John Hollinger and the perils of fatally flawed analysis

Let me clear this up off the bat: I am a big Hollinger fan (for the most part). That said, I have had some more serious problems with his work as he has upped his column load- it appears that he seems to be writing more “You see this and think _____, but that’s not right”

Today brought another one of those columns, with the subject being the Miami Heat’s potential addition of Carlos Boozer or Lamar Odom. John proceeds to go through why those players would not be substantial upgrades for the team. However, he either deliberately or accidentally (I’m not one to state motive) falls in the trap of misusing his data to make his point.

In the piece, he compares both men’s production to Michael Beasley, the #2 pick in last year’s draft. Ignoring the fact that Hollinger’s PER essentially ignores defense, he relies on it to show that there is no meaningful difference between Boozer and Beasley and that the Beasley actually outperformed Odom. Unfortunately, Hollinger commits one of the fatal flaws while crafting his argument: false choice. He chooses to create a situation where the minutes that Boozer and Odom are taking come from Beasley when the fact of the matter is that they would come from other players. Not surprisingly, the players who Boozer and/or Odom would actually be taking the minutes from were far less productive than the three players analyzed for the piece.

Assuming that the only two positions in play are Small Forward and Power Forward (even though Boozer would inevitably play some C considering Miami’s depth there is deplorable), there still are 96 minutes to divvy up to different players. The players who played those minutes this year and finished the season with the team were:
Michael Beasley (24.8 MPG @ 17.28 PER)
Udonis Haslem (34.1 MPG @ 13.10 PER)
Jamario Moon (25.9 MPG @ 13.35 PER)
James Jones (15.8 MPG @ 8.43 PER)
Yakhouba Diawara (13.5 MPG @ 6.4 PER)
Dorrell Wright (12.2 MPG @ 3.77 PER)

Those are the players that Odom and Boozer are replacing in real life. In fact, since Beasley absolutely would not be a part of a Boozer trade, it would be more logical to consider his minutes more “locked in” than the others, especially coupled with the fact that he was also the most productive of the group and has the most long-term potential.

Even if you give Beasley the 31 MPG he played in April of last season, there are still 65 minutes to be had for other SF/PF players, and Lamar Odom and Carlos Boozer are dramatically more productive than the rest of the Heat roster, particularly those still around since Jamario Moon left for the Cavs.

While all of the analysis is flawed, the most jarring (with the exception of his idea that Boston would trade Rajon Rondo for Michael Beasley with KG, Sheed, and Perkins on roster) is Hollinger’s piece towards Lamar Odom. Despite Hollinger’s classification of Odom as a PF, he would play a vast majority of his minutes at SF for Miami, a fact that John mentions. Looking objectively at where Lamar’s Minutes would come from, the most likely culprits are out-of-town Jamario Moon and Dorell Wright. Incidentally, Lamar’s typical 38 minutes per game (I’m understating his 5-year average to be nice to Hollinger and to account for his slightly older age) exactly connects with taking these minutes away from those two players.
Since I value showing my work:
Lamar Odom (38 MPG @ 16.6 PER)
Jamario Moon (25.9 MPG @ 13.35 PER) + Dorrell Wright (12.2 MPG @ 3.77 PER)

Balancing out the PER’s for those two players for the minutes they played gives a PER of 10.28 . The upgrade from their combined 10.28 PER to 16.6 is the equivalent of going from Chris Bosh to Jason Maxiell. If you switch Dorrell Wright to Diawara (more favorable to Hollinger), the difference is 5.63, the margin between Kevin Durant and Grant Hill.

For Boozer, the best man to use for the comparison is Udonis Haslem, since any trade for Boozer would almost definitely have to include him and they play the same position. Of course, it is worth noting that giving Boozer Haslem’s minutes would also necessitate a move of Boozer to C or Beasley to SF for part of the time (31 MPG +  > the 48 MPG at Power Forward), so feel free to discount the #’s slightly. The difference in PER from Boozer to Haslem is 4.18, which is less than the margin between Nene and Andrea Bargnani or David West and Nick Collison.

When you use the players who would actually lose minutes instead of the steady guy who will get his minutes regardless of what personnel changes Miami makes, you see the true value in pursuing Boozer and Odom.

1 Comment

Filed under Danny, NBA Basketball, Sports

One response to “John Hollinger and the perils of fatally flawed analysis

  1. Pingback: Daily Links 7/25 | The Epic Sports Wagering Broadcast

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s