I think his statistics are overall pretty solid at predicting what a team is likely to do in the regular season. He uses smart, solid indicators - point differential weighed more than record; strength of schedule is factored in since you expect, on the whole, that schedules will balance out over the course of a season, so if you played a tough schedule it should lighten up and vice versa; paying attention to home games v. road games since teams typically are better at home than on the road and that will balance out by the end of the year; giving greater emphasis on what a team is doing lately than what it did two months ago. They're good indicators for how good a team has actually been this year, and what they are likely to do over the remainder of the regular season.
There are obvious flaws - I agree, there's no way Hollinger can reasonably think Philly is the fourth best team in the NBA (is there?). And a big point differential outlier can have a significant but erroneous impact. For example, both Boston and New York played Portland at home in mid-March - the Celtics led by 35 at the half and 34 after 3. The end of the Celtics bench was outscored by Portland's by 16 in the fourth quarter and Boston got a plus-18 margin of victory. The Knicks led by 26 at the half and 22 after 3. Their bench beat up on the Blazers bench and New York got a plus-42 margin of victory. Neither victory was more impressive than the other, but the Knicks get an extra 24 point margin - at this point, that's a half a point difference for the season point differential, and a 2 point difference for the last quarter of the season and would have a significant impact on Hollinger's rankings.
But overall, I think they're reasonable. And I think there was reason for an impartial NBA fan to question the Celtics. Until that Milwaukee game, the Bucks were hot, the Knicks were hot, the Celtics were lukewarm, all three were tight in the standings, but the Celtics had by far the toughest stretch of schedule remaining. His model had the three finishing with the same record, Boston losing out on tiebreakers. (And importantly, it wasn't just his numbers. He was asked point blank in his chat how he specifically thought the season would play out, and he said Milwaukee 7, New York 8, Boston 9. That's not unreasonable for an impartial fan to think.)
I do think his statistical models suffer once you get to the postseason. At that point, the game changes (a more grind it out style works better than a free wheeling style as defenses buckle down), the teams change (benches are shortened), and circumstances change (no back-to-backs, game planning for only one team for two weeks instead of adapting to new teams on a nightly basis). And the intangible quality of "knowing how to win" becomes somewhat more important than the ability to just overpower as there are closer games typically - this poses a huge problem for Philly specifically, which has more 20 point wins than any other team in the NBA, but has not won a game within five points.
Topic: Hollinger Wrong - Cs won't be in lottery (Read 9633 times)