Wednesday, September 10, 2008

Most Valuable Column: Individual Performance

Josh Tucker is the author of Respect Kobe. His Most Valuable Column seeks to track the MVP race and in preparation of the upcoming season's race, he's breaking down the elements involved. Today he covers numbers, flash, and exposure. Like talking about cameras, only with lots of Kobe defense. Enjoy. Let him have it in the comments, stat-heads and Kobe haters!

First, apologies for my prolonged absence to all 3 of you who, in early September, already care about the MVP Race. Though I often consider it a fairly unimpressive excuse, the last several weeks have been absurdly busy for me. I'm back, and I have a plan for discussing all of the various criteria that could be applied to the MVP decision before the season starts.

In the last edition of the Most Valuable Column, we discussed what I have come to refer to as the 50-Win Standard. I presented evidence of an overwhelming and thoroughly established trend in favor of the standard, discussed the likely reasons for existence, and presented my argument for why it should not be changed.

Therefore, since the last incarnation of this column focused on team success as a measurable criteria for MVP candidacy, today's will focus on individual success.

Cheer up, LeBron*, it's time to talk about Individual & Statistical Dominance!

The reasoning behind this criteria is obvious — at least, I should hope so. (If it is not, then I fear that critical thinking may not be your strong suit; perhaps you'd be better off reading a Bill Simmons column, for which you need only take on the role of the baby bird, opening it's mouth to accept whatever its mother vomits up — an entertaining, if completely unappealing, process.) As those who oppose the 50-Win Standard have often argued, the MVP is an individual award; therefore, it only makes sense to consider the candidates' individual performances.

To an extent, those who make this argument are quite right. I think they take it too far, often arguing that only individual performance, and not team success (and, therefore, a player's impact on said team's success), should be considered in determining the MVP. As you have seen, I clearly believe that team success should be a significant factor.

Nonetheless, it is obvious that individual success must also be a factor. Few players can score at the level that Kobe Bryant and LeBron James do. Few can rebound like Dwight Howard, or assist like Steve Nash and Chris Paul. Fewer still can perform several of these skills well.

In addition to these three primary statistical categories, which tend to receive most of the attention, consistent defense is often a rare commodity in the NBA. And of course, there is the question of how all of these statistics should be evaluated, so as to keep them in context.

Also to be considered is the role that a player plays on his team — credit must be given to the players who score at a high level while being forced to create most of their shots, more so than to the players who score in high volume on shots primarily created for them by a teammate.

Finally, while we're on the topic of individual success, what of showmanship and a player's performance on national television? It may seem like a nonsensical concept at first, but it's not necessarily invalid to suggest that a player should be rewarded according to the size of the stage on which he plays. Since impressive performances in the Finals are valued more highly than those that come during the regular season, shouldn't we apply the same standard to regular season play, valuing more highly an impressive performance on national TV than one that is not?


First, let's talk about statistics. Clearly, they are the easiest way to measure individual performance... or at least, the most obvious and accessible. On the other hand, virtually any NBA fan who has engaged in a discussion comparing two players that centers on statistics understands that they are limited, and often do not tell the full story. Furthermore, it can be very easy to manipulate statistics, with each side often using them only in ways that support their case. Which statistics should be used, and in what context should they be viewed? Should certain statistics be valued more highly than others?

Before we go any further, let's make one thing clear: In this space, we are not looking for a single number that represents the cumulative statistical value of a player. By their very nature, I am highly skeptical of any attempt to sum up a player's value, production, or efficiency in a single number. To me, any such metric — at least at this point in time — is experimental, likely flawed, and more valuable as a thing of intrigue than as an authoritative manner of comparing players (though if anyone can convince me otherwise, HP stat guru David Sparks is the guy that could do it.) While my goal is to establish a method for determining the MVP, and numbers will factor into that method, subjective evaluation remains very highly valued. Thus, the method employed here, in the Most Valuable Column, for determining the MVP will be a clearly articulated, but nonetheless subjective one.

