Basketball on Paper by Dean Oliver

[MyTurn2Raise]

saw the debate on top 30 finals teams of all time and it reminded me that I read Basketball On Paper by Dean Oliver during some of my vacation time last week. Amazon review

Oliver shows how he and others are trying to transfer the stat revolution in baseball to basketball. Of course, this is exponentially more difficult as basketball is a team game with an even more difficult to measure defensive component verse the essentially individual confrontations in baseball.

The book is actually about 3-4 years old and I'm amazed it took me to know to read it.

It was a fantastic read. One sees a thorough explanation of the points per possession (for those that got into KenPom) metrics. The work did a good job at pointing out what causes teams to win. It also weighs in on the best (and worst) teams of the last 30 years. I enjoyed using objective numbers like standard deviations above the avg for points per possession, etc. It makes theoretical sense and boils down to a solid comparison. The author did more than present tables. He also evaluated meanings. He looked for common factors of the best (and worst) teams. The book ends up being a riveting historical look at the last 30 years of the NBA. However, I do wish Oliver would've adjusted the ratings based on schedule as KenPom now does for his college rankings.

From the team, Oliver than tries to dive into the ultimate unachievable quest--individual player ranks. Some great work down here that highlight Jordan, Bird, Magic, etc. I even learned some about Cynthia Cooper, though I'm not proud of it. In addition to the greats, Oliver looked at the specialists (Rodman, Kerr, etc) and the 'freaks' (Bogues, Bol, etc). Oliver was quite forthcoming with the different measures used to evaluate individuals, their strengths, and their flaws. I commend him on his honesty, however I wish he used more rigor in determining the individual weights he used for stats. For instance, he ridicules some sports economic papers that derived weights of offensive rebounds being worth 4 times a made basket. Granted, that doesn't pass the laugh test. But, I wish Oliver used the same type of mathematically rigorous methods for finding his weights. Perhaps, he could've used some bounds and LaGrangian optimizers or something. His work is still quite amazing. The contributed points per game and win shares work is fantastic.
While I'm upset that Derek Harper didn't come off better through the analysis, Oliver's book is exceptional. He brings a great mathematical improvement to traditional stats. He allows one to compare across time. He tells the stories of great teams and players as their numbers highlight. He applies new numerical methods to a historical survey of the last 30 years of the NBA. It's Fantastic!

In summary, this book is the Baseball Between the Numbers equivalent for basketball.


PS--it's funny reading the chapter on "Should I firebomb Billy Donovan's House?" 3 years later!