Lebron or Kobe: Is the MVP Race a Dead Heat?

[mdh88]

Lebron, with his improved game (especially his defense) and the Cavaliers winning ways, has been the MVP favorite almost from the beginning of the season with Kobe a close second. Now, with the Lakers recent 6 game East Coast win streak (without injured center Andrew Bynum) including wins over the Celtics and the Cavaliers (giving the Lakers a 2-0 season sweep against both teams), the talk is about Kobe. The players interviewed at All-Star weekend and the media now appear to be split. I would call it a dead heat, but if a gun was to my head and I had to vote today, I would vote for Lebron because I think Kobe has more talent around him. However, much of the media often just votes for the best player on the best team (which to me is lazy thinking), so if the Lakers still have the best record in the league at the end of the season, I expect Lebron to win in a very close battle. Who gets your vote? Who do you think is the better player right now? Lebron vs. Kobe -- still the best debate in sports!

[bukowski27]

i'm doing research for a graduate school econ class forecasting the nba mvp.

i'm using a binary logistic model for all past mvp vote-getters to determine which variables are important in mvp voting.

my model predicts than lebron has an 80% chance of winning.

the model has never failed when having such a high probability. it has failed in the past with Steve Nash's mvps, malone over jordan, hakeem over shaq, and julius erving over larry bird.
_________________
chuck

[mdh88]

Chuck, what are the factors used in the model? Last year I got the first 5 MVP candidates in the exact order and I had no system. This year, I predict: 1. Lebron 2. Kobe 3. D-Wade 4. D. Howard 5. Chris Paul 6. Tim Duncan 7. D. Williams 8. C. Billups 9. B. Roy 10. P. Pierce

[bukowski27]

my model found PPG, RPG, APG, team win percentage, and POINTGUARD significant. I tested for blocks, steals, race, and difference in win percentage expectation, but none of the latter were significant predictors.

As for last year, my model got the first 4 correct. I had D. Williams fifth.

The model is pretty intuitive. Obviously, all else equal, points per game and winning percentage will put a certain player over the top.

i don't know if you're into statistics and econ, but i used a binary logit model to come up with my forecast.

as i mentioned before, my model has never been incorrect at such high probabilities as Lebron's 75+%.

If he doesn't win, it will prove that my model has a major flaw.
_________________
chuck