To get back to the question at hand: Which statistics are we looking at? First, I suggest we divide the statistics up into offensive and defensive stats. My primary reason for doing so is that statistics that are classified as "defensive" are extremely limited and, in fact, often misleading. Steals, for example, are often an indication that a player is player the passing lanes, rather than playing solid man-to-man or team defense. While that may have been a sound and very positive strategy for 2008 Gold Medalists Team USA, it is not often a sound defensive strategy in NBA basketball.

Meanwhile, blocks are somewhat positional by nature. Shots closer to the basket tend to be easier to block, and at the same time, defensive players who play closer to the basket tend to be taller and heavier. Both of these factors lend to the fact that shot blocking is a skill that is rarely, if ever, expected of perimeter players, regardless of any given perimeter player's defensive tenacity.

There are attempts being made to develop more advanced metrics which could measure a player's defensive impact more reliably. Of these efforts, defensive Adjusted +/-, which tells us how a team does defensively while a player is on and off the court, is the only one I feel comfortable endorsing as a valid measure of a player's defensive tenacity. It is also, according to our stats guru, David Sparks, considered to be the only metric that captures defense in any real way. Still, even Adjusted +/- tells us little about the differences between a player's team and individual defensive abilities, or one player's willingness and ability to take on the other team's top player, and successfully shut them down (or at least make 48 minutes of their life more difficult than usual). Therefore, I would suggest that individual defensive valuation must be a combination of subjective evaluation and the use of some advanced defensive metrics — namely, Adjusted +/-.

(As a quick aside, offensive Adjusted +/- ratings are also a good metric. Unlike PER or wages of wins, which attempts to derive an overall number based on box score stats — who decides how valuable a rebound is, or an assist? — and fails to account for all the things a player can do that do not show up in the box score, Adjusted +/- attempts to show a players complete and overall effect on the game, as reflected by the score. After all, if a player has a positive overall influence, even if it is in ways that do not show up in the box score, shouldn't that be reflected in the game score?)

Next, let's talk about rebounding. As a composite number (total rebounds = offensive rebounds + defensive rebounds), it is neither an offensive nor a defensive statistic. Nonetheless, it is often used as the third defensive statistic, along with steals and blocks. Again, however, rebounding is somewhat a function of position. Guards are not expected to be filling the paint, boxing out bigger guys, and securing rebounds — if they did so, there would be no one to receive the outlet pass. They are also shorter and lighter than players that play close to the basket. So, once again, both size and position on the court directly affect this statistic. Yet, we wouldn't conclude that Al Jefferson is a better defender than Bruce Bowen, even though Jefferson rebounds at a much higher rate than Bowen.

Finally, let's talk about offensive statistics. Scoring obviously comes to mind, and it is important. After all, putting the ball through the hoop is a crucial component in any successful game strategy. It is also important, however, not to overvalue scoring. Sheer point total should not be the primary concern here. Instead, a player's ability to score when his team needs it most, and his ability to do so with little or no help, should be important factors. Note that this does not necessarily refer to "clutch" shooting (we will address that topic at a later date), but instead refers to a player's ability to score when his teammates are unable to, ending or preventing a scoring drought, and his ability to create his own shot and score with little help from his teammates.

Assists are also very important, but can be directly affected by the offensive system employed by a team, as well as a player's role in that system. Some teams run a point-driven system, in which a point guard (or sometimes a point forward) brings the ball down the court, sets up the offense, and delivers the ball to his teammates in positions from which they can score. Other teams, such as the L.A. Lakers with their use of the Triangle Offense, employ a system that specifically prohibits any one player from dominating the ball. Naturally, a point guard playing in the former system would dish out significantly more assists than he would playing in the latter. Meanwhile, LeBron James, who plays the role of point forward for Cleveland, should be expected to tally more assists than Carmelo Anthony or Richard Jefferson, who, despite playing the same position (small forward), do not fill the point role for their teams.

In addition, centers and power forwards are almost never expected to accumulate high assist numbers — it's simply not their role, and it would usually be counter-productive for them to attempt to take on that role. That is, for a center to attempt to average high assist numbers would often make him less valuable to his team than he would be while playing a more traditional center role.

