Science and The Bible

Credit where credit’s due: Simmons couldn’t have been more right when he declared that this years MVP candidates form the deepest pool in league history. Of his top five frontrunners – Nash, Wade, Dirk, Bron, and Kobe –, each marshals a claim that is utterly rare and singular. And yet despite (or perhaps because of) this fact, all share a legitimacy that is alarmingly uniform, prompting every Page2 pundit this side of Jericho to anoint their various victors with absolute certainty. Depending on who you listen to, the MVP is either definitely Nash, definitely Wade, definitely Lebron or definitely Kobe. But while the pool of legitimate contenders may indeed be deep, the pool of legitimate arguments is far less so, and as is often the case with the thorniest multiple-choice puzzles, a process of elimination may be the shortest road to grace. With this we present, in descending order of obviousness, the three worst arguments for this year’s Most Valuable Player award:

(1) The Al Gore
“He won it last time, he should win it this time”
Player: Nash
Pundit: Chris Broussard (see also: Marc Stein, Tim Legler)

For the last three months, Nash has enjoyed the top slot on the majority of MVP lists. 52 wins, brand new roster, no Amare – the talking points are well-rehearsed. Yet strip away this pomp and pageantry and the conventional wisdom of virtually every pro-Nash partisan ultimately boils down to following inferential foxtrot:

He deserved/won it last season
He is playing even better this season
Ergo, he deserves/should win it this season



However elegant, this nifty little syllogism obscures one exceedingly simple flaw - so simple, in fact, we’re stunned to be the first to directly call it out. Simply put, this year’s MVP race IS NOT THE SAME AS last year’s MVP race. Last year, the only other contender was Shaq. This year, its half the league. The fact that Nash’s 04-05 season was the better than Shaq’s 04-05 season doesn’t necessarily imply that Nash’s 05-06 season – however much improved – is better than Lebron, or Wade, or Kobe’s 05-06 season. In short: Not all seasons are created equal.

(Note: what this defeats is not an argument for giving it to Nash, but the argument against NOT giving it to Nash, i.e. “how can Nash not get the MVP - he’s playing better than when he won it last year?”. But to the extant that every positive pro-Nash argument seems to rely on this negative one, we may safely conclude that as goes The Gore, so to goes the candidate.)



(2)The Bob Graham
“Great Stats + Viable Team = Winner”
Player: Dwyane Wade
Pundit: John Hollinger

Though recent events may have slowed his momentum, Wade has held his place alongside Nash at the very front of the MVP field. Back in mid-March, no less an authority than John Hollinger anointed him the hands-down MVP with the following, seemingly bullet-proof line of reasoning: Wade is the only candidate putting up prolific numbers (27-7-6) on a contending team. Of course, Bron’s Cavs have since called into question the second postulate, while Wade’s slump from a 1st to 4th PER rating has all but undermined the first. One could argue that this alone renders the Hollinger approach suspect - if his conclusions are so precarious, so must be his method. But there is also a broader point to be made about the incoherence of statistical comparison itself. Wade may average more assists than his rival 2 guards, and may use his possessions more efficiently, but how much of this is a function of his superior teammates as opposed to his own divine right and providence? As mighty an offensive powerhouse as Zydrunas Ilgauskas may be, he is hardly the second-option that Shaq is. Is it merely a coincidence that Shaq’s team has finished in the top 10 in total assists for 12 of his previous 13 seasons?; or that when he left the Magic in 1996, they fell from 2nd to 23rd in that category? This isn’t to say that Wade’s passing is necessarily less impressive than Lebron or Kobe’s – just that it is incomparable. Put simply, not all statistics are created equal.



