LeBron James, Dwight Howard, Derrick Rose, MVPs: Objectivity vs. Subjectivity

Kelly ScalettaFeatured ColumnistApril 2, 2011

CHICAGO, IL - MARCH 25: Derrick Rose #1 of the Chicago Bulls drives against the Memphis Grizzlies at the United Center on March 25, 2011 in Chicago, Illinois. The Bulls defeated the Grizzlies 99-96. NOTE TO USER: User expressly acknowledges and agrees that, by downloading and/or using this photograph, User is consenting to the terms and conditions of the Getty Images License Agreement. (Photo by Jonathan Daniel/Getty Images)
Jonathan Daniel/Getty Images

The blogosphere and their champions like John Hollinger point to "objectivity" and stats like PER and Win Shares and the like and argue that there isn't a single "objective measure" to vindicate that a Derrick Rose MVP run is a viable one. 

Now truth be told, if you're looking at Win Shares it's hard to make a case for Rose. He's sixth in Win Shares and 10th in Win Shares per 48 minutes. He's only 13th in PER. Dwight Howard and LeBron James are far ahead of Rose in both categories. This is viewed as "objective" proof that they are "better" players and more "valuable" than Rose. 

I mean how can you argue with that? You look at the numbers and the number next to Howard is bigger than the number next to Rose. The number next to James is bigger than the number next to Rose. It really just comes down to the simple fact that their number is bigger than Rose's number and that proves the case, objectively. 

Except for one thing—that number isn't as objective as you might think. It has the benefit of being a number, which makes it appear objective, but that doesn't mean it is.

In obtaining those numbers they combine different numbers. Those numbers are derived from other numbers like assists and rebounds and points, so that gives an appearance of objectivity too, but the fact is that what it is a subjective argument disguised as a number.

Why? Because the values of those different numbers are assigned according to subjective valuations of what is or isn't valuable, or what has more value than other things.

What's the value of an offensive rebound versus a defensive rebound?

What's the value of an assisted field goal or an unassisted field goal? Are they the same?

For that matter are all unassisted field goals the same value? If a player drives through traffic to get a dunk, is that the same as a player who gets a tip-in when they're all alone under the basket?

Are all assisted field goals the same value? If a player makes a cut to get away from his defender and create space for himself, when he gets the ball and stops and pops, is that the same as a player that has a pick set for him who catches a pass and stops and pops?

Is a player that takes risks by shooting and also gets extra steals, but in turn lets his man get by him more often than he should helping his team? The reward shows up in the box score but the risk doesn't, and can mask a bad defender as a good one.

Is a missed field goal right before the buzzer at the end of a quarter from half court—a good basketball play—counted any different than a wide-open, missed three-point attempt? 

If a player sees he has a man all alone under the rim, stops and pops a three, it's a good play. If he misses it, the man under the rim gets a rebound and puts it back in for an easy two; if he makes it, it's three. All reward and no risk make it a good play. 

Part of the problem with these combinations of different stats into a pretty number don't deal with any of the above nuance. They might give you a rough picture, but sometimes that rough picture can be misleading. 

In addition to treating all "like stats" the same i.e. an assist is an assist, there's the task of assigning differing stats differing values. Since basketball is a game of "roles" it's important that no one "role" is so emphasized as to devalue the importance of other roles. 

So when we look at the leaders in Win Shares and realize that there is only one guard among the top 30 players in the league, it's easy to conclude that a problem with Win Shares is that it devalues the importance of the guard position, since roughly 40 percent of players in the NBA are guards. 

Stats have differing values and how you place that value is subjective. How much an assist is worth relative to a rebound and how much a missed field goal is worth, etc. is what advanced basketball stats are all about. But in so many ways, it has barely begun. 

People confuse things like PER or Win Shares with stats. They aren't stats. Stats are things like assists, rebounds, points and missed field goals. Those are objective things that actually happen on the court. The "tell-all" advanced calculations like PER and Win Shares are subjective valuations about how to combine all those stats into a single number. 

They aren't "objective numbers," they're subjective arguments disguised as numbers. 

No player in a game has a PER or gets a Win Share. Let's not pretend that these amalgamations of stats don't have holes either. If you want to know how flawed they can be just look at Peja Stojakovic.

Win Shares are intended to show how many "wins" a player adds to his team. He played two games in Toronto, totaling 22 minutes. His Win Shares for those games were 5.42. In two games, did he help the Raptors win nearly five-and-a-half?

I know small sample sets can skew the numbers, but when you can skew them that much there's something wrong.

There are many subjective calculations that go into things like Win Shares. They normalize minutes played (thereby reducing the value of actually playing minutes), make all scoring, assists and rebounding the same, and they weight the relative value of these things to one another.

Then there are things they either exclude deliberately or just didn't think of including. The ability to create your own shot is arguably the most important skill in basketball because it spreads the court, opens up shots for teammates and commands double-teams. That skill is not reflected in Win Shares or PER. In fact not only is it not valued, it's devalued by emphasizing the things that doing that creates for other players. 

Does that mean that PER and Win Shares are wrong? Maybe. Maybe not. It's a matter of opinion. That's where the crux of the matter is though—it's a matter of opinion. PER and Win Shares create the illusion of objectivity because they are all displayed in a pretty, shiny number, but that number is really just an opinion. It's not a direct opinion of a player, but it's an opinion of what goes into making a good player. 

No self-respecting snake-oil salesman is going to tell you what his product is; he's going to tell you what you want to hear. There's something that reasonable people have—a desire to be objective—and PER and Win Shares give us that illusion. Don't buy the snake oil. It's still an opinion.

It's funny because people will say that Hollinger is entitled to his opinion. I agree. He's entitled to his opinion. I just wish he would admit that's what it is. But then you wouldn't buy the snake oil.