# Missing the Obvious on KG

Who is the best player in the NBA? The search for the best typically begins with the best team. This year, the best team is the Boston Celtics. And the best player on that team is Kevin Garnett. So should KG be MVP?

Sportswriters – the people charged with choosing the MVP – appear confused by KG’s candidacy. Certainly they see that Boston is on the verge of the biggest turn-around in NBA history. And just as certainly, people understand that it’s Garnett who is the most responsible for Boston’s ascension to the top of the league.

Given these facts, why does it appear that Garnett is lagging in the MVP polls? The problem is in the numbers. KG is posting the following per-game averages:

18.9 points

9.4 rebounds

3.5 assists

1.4 steals

1.3 blocked shots

Certainly these are “nice” numbers. But how can the MVP in the league fail to average at least 20 points per game, or double figures in rebounds or assists? After all, this same player had the following career averages heading into this season:

20. 5 points per game

11.4 rebounds

4.5 assists

1.4 steals

1.7 blocked shots

Garnett has been at least a 20-10 guy in every season since 1998-99. Since his numbers have “clearly” declined this season, how can anyone vote KG as MVP in an apparent down year for him?

**The Playing Time Illusion**

Last year Eddy Curry was considered a much improved player. As I noted, though, all that had really change was Curry’s minutes. On a per-minute basis he was actually a worse player.

The playing-time illusion is not just confined to Curry. Mikki Moore was able to cash in on this illusion with the Sacramento Kings last summer.

Garnett’s story is characterized by the same illusion. But for him, this illusion is not providing any benefits.

For his career Garnett has averaged 38.3 minutes per game. This season his minutes per game have fallen to 33.5. As a result, Garnett’s per game averages in a number of stats have also fallen.

What would KG’s numbers look like, though, if he went back to playing 38 minutes a game? The answer is in Table One.

From Table One we see that KG would have the following per-game averages if his minutes per game had not declined.

21.7 points

10.8 rebounds

4.0 assists

1.6 steals

1.5 blocked shots

Would these numbers help KG’s candidacy? I think if KG was once again a 20-10 player he would be the favorite for the MVP award. But because sportswriters are missing the obvious — in other words, per-minute numbers are a better barometer than per-game stats — Garnett has become the John Edwards (or Mitt Romney depending on your persuasion) of the 2008 NBA MVP vote.

**Has Garnett Declined?**

Okay, KG looks better if he played more minutes. But exactly how does his performance – on a per-minute basis – compare to his past?

For an answer we turn to Table Two.

**Table Two: The Boston Celtics after 75 Games**

Table Two offers two projections of the Celtics. The first looks at how many wins we could expect if each of the Celtics (except the rookies) played as well as they did last year. The second looks at how each player has played this year.

Judging by last year’s numbers, the Celtics should have expected to win 54 of their first 75 games (or 59 for the season). This year the team has done even better. Although many players have improved a bit, about half of the leap forward can be tied to the play of Kendrick Perkins. Perkins only posted a 0.042 WP48 last season. In 2005-06, though, his mark was 0.156. Had we known Perkins was going to return to his form from two years ago we would have forecasted 64 wins for Boston this year.

Like Perkins, KG has also improved. Yes, although his per-game numbers are lower, his per minute – or per 48 minute – production has increased slightly. Last year KG posted a 0.330 WP48. This season, primarily due to improvements in his shooting efficiency, Garnett is posting a 0.361 WP48. Such a mark leads the Celtics. So Garnett is indeed the most productive player on the best team.

**KG for M2P?**

But is Garnett the Most Productive Player (M2P)? His lack of minutes is going to hurt his wins production. If we focus on WP48, though, we see ….. okay, he is not the leader there either.

No, I don’t think KG is the most productive player in the league. I will say he is more productive than Kobe Byrant (just had to mention that). And Garnett is more productive than all but two or three other players (and maybe next week I will talk about the leaders). The MVP and M2P awards, though, can only go to one player. So if one player offers more than you, then you are probably not the best choice for the M2P award (of course MVP is not defined so it is hard to say if KG is not really the best choice).

**A Summary**

To summarize, if the MVP and the M2P are the same player, I don’t think KG should be MVP this year. I do think, though, that he should be a significant part of the MVP conversation. But the obsession with per-game stats is stealing some of Garnett’s thunder.

Let me close by noting that I don’t think that people should solely look at WP48 or just per-minute stats. If you did that, Jerome James – who posted a 1.341 WP48 – would have been the first half MVP. James, though, only played five minutes in the first half of the season, so his WP48 doesn’t really mean much.

Although I do think people need to look at more than per-minute numbers, I also think people need to stop focusing solely on the per-game stats. Specifically, when we are looking at players who played at least 30 minutes a contest, we shouldn’t penalize players whose minutes are closer to 30 than to 40. Such penalties — as we see in the case of KG — can easily cause us to miss the obvious.

– DJ

Our research on the NBA was summarized HERE.

The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score

Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:

Simple Models of Player Performance

What Wins Produced Says and What It Does Not Say

Introducing PAWSmin — and a Defense of Box Score Statistics

Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.