(3) The Howard Dean
“The Greater the Contributions, the Greater the Man”
Player: Kobe Bryant
Pundit: Bill Simmons

With our choices narrowed to three, Simmons presents us with a logic that comes nearest of any to coherence. Eschewing historicism and fetishism alike, he relies on this simple formula: if we replaced candidate X with a decent player at their position for the entire 05-06 season, what would be the effect on the candidate's team? For example, Simons substitutes Mike Miller for Lebron, and predicts that the Cavs would have won 27 games instead of 50. Similarly, he substitutes Jamal Crawford with Kobe, and predicts that Lakers would have won 18 games instead of 45. Simmons doesn’t propose a substitution for Dirk, but lets assume that had he been replaced by, say, Zach Randolph, the Mavs would have won no more than 42 games this season. With this formula, Simmons believes he can measure – and thus, rank – the quantitative impact that each MVP candidate has had on their respective teams.

Dallas with Dirk = 61 wins
Dallas with Randolph = 42wins
Dirk’s Value = +19 wins

Cleveland with Lebron = 50 wins
Cleveland with Miller = 27 wins
Lebron’s Value = +23 wins

Lakers with Kobe = 45 wins
Lakers with Jamal Crawford = 18 wins
Kobe’s Value = +27 wins

According to Simmons, Kobe’s +27 wins makes him the #1 pick for MVP. Lebron is the #2 pick with +23 wins, while Dirk is #3 with +19 wins. Yet to reach this conclusion, Simmons must assume that all “wins” are completely alike. This clearly isn't the case: a win against Detroit is qualitatively different than a win against Charlotte. Similarly, it is easier to improve from 20 wins to 40 wins than it is to improve from 40 to 60: the first improvement entails beating teams like Houston and Philly; the second improvement entails beating teams like Denver and Phoenix. Ceteris paribus, the closer a team gets to 82-0, the more difficult each additional win becomes.

When evaluating the impact of MVP candidates on their teams, the question must not only be “how many wins do they add?”, but also “what kind?”. In an effort to make Simmons’ model more sensitive to the qualitative differences between wins, I’ve compared the 05-06 Mavs, Cavs and Lakers with those teams that best approximate their (predicted) MVP-less records: Utah, Toronto, and Portland. For each pair of teams (i.e. Dallas/Utah), I’ve categorized their regular season wins according to the quality of the opponent. Thus, of the 19 wins separating Dallas and Utah, exactly how many are against “Top Teams”, how many are against “Average Teams”, and how many are against “Poor Teams”. In this way, we can identity the qualitative distribution of the total wins each player adds to their teams.



In the table above, we see how the quantity of wins added by each candidate varies in terms of the quality of opponent. Of the +19 wins Dirk adds to the Mavs, the majority (10) are against Top Teams. In contrast, the majority of Lebron’s +23 wins are against Average teams, while the majority of Kobe’s +27 wins are against Average/Poor teams. The question of “whose wins are more valuable?” will have a different answer depending on how we value each type. If the values are something like (Top = 2)(Average = 1.5)(Poor = 1), Kobe’s +27 will still make him the #1 pick, followed by Lebron at #2 and Dirk at #3. If the values are more like (Top = 5) (Average = 3)(Poor = 1), then the order will be reversed: Dirk #1, Lebron #2, Kobe #3. In short, once we recognize that not all wins are created equal, Simmons model looses its objective determinacy, and the MVP race is once again up for grabs.



Although Simmons’ model gives no clear-cut answers, personally, I believe that Dirk’s +19 and Lebron’s +23 wins are ultimately more meaningful than Kobe’s +27. (I also think that the latter figure is exaggerated: would a similarly-coached Laker team of Crawford Odom and Parker really win 3 fewer games than this year’s Blazers? would they win 9 fewer games than Mike Miller’s Cavs?). So for me, the choice comes down to Dirk and Bron. On the basis of last month's performances, it isn't even close - Lebron wins hands-down. But insofar the award considers the entire season, I think the choice of Dirk is fairly clear. No one else has played more consistently fearsome, more consistently champion-like, or more consistently Dirk than Dirk. Even when his teammates have struggled with injury, he has continued to bring home wins, going 13-7 without Howard, 19-7 without Stackhouse, and 20-1 without Daniels. And if expert opinion has ignored his candidacy, this is only because its arguments – not the man himself – are lacking. So act quick and spread the word: Dirk is this year’s MVP.