Bucks Diary

Monday, April 07, 2008

Bucks Diary weighs in on the MVP race


The latest outstanding post on the WoW Journal concerns this season's MVP race, evaluated from a "Win Score/ Wins Produced" perspective. As those of you who read this blog know, I am a big proponent of the value of this method. So much so, in fact, that I made an unsanctioned addition of my own to the method by considering each player's opponents Win Score, which I refer to as that player's "Defensive" Win Score. For those of you who are interested, I explain below how I came up with the idea for this addition, how I figured out the value of defense to wins produced, and how it works on a team-by-team basis, and you can make your own judgments about its validity. For now, let me reveal my results.


Discussion

I guess the interesting aspects of my "Shadow M2P" results are the points where I agree with and diverge from Professor Berri, so I will focus my discussion there.

1. Dwight Howard is the man.
In both evaluations, Dwight Howard comes out on top. The big man is having a whale of a season. He simultaneously stands among the top offensive and defensive centers in the sport. There is no question Orlando would be where the Bucks are right now without Dwight Howard. He is the best basketball player in the NBA on every level I consider significant. He has the highest "Combined Win Score" and the highest "Wins Produced". He holds the same distinctions in Professor Berri's normal Wins Produced calculations as well. You can't do much better than he has done.

2. I like Kobe more.
If I am reading the WoW Journal correctly, Professor Berri views Kobe Bryant as a slightly overrated product of the mainstream media's overvaluation of scoring... essentially the role originally played in the "Wages of Wins" by Allen Iverson. That could be, but according to my calculations, he may be doing a bit more on the defensive end than is generally acknowledged. I can't say for certain that he's a defensive wizard, but I can say this: the counterpart shooting guards the Association matches against him have a Win Score average that is way below the overall average for that position. And if that isn't "defense" I'm not sure what is. And its benefiting the Lakers, so I have to take a higher view (albeit not as high a view as Kobe has of himself).

3. KG has made the most of his time
If Kevin Garnett had played an equal amount of time as Dwight Howard, the two would be neck-in-neck for top Win Producer. And, if you consider the impact KG's presence has had on the improved defensive performances of so many of his teammates, you might "spot" him the difference and award the MVP to the Big Ticket.

4. We disagree on Camby
The one point where there is clear blue water between the Professor's results and mine is the evaluation of Marcus Camby. We both agree Camby is an outstanding "Win Score" producer, mainly because of his voracious rebounding and his efficient use of offensive possessions. But, somewhat shockingly, I think he's a very poor defender who is giving a lot of his statistics back on defense. I think he drifts around too much on that end of the court, and consequently allows his "covers" too much freedom to produce. And I think its hurting his team. I can hear the counterargument: What about all his blocked shots and all his alleged "alterations"? I can only say this: while he may be helping suppress his teammates "Defensive Win Scores", and that is by no means clear from the numbers, what is clear is that the other team's centers have been let free to inflict damage on the Nuggets. Now, I'm not saying Camby isn't a huge asset to the Nuggets... I'm saying he's taking some of the luster off his statistical contributions by permitting his counterparts' to do too much of the same.

5. Chris Paul is a freak
What Chris Paul is doing at 5'11 and, I don't know, 165 pounds soaking wet, is beyond comprehension. His defense is not as good as the other candidates save for Camby, but his Offensive Win Score is the best of all. He's a marvel.

I'll make more points on this whole thing in the coming days. For now, the background "foundational" information I promised is next.

Backstory: How I stumbled upon Defensive Win Score

If something doesn't make logical sense to me, it will bug the shit out of me until I resolve it. Essentially, that's how I came up with Defensive Win Score.

Here's specifically what happened. I accidentally discovered a page on draftexpress.com that keeps track of every team's cumulative "Win Score / Game" average. I looked at it and noticed Phoenix had the best average by far. This made no sense to me. If they have the highest per game average, that means they are the most "above average" team from a Win Score perspective, and it logically follows that they should have the most wins, but they don't have the most wins.

After all, if you can explain with a high degree of certainty each team's Wins Produced by ascertaining each individual player's Win Score average and then adjusting it by the average production at that player's position, then you ought to be able to do the same thing taking the entire team's average and simply adjusting that against the entire NBA's team average. What I mean is, if you are accounting for every single player minute -- which Win Score/Game does -- and you are comparing it against the NBA per game average, you are in essence "adjusting for the positional average" of every player on your team in one fell swoop. So you should get the same results on the aggregate level that you get on the "cumulative individual" level. But you don't.

For example, at the moment Phoenix's "Win Score/Game" average is a staggering 56.0. That is 13.7 aggregate Win Score points above the NBA aggregate average of 42.3. If you divide that by 5 (so that you are, in essence, considering the entire team as one player), you have a collective Phoenix Suns player who is +2.74 Win Score points above average. Theoretically then, you ought to be able to take that figure, put it through the proper "Wages of Wins" formula, multiply that result times the sum total of all Phoenix Suns player minutes to date, and come up with something close to the team's aggregate win total. But you can't. It doesn't work. Phoenix is so far above average it should be undefeated. It took me a couple of days to figure out how this could be so.

I came to the conclusion that it must have something to do with the productivity of their opponents... in other words, defense. So I looked up each team's "Opposition" statistics and used them to figure out each team's "Opponent Win Score/Game" average, and once I did that, I did the same calculation as I did in the preceding paragraph. Doing so, I quickly discovered that each team's actual win total was located almost exactly in the middle of the two results.

Take for example the Boston Celtics. Their "Offensive" Win Score is 46.4, which translates into a record of 46.6-29.4. Their "Defensive" Win Score is a mind boggling 29.8, which translates into a record of 72.8-3.2. Together, their predicted record by Combined Win Score is 59.7-16.3. Their actual record is 61-15, so its pretty close. I found similar results throughout the Association.

As a result of my "50/50" findings, I concluded each team's straight "Win Score/Game" average functionally reflected their offense's contribution to victory (thus "Offensive Win Score") and each team's "Opponent's Win Score/Game" average functionally reflected their defense's contribution to victory (thus "Defensive Win Score"). That's how that all came about.

Later, after I found out about 82games.com's "Counterpart Production" information for every player, I used that to "individualize" Defensive Win Score. Then I used the same "half and half" method from above to determine each player's "Combined Wins Produced".

A few more general points

1. Regular Win Score is still the more reliable method of evaluating personnel. Defensive Win Score varies too much from year to year to have any great use for predicting how a player will do in the future. Regular Win Score, by contrast, is remarkably consistent. In simple terms what I mean is "Combined Win Score" is more like a player's report card, whereas Regular Win Score is more like a player's standardized test score... if you understand my analogy.

2. I think the variance in Defensive Win Score performances is due to the nature of defense in the sport of basketball. Basketball, unlike most every other sport, affords players the opportunity to relax on defense without any great social sanction. In baseball, if you relax on defense, you commit errors for the world to see. In football, players are specifically assigned to defense, so relaxation could cost a player his livelihood. But in basketball, you can often drift around and conserve energy on defense, without necessarily incurring individual sanction (especially since basketball, unlike other sports, keeps no specific individual defensive statistics). Thus when players are stuck in hopeless situations, as Kevin Garnett and Paul Pierce were last season, human nature is to relax a bit. Conversely, when you put them on a good team, as the two are on this year, you shouldn't be surprised to get a much better result.

0 Comments:

Post a Comment

<< Home