Hoops Lab: Hoops Lab-Week 22

Statistical Revolution

Last week an event called the MIT Sloan Sports Analytics Conference was held on the campus of MIT, one of the most prestigious schools in the country. Its purpose was to discuss the increasing role of analytics (read: stats) in the sports industry. Although the conference was for many sports, one of the main branches of the conference talked about the growing number of advanced stats and techniques being developed to gain a better understanding about basketball.

I would have absolutely loved to be there.

As those of you know that have read my column regularly, I am heavily in favor of these types of advanced techniques. It has gotten to the point that it's hard for me to even have a basketball argument with people that aren't well versed in advanced stats, because it seems like they're missing so much of the game. I know these stats aren't perfect, especially when taken as individual measures, because each one has its own flaws. But since there are so many different stats now that take so many different directions for breaking down the game, it's getting harder and harder for the "just watch the game!" guy to defend his case.

Just recently I had a long, drawn out argument with someone about just how good Jason Terry and Josh Howard were as supporting players on the 2006 runner-up Mavs versus Paul Pierce and Ray Allen on the 2008 championship Celtics. The argument touched off because one side said that of course Pierce and Allen were better, because as future Hall of Famers it was just self evident that they were better. The other side, though, pointed out that there was no way you could look at their box score stats from their championship runs and say for sure that one pair was much better than another. In fact, let's play that game: see if you can tell me which is which:

Pair A: Averaged 17.8 points on 45% FG, 5.2 rebounds, 2.6 assists, 1.1 steals, .4 blocks

Pair B: Averaged 17.7 points on 43% FG, 4.4 rebounds, 3.7 assists, 1.0 steals , .3 blocks

I'll give you a penny if you can look at those box scores and tell me which was Terry/Howard and which was Pierce/Allen without looking it up. But the perception is that one pair was so much better than the other that the difference should be obvious. Anyway, the discussion went back and forth with both me and my opponent pulling out all types of advanced stats to support our points and really break down the difference between the two. The end result was that we each agreed that Pierce and Allen performed better, but it was a lot closer than people would think, and the regular box score stats weren't enough to illustrate the difference.

And in the end, that's the point of all of these new advanced stats: to be able to show and prove elements of the game that just aren't well captured with regular box scores. To me, the main thing that box score stats are good for these days is playing fantasy basketball. And for that reason, I still love looking at things like points, rebounds and assists. But if you really want to talk about what characterizes good, productive, winning basketball, and all you have in your arsenal are some fantasy stats, then these days we really don't even have much to talk about.

Situations to Watch and quick hits