[HBR Blog] Success Comes From Better Data, Not Better Analysis

I hate HAAAAAATE the concept of running a team off iPads. Thats so LAME. Geeks think they're funner than they really are. We watch sports to get AWAY from the measured life. We do watch for the "I told you so" but not at the cost of spontaneity and "treating human man like asset".

Oddly I find myself seeing the rationale behind Morey's moves more and more. Rockets were criticized for their LENGTHY coaching interview process. But after seeing the end result its hard to knock them for being diligent and detailed. I think thats a result of the quant approach. And the Rockets can actually explain their moves better that way.

The Lakers who operate traditionally had a QUICK coaching search. And Laker fans HAAAATE the Mike Brown hire that came out of nowhere. Kobe went on silent boycott from the press for 3 months after Brown was hired. Lakers fans AND players woulda MUCH preferred the way the Rockets approached it. The Lakers the best franchise in the league have to do more PR spin than the lowly Rockets.

So I guess thats what I'm getting at, quant isnt THE way. But quant = diligent if nothing else.

 

The competitive advantage of analysis is worthless in the long run without proprietary knowledge..... whoop-t-do. How about study the long term effects of drafting under-sized players or the competitive advantage of obtaining injured players or maybe how only the gym rats will truly be great. This movie sounds like it will be full of obvious facts. I could care less about pure statistical analysis he spews when it's study of human nature that really gets results on the championship level. We need a GM that does more than squeeze pennies out of people, we need a guru that understands what drives individuals to greatness. Anything else is a waste of time if you are trying to be crowned.

 

This reminds me of a presentation at a Google conference where a dude talked about getting accurate results from data.

He said that they tested several categorizations of algorithms and found that trying to create a superior algorithm was not that helpful but rather that increasing the size of data produced accurate results on a much better scale (theory and analytic approach were not as big of a deal as just scale of data).

In some instances, "inferior" algorithms actually outproduced their "superior" counterparts after increasing the data by a degree or two.