in general). I'm anticipating the counterargument that teams did not aggressively contain LeBron the same way because he's a good passer and would find the open man every time. This is valid, and I am not sure the best way to address this yet.

 

I've fleshed out the argument about LeBron's finals appearances. The conferences had huge discrepancies in talent over the years. In fact you could only make the 8th seed in the West with 50 wins at one point, but you could make the playoffs with a losing record in the East. This is compounded by the fact that Eastern conference teams play each other in the regular season more, inflating their records.

LeBron's teams also weren't as bad as people say. They were designed to put shooters and defensive players around him so he could penetrate and kick. This concept isn't new, and began with Allen Iverson in 2001. Some of those Cavs teams were ranked near the top of the league in defense, and we all know LeBron was not an exceptional on ball defender back in 2009 (and even so could not be responsible for his entire team being a well-oiled defensive unit). And in fact, most of LeBron's teams that were successful actually had at least 1 current all star and usually an all star that could be successful as a primary option on their own team. Kobe's teams post Shaq didn't always have that.

I can provide all the numbers and details about the offenses to back this up. But from the eye test we all know this to be true.

 

I'm brainstorming some of my arguments on why advanced statistics are essentially meaningless in absence of the narrow context under which they were developed. To preface this, just know that I work in a field dominated by statistics and numbers, all of which are 10x more complicated than any of these advanced stats. So when I say most of these advanced stats are bogus, I'm not doing so lightly. It's all a bunch of BS that came after that book MoneyBall (baseball and basketball are completely different games). You'll get more insight into my article, but I'd encourage you to look at a few things: See my thoughts below:

1. What is the formula behind these statistics? For example, win shares is based on something called marginal offense, which equals (points produced) - 0.92*(league points per possession)*(# of offensive possessions). Where did this 0.92 come from? And why are these the only numbers that factor into this statistic? Extremely arbitrary.

2. Look at who developed these stats. John Hollinger, for example, created PER. Dude has been a LeBron nuthugger for ages, and if you read his writing, it's quite obvious he's a moron. Most Laker boards have repeatedly confirmed this since I've been around (mid 2000s onward)

3. Look at rankings produced by these stats. For example VORP (this season) has Kyle Lowry and Jokic far over Durant and Curry and even Giannis. WTF? Obviously, common sense tells us that VORP shows nothing about who is the better player.

4. If you look at advanced stats, there's inherent biases in them. Offenses now are designed to optimize on these advanced stats (at least in some cases, like the Rockets) - making them more applicable to how the game was played now, but less so 10 or 20 years ago (Kobe's era). The stats reward a certain type of game (not better or worse, just different) - and the type of game it rewards is closer to what is seen now on the floor vs. Kobe's prime, for example. In statistics terminology, you have to understand what outcome is rewarded, and what the optimization metric is. These align a lot more in the modern era.

5. Reverse causality. If you use award winners to justify the validity of advanced stats, you have to realize the media, who have boners for these advanced stats, will use them to vote for award winners. It's no coincidence that Gobert lead the league in DRPM this year, and won DPOY. That's not to say he didn't deserve it, but 5 years from now, people will be looking back and saying "Hey, DRPM is a good stat, because Gobert lead the league in it in 2018 and won DPOY." But part of the reason he won DPOY was BECAUSE he lead the league in DRPM. Reverse causality.

Game 7 of the WCF is a prime example. Morey has his team shoot 3's because TS% dictates that the expected value of PPP will be higher when you take a 3 vs. a 2. Now what happened. The Rockets failed to recognize CONTEXT and that they were ice cold from 3, and went on to miss 27 straight, which 538 calculated to have odds of 72000:1. That's basically 0. Now what's more likely, the strategy based on advanced stats was wrong? Or that an event with essentially ZERO probability occurred. The Rockets failed to understand that the Warriors adapted their defense to the Rocket's strategy, that their players were out of rhythm, and that there was a lot of pressure that night. The fundamental assumptions underlying their strategy didn't hold anymore, but they stuck with it, and gave up the best chance they had at a ring since Hakeem was playing for them.

The coefficients used in these models were developed on data with underlying assumptions, but if the assumptions (or context) changes - then the statistic isn't as robust anymore. And a collection of garbage, extrapolated statistics leading to a conclusion isn't any better than a single garbage statistic saying the same thing, either.

Problem is you have writers like Bill Simmons, Zach Lowe, etc trying to act a lot smarter than they are (often with clear anti-Kobe biases) and quote these statistics like they are some gospel without understanding what they are actually saying. Most of them don't even understand BASKETBALL (different offenses run in NBA games, for example). I doubt they could tell you the optimal coverage off a pick and roll to stop a single side tag, for example. And if you ask people who understand these things (coaches, former players) who they would take, they usually choose Kobe over LeBron.

Basketball is too complex a game to be reflected by any one statistic, or even a collection of statistics. There are intangibles, offensive and defensive schemes, and other factors that simply cannot be modeled by a regression or mathematical function. It is far more sophisticated than any board game, and that's why they are nearly meaningless without the narrow context under which they were developed.

I still need to flesh out these thoughts a lot more clearly, but this is where I'll be going in the article.

 

Props to your efforts BFC on this article and your work on the subject...but its just a subject with no definitive conclusion to it imho.

KAJ said it best, there is no such thing as a GOAT or how to conclude that one exists. Its impossible to reach that conclusion and the primary way fans try to debate this topic is through stats, which like you said, is fundamentally flawed.

Aside from skewing numbers to favor your argument, the numbers will continue to change as the game continues to change. Players will continue to break offensive records (Bron racing towards KAJ as the alltime scorer) because the game has been tilted to favor scoring now.

Yesteryear favored the big man, defense and hard fouls....today's game favors splash more than smash. Bron's stats are clearly impacted by that and when you compare them to Kob's, Jordan's or KAJ's...you simply can't make that comparison even if the numbers say otherwise.



Still, good luck in your article BFC.

 

Completely agree it's subjective. And I have no problem if people put LeBron over Kobe all time. My problem is when people say "OMG LeBron is the GOAT cuz of muh WS and TS% while Kobe is fringe top 15 all time maybe".

There's a valid case for either, that's the point. But the media seems to thing he's leapfrogged Kobe and it's now only between him and Jordan. While that makes for a better headline today, I want to show that in depth analysis tells you that Kobe isn't a slouch, either, despite what the advanced stats say.

 

Kobe had to deal with a more defensive oriented landscape and had to play/develop alongside Shaq. Kobe also had to deal with a coach that would put him in check all while playing in a tougher conference. Bron's never had to do any of those things.

For a player that has never been considered a great offensive player to now have the ability to threaten KAJ's scoring record tells me enough of how the NBA landscape has changed and continues to change. Props to Bron for exploiting it, but to use that to deter Kob's personal accomplishments and have him sit out of the top 5 player discussion...