From 2009: NBA MVP Statistical Analysis

Over a couple of beers the other night with some friends, I was discussing the apparently three-pronged race for MVP in the NBA this year. Unless you've been living in Bedrock, Mississippi the past year, you know that LeBron James and Kobe Bryant are the favorites, with Dwyane Wade making a stronger than usual dark horse case. One of my friends was appalled at how obviously this is a case of media favoritism, and that other players were more deserving but stood no chance because of little to no media exposure. Fair argument, but you can't deny the other three's performances. Another friend tried to make the argument that the MVP HAS to contribute to a winning team, otherwise they cannot be considered for the award. After drinking a few more beers and listening to their arguments, something occurred to me.


Look at the name of the award. Most VALUABLE player. Working in outside sales, I'd like to think I know the meaning of the word value. So how does that translate into basketball? It covers quite a bit of ground when you really get down to the etymology of the word, but what about the substantive pieces of the argument? That's where you can take the ball and run with it: statistics.

I took a very scientific approach while figuring out how to answer the question. So my hypothesis is, "The MVP of the league is the player who hurts his team the most if he does not exist." This means I have to compare the player in question with his backup, and compare the differences in production. At the same time, I had to make some admissions about my argument that I cannot compensate for, so allow me to give you a "disclaimer."

1) You obviously cannot find a stat for intangibles.
2) I fully realize that if Yao Ming doesn't play, the Rockets go big with Hayes and Landry, and not exclusively Mutombo. This holds true for every team.
3) I know Kobe simply being on the court changes defenses by merely taking on double teams, opening up shots for Bynum, Gasol, etc.
4) I realize that, no matter what, any statistical analysis involving imaginary extrapolation cannot be taken for the unforgiving truth.
5) I realize I'm going to sound like an idiot for even writing this.
6) Yes, I woke up and felt REALLY OCD when I wrote this.

Regardless of all of those things, let me tell you what I did, and how interesting the results turned out to be. For our first example, let's look at an analysis of Brandon Roy, and his exact position backup as listed on the depth chart, Rudy Fernandez. Again, I know the Blazers can play them both at the same time, but I picked the player who's playing time would go significantly upward were it not for the mere existence of the first stringer. Let's take a look at Roy's stat line. Brandon Roy
22.6 PPG, 4.7 RPG, 5.1 APG, 1.1 SPG, 0.48 FG%, 37.2 minsPG

And to compare, let's look at Fernandez's stat line. Rudy Fernandez
10.4 PPG, 2.7 RPG, 2.0 APG, 0.9 SPG, 0.43 FG%, 25.6 minsPG

Then I ask, "What would Rudy Fernandez's stats look like if he played the same minutes per game as Brandon Roy and maintained the same field goal percentage?" You can't statistically model pressure and clutch play, but proportions help out quite a bit. Extrapolating those stats for the same minutes as Roy gives us this:

15.1 PPG, 3.9 RPG, 2.9 APG, 1.3 SPG, 0.43 FG%, 37.2 minsPG

Certainly better statistics, but still not a match to Roy's line. Obviously this doesn't mean everything, but it has to be important to a degree, right? To make sense of all of this, I took all the differences in stats between the two (Roy's actual vs. Fernandez's altered) and took the percentage difference between the two, then I averaged all of those percentages and took the final count. On my first run, I realized that the numbers were all over the place, some saying that a player's backup actually would have done better in his starter's place.

Knowing that can't be 100% true, I numerically took into account the importance of certain statistics:

-For the percent difference in assists per game, I divided that total by 3 and used that as a modified total in the final percentage.
-For the percent difference in steals per game, I divided that total by 10 and used that as a modified total in the final percentage.

Still referring to the case of Roy vs. Fernandez, these are the percent differences I came up with:

PPG: 33.186%
RPG: 17.021%
APG: 14.379%
SPG: -1.82%

And I also took the straight percent difference between field goal percentage, obviously needing to factor that in:

FG%: 11.458%

Thanks to MS Excel, this took me a lot less time than it would've taken someone in 80's to do. This is what the final difference number looked like: [(33.186+17.021+14.379-1.82+11.458)/5] = 14.845%

I know, right? Are you tired of reading yet? Hold on, I'm just getting to the good part.

I ran this process for 11 other cases of legitimate MVP candidates, and compared all the percent differences. Holding all of the above rules to be true for each case, I present you with my final ranked list of difference percentages. The higher the percent difference, the higher the ranking.

1) Chris Paul vs. Antonio Daniels (27.049%)
2) Dwyane Wade vs. Daequan Cook (25.884%)
3) Yao Ming vs. Dikembe Mutombo (25.394%)
4) LeBron James vs. Tarence Kinsey (23.237%)
5) Kobe Bryant vs. Sasha Vujacic (20.833%)
6) Dwight Howard vs. Marcin Gortat (16.826%)
7) Tim Duncan vs. Kurt Thomas (16.312%)
8) Brandon Roy vs. Rudy Fernandez (14.845%)
9) Deron Williams vs. Brevin Knight (12.911%)
10) Dirk Nowitzki vs. Brandon Bass (9.885%)
11) Kevin Garnett vs. Leon Powe (6.827%)
12) Derrick Rose vs. Kirk Hinrich (5.963%)

I know how much Derrick Rose means to the bulls, and I know how Brandon Bass being a good bench player shouldn't diminish Dirk Nowitzki's value to the Mavericks, but I thought it was interesting how the supposed three pronged race still showed up in the top five of my list, which completely took intangibles out of the argument.

And when you think about it...if Chris Paul wasn't on the Hornets, would they even be in the playoffs? Someone has to lob up those alley-oops to Tyson Chandler if they're ever going to score...and it tickles me pink to see Yao Ming at #3 on my list.

Let the arguments begin.