Finally, efficiency is another frequently touted offensive statistic. Usually, efficiency is expressed in the form of Field Goal Percentage (FG%) and Free Throw Percentage (FT%). Even this, however, is not without limitations. Centers, for example, play very close to the basket; therefore, the vast majority of their shots are considered "high percentage shots" — that is, they are expected to make a high percentage of them. Perimeter players, on the other hand, typically take more shots farther from the basket than do front court players. Therefore, the unadjusted FG% of guards tends to be significantly lower than that of centers.

Unfortunately, this is not an accurate measure of offensive shooting efficiency. For starters, the three-point shot is a significantly lower percentage shot than the close layup, but it is also worth more points. Consider the following example:
A point guard takes 10 shots, all three-pointers, making four of them. His FG% is 40%, and he has scored 12 points. Meanwhile, a center also takes 10 shots, all close to the basket, making six of them. His FG% is 60%, but he has scored the same number of points as the point guard: twelve. Therefore, while the point guard may have shot a lower FG%, his offensive efficiency is equal to that of the center who shot a higher FG%.
In addition to all of this, FG% does not account for made and missed free throws; instead we are required to use FT% a separate statistic. This can result in the free throws being undervalued. It also creates a disconnect between free throw shooting and overall shooting efficiency. A play on which a shooter is fouled and does not make the basket is not counted as a Field Goal Attempt (FGA) for that player in the box score. Nonetheless, it was an offensive possession that the player used, and making those free throws should be considered equivalent to making that shot.

This is where John Hollinger's True Shooting Percentage (TS%) comes in. TS% counts three-point shots as being 1.5 times as valuable as two-point shots, such that shooting 40% from beyond the arc is made equivalent to shooting 60% from inside the arc. It also factors in free throw shooting, such that two free throws equal roughly one shot attempt (.44 to be exact, accounting for and-1 free throws), and a player's FT% affects their TS% as though every pair of free throws were a regular shot.

Thus, while Chris Paul tends to shoot a few percentage points higher in FG% than does Kobe Bryant, the two players' were equally efficient in 2007-08, both shooting at .576 TS%.

Given all of the above, here is my proposal with regards to measuring individual performance:
  • Defensive ability and performance should not be measured in steals, blocks, and rebounds. Instead, they should be evaluated with a combination of defensive Adjusted +/- ratings and good, old fashioned, subjective observation.
  • Similarly, offensive Adjusted +/- should also be considered. It is not a be all, end all, as good players with poor teammates will tend to have higher scores than those who play with better teammates — for example, a team with a very solid second string, capable of holding its own against for a while against many first string lineups, might skew each other's +/- numbers. Nonetheless, this is a good metric, and deserves consideration.
  • A player's ability to score should not go beyond his scoring average, but should instead focus on his ability to take on the offensive load when needed, even when receiving little offensive help in their quest to keep their team in the game. (i.e., That for which Kobe Bryant and LeBron James are well known, and Kevin Garnett is not.)
  • Assist numbers should take into consideration a player's position, and the system in which they play. A player who plays in an offensive system that tends to reduce individual assist numbers (i.e., Kobe Bryant) should not be penalized, nor should one who is not asked to be the primary ball handler or distributor (i.e., Dwight Howard, Kevin Garnett).
  • Rebounding numbers should be evaluated according to position. Players should not be compared across positions, comparing a center to a point guard or a power forward to a shooting guard. Instead, we should look at where a player ranks at his position. The fact that LeBron James does not rank in the top 25 in the league in rebounding is meaningless. The fact that he ranks #1 among small forwards, and by a comfortable margin, is what tells us that he is an excellent rebounder.
  • True Shooting Percentage (TS%) should be used to evaluate a player's offensive efficiency, instead of Field Goal Percentage (FG%) and Free Throw Percentage (FT%).

These are the guidelines I propose for evaluating statistical performance. As I've already mentioned, some of these things require a certain degree of subjective evaluation, and subjective judgment is required to determine how they all fit together, and how they fit into the larger picture of the MVP race.

