I’m not sure, but calling someone a “stat-head” seems pretty self-important. Berri didn’t have to pound his chest after cracking the NBA box score with nuclear physics. Dean Oliver had already done it with basic math.

re: Hollinger & Rosenbaum – no.

I find the analysis pretty retrospective.

1. I think strength of competition in the league has generally improved over the past couple years and especially in the Eastern Conference.

2. Even throwing out last year’s monstrosity, it’s pretty debatable whether the Bulls had or have the room for steady improvement (meaning, they relied on Wallace and PJ Brown a lot a couple years back, and their other players/additions didn’t improve enough to offset the decline and departure of those guys.

in seeing how that Detroit series looks on paper. Not to mention closing the regular season off losing that very much needed game against the Nets. Our last “good” season is a hard one to figure out for me, as hard as our terrible year. We mad a second half push for the playoffs, but the majority of teams we played that that second half were below 50 teams. That’s just how the schedule fell and how we looked so “strong” going into the playoffs. Then we played the weakened Miami in the first round and killed them… Perhaps that partially added to the predictions we’d do even better the next year and end up in the eastern conference finals. I’d suspect that statistically we’d look good against that last string of below 50 teams and good against the weak Miami, but the complete unraveling in the Detroit series seems to have been ignored in the predictions of most people.

And we often lose sight of that. A statistic is a recording of a past event. It describes something that happened. It can only determine the future insofar as what happened in the past is reflective of what happens in the future.

The other thing people lose site of is just how endogenous the variables are we’re dealing with. The output of a basketball team is winning by scoring more points than your opponent, and you’re only one of about 10 guys that take court for your team. Getting a definitive statistical measure from that is simply not possible, especially when the different guys on the court have different roles. It’s a bit like asking which is more responsible for cutting the grass, the lawnmower or the guy pushing it. You can’t separate the two very easily.

Using advanced statistics attempts to separate them out, but you basically run into a few problems:
1. Not enough data points until you’re well on in a player’s career. Since definitively finding out about a guy doesn’t matter much after ten seasons, what’s the point?
2. Not enough situations to really see what’s going on. Like, there’s no way to really statistically say Jamal Crawford sucks if you acknowledge the fact he had to play next to guys like Ronald Dupree and Chris Jeffries. Or to separate how Deng and Gordon play with a good interior defense vs. how they play with a crappy one.

All this is not to say that the stats aren’t worth looking at, but in my opinion they can only serve as the starting point for discussion. You look at the numbers and then look at the obvious non-statistical factors. When making a prediction, I try to look at a player’s statistical output using various measures, then factor in:

1. Age. Is he getting better or worse? I was pushing to trade Wallace by halfway through his first season here because it seemed obvious he was falling off quickly, although the stats didn’t reflect this tremendously until this past season. The trends in the stats, however, reflected it even in his first year.

2. Teammates. What quality guys is the subject playing with, and what roles do they fill. Is the player being asked to do more or less than he should in ways that are likely to show up in his stats?

3. Competition. Who did the subject accumulate his stats against, and how does he rate out compared to what other, outwardly similar players generate?

Which isn’t much of an explanation.