[SabermetricResearch] Basketball statistics are too flawed to work

People who think stats is "worthless" because of all the noise inherent to statistical data forget that eyeball watching has even more noise than stats, even small sampled stats. Just look at the GARM and you can see the results of eyeball evaluation after any game. The eyes have the tendency of being attracted to some things and neglecting some other things. I don't know why they think it is somehow more accurate than statistical analysis.

 

I noticed this sentiment from several posters besides yourself. The thing is, those things do show up in statistical analysis. In short, if a well-defined, observable event occurs, it can be tracked for use in statistical analysis.

 

Morey talks about this in one of his interviews. It might be in the No Stats All Star article from the NY Times linked above.

 

I believe you really need a combined approach - statistical and traditional scouting from watching the actual thing. Stats don't tell everything, i.e. LeBron is a whole lot more than what his stats represent. A lot of the times the player who "made" the play successful isn't the one shooting or the ball handler..how the player screens off other people, box out, drew defenders away, etc, all needs to be caught by the naked eye.

But stats can also be very useful. Morey's done that to perfection with Brooks/Landry etc. Numbers don't lie.

 

Now Brook Lopez has a good reason why he only got one rebound. He didn't want to show up his teammates.