One thing is for sure: Individual statistical performance is far from the only factor to be considered in the quest for the MVP Award, but nonetheless, statistical dominance should clearly be rewarded.

Showmanship & Prime Time

Before we wrap this up, let's talk a little about two elements of basketball superstardom that are more elusive and less quantifiable: showmanship and prime time performances. At first, you may think (as I did) that this is a silly discussion. After all, the MVP award is presented to the most valuable player, not the most impressive. And that value is value that he adds to his team, on the court — not value that he adds to the networks' or franchise's corporate bank accounts. So why should showmanship and how a player performs on national television (as opposed to games televised only locally) matter at all?

With regards to showmanship, I'm not sure that I have an answer for you. Why does it matter? I don't know. There are, of course, certain occasions on which showmanship does actually benefit the team. However, these are fairly rare, especially during the regular season, and most of the time it is of little benefit to the team. So why are we talking about this? Because, by all appearances, some members of the voting sports media seem to be significantly influenced by it.

Mr. Paroxysm himself, for example, has expressed to me his belief that Tim Duncan would have won more MVPs were he flashier, drawing more attention to himself. Certainly, it can't be denied that the media — much like the vast majority of fans — gives special attention to those who are not only effective, but to those who look good doing it.

Consider also Kevin Garnett. Half of the MVP argument for Garnett last year was his "intensity" on the court — and yet, when asked for evidence of said intensity, his proponents usually pointed to his terrifying yells and menacing grimaces, muscles clenched and throat tensed. While this visual display does not necessarily indicate intensity and focus any more so than the determined drive of Kobe Bryant or the patient relentlessness of Tim Duncan, it is visually appealing to fans and media types alike, and effectively earned Kevin Garnett the overstated title of "most intense player ever!" last year.

And what of prime time performances? Earlier, I made a point that bears repeating. Over and over again, we as sports fans have proven that we value an impressive performance even more when it is delivered on a large stage. An impressive feat in the Playoffs is worth more than the same feat in the regular season; in the Finals, it would be worth even more. And though you don't have to agree with it, you have to admit that it is a potentially valid position.

It's also completely wrong.

Listen, we know that the voters are influenced by showmanship, and we know that they are influenced heavily by performances on national television. After all, many of them probably don't watch much NBA basketball beyond what is nationally televised. They're getting paid the big bucks, and the most popular matchups — the ones that generate the most hits for their columns — tend for the same reason to be those that are on national television.

But that shouldn't matter.

Here's something you're going to "hear" me say a few more times before this is all over. Value is what we're after here, and more specifically, value as it relates to the team. (Not the team business office or the team ownership, but the team of players on the court, and their basketball success.) Since a player's value to the team is only meaningful if it is felt by the team, anything that factors into the MVP discussion should also be a factor in contributing to overall team success.

Except in the rarest of occasions, neither of these things do that. As Tim Duncan has proven, showmanship has little effect on basketball success. And as Golden State can attest, after winning 48 games and still barely missing the Playoffs, a win that no television broadcasts is every bit as valuable as a win that is shown on TV screens across the nation.

I make no secret of the fact that I am a Kobe Bryant fan, trusting that you will recognize the validity of the points I make, regardless of how you feel about the player I claim as my favorite. So hopefully it means something to you that, as an avid supporter of Kobe Bryant — who excels at both showmanship and prime time performances — I'm here to tell you that neither of these things should factor into the MVP discussion at all.

Not even just a little.

As always, thanks to those of you who have read all 3,481 words of this post. This can be a lengthy endeavor at times, but my hope is that, by the time the season starts, we will have addressed all of the criteria that should factor into MVP consideration, enabling us to make the best possible evaluations once the season starts.

Cheers, and here's to those players whose individual dominance rewards our dedication to this team sport.

* LeBron James fans, please don't take this as a dig at LBJ — I'm a big Kobe Bryant fan, and a year or two ago, this would have read, "Cheer up, Kobe...."

Add to Technorati Favorites