Celtics are number 2 on a power ranking??????

[cfan79]

I don't get Hollingers thinking by putting the C's at number 2. We've gotten 16 in a row thus far and are 24 and 2.

http://sports.espn.go.com/nba/columns/s ... iem-081218
http://sports.espn.go.com/nba/hollinger/powerranking

[MalReyn]

It's just the way Hollinger's computers spit out the numbers. It's not his "thinking" at all, purely his statistical ranking system, the flaws and benefits that go with it.

Sagarin has his own computer rankings calculated differently that have the Celtics first:
http://www.usatoday.com/sports/sagarin/nba0809.htm

It's like the BCS computer polls. They have a half-dozen different systems that rank teams different, no human input whatsoever

[EJay33]

His rankings also say that Zydrunas Ilgauskus is the 12th best player in the NBA.

[KyleCleric]

His rankings also say that Zydrunas Ilgauskus is the 12th best player in the NBA.

[s1ickd]

The fact that he thinks that winning close gaames decided by 5 points or less, or overtime games, are a "50/50 proposition" and that the celtics have been "fortunate" to be UNDEFEATED in these type of games is RIDICULOUS.

ps. he then states the cavs are 1-2 in these types of games.

doesn't nearly every game int he playoffs get decided by 5 points or less at some point in the 4th wuarter? what a fraud.

[Dave_From_NB]

They don't hang Power Rankings in the rafters. They're kind of fun, but mean close to nothing.

[Taget]

Celts do better in winning championships than they do winning "power rankings" and pundits. Also keep in mind power rankings are like the US News ranking of colleges. Is Harvard really worse in the years they are ranked 3rd? You have to change things around every so often just to have something unique so people pay attention to it.

[GuyClinch]

I think it's fine. I do think the C's are deserving of the #1 status but they rank #2 because they don't win by a very big differential.

Now historically speaking - in general teams that win by alot win championships. And actually that differential is more important then the win loss record. (though of course win loss record is important as well.

People claim that his system is 'flawed' and of course it is. But the real problem is statistics in general are not entirely predictive. You can't just look at past performance (especially in the years past) and determine future results. In otherword no 'tweaking" of his model is going to be perfectly predictive.

It's like those equations that determine the "best" player. They don't really work becaues they don't talk about the fit between players on certain teams. No fudging of the variables is going to solve the problem.

[ryaningf]

We could post 74 wins and still be #2 in the Hollinger rankings if we don't get our point differential up. Maybe that would make him tweak his system.

[Rocky5000]

There are several problems with his system. The Cavs lost to Atlanta by low single digits last week, and we beat them by 3. But to the computer, it makes very little difference since it will only affect the point differential by a small amount.

There's also the fact that his whole purpose of this is to supposedly remove the impact of wins and losses and use point margin and the quality of the opponent to determine rankings. This is fine, if you beat the Celtics by 20 or the Knicks by 1, it should make a difference. However how he deals with the quality of the opponent, isn't by looking at win-margin,it's by Strength of Schedule, which is based on WINS and LOSSES. So which is it? Are wins and losses a good indicator of the quality of a team or not?

[floyd]

He should (and he may, I don't know) weight the point differential less as you get into blowouts. What's really the difference between a 20 point win and a 30 point win? The scrubs played better, that's what.

What would make the most sense is to have an ordinal factor, like close, not so close, and blowout. And it would be better to determine that by watching the game rather than the final score as fts distort and garbage time can really distort how close a game is.

[GuyClinch]

We could post 74 wins and still be #2 in the Hollinger rankings if we don't get our point differential up. Maybe that would make him tweak his system.

Its still not going to work. Imagine if the C's go 80-2 and win the championship but with a point differential of 5. They beat the Lakers who go 70 - 12 but with a point differential of 15. So he rebuilds his system and weights point differential far less and losses far more.

That's still not going to help when teams like Dallas lose to teams like Miami. It's not that the variables need to be fudged is that the statistics used just aren't perfectly predictive. Sometimes the teams with higher point differentials DO win the championships. But not always..

Just as the NE patriots. Almost all the statistics had them pegged to win..

[SuigintouEV]

They don't hang Power Rankings in the rafters. They're kind of fun, but mean close to nothing.

That's kind of true, but look at the pythagorean w/l of the last few championship teams...

2008 Champs - Celtics 66-16 (1st in NBA) vs 67-15 (theoretically 1st in NBA)
2007 Champs - Spurs 58-24(3rd in NBA) vs 64-18 (theoretically 1st in NBA)
2006 Champs - Heat 52-30 (5th in NBA) vs 52-30 (theoretically 5th in NBA)
2005 Champs - Spurs 59-23 (tie for 2nd in NBA) vs 63-19 (theoretically 1st in NBA)
2004 Champs - Pistons 54-28 (5th in NBA) vs 59-23 (theoretically 2nd in NBA)
2003 Champs - Spurs 60-22 (tie for 1st in NBA) vs 57-25 (theoretically 2nd in NBA, + dirk missed the last 3 games of the series that year)

I'd say, if not much, it's still meant a more than actual regular season record

The thing that does separate the celtics from most other "theoretical seconds that finished 1st" like the 07 mavs and 05 suns ??

Defense that's more than just coaching, it's players who are really damn good at it, that's what!

(Oh, and actual championship experience, poise, being really damn good, and steadily moving up from like 20th in offensive officiency to open the season to being like 6th now... having beat actual competition not the Knicks, etc)

[MyInsatiableOne]

Hollinger is a ****...in one breath he'll talk about how amazing the C's are, then he'll justify them being 2nd or 3rd in his moronic power rankings. He's trying to BCS-ize the NBA..."strength of victory"? "Quality of opponent"? If so, Cleveland wouldn't even be top ten having beaten 2-3 teams with winning records during their 11 game streak...

[GuyClinch]

^^^ Hollinger is like that. He is smart enough to know his stats don't exactly work - 100% of the time. So if confronted directly he downplays them. But at the same time he wants us to care about his lists.

He will say things like - well PER isn't meant to measure how good a player is overall. But here is my list of top 100 guys ranked by PER. It's the same thing with power ranking.. Let's face it they are just imperfect stats. It's not really his fault though - as I said before no matter how you tweak the numbers your not going to get perfect predictions. Sometimes teams feature guys who can step it up and weren't going all out during the regular season so those stats he relies on get rendered useless.

[Willie Beamen]

Why do people get all caught up in "power rankings"? Outside of the BCS power ranking are meaningless. Ok so some guy has the Celtics rank the number 2 team in the league. Why should we care about that. According to the teams record they are the number 1 team in the league. But that really doesn't matter because the last I seen championships aren't won during the regular season. So if Hollinger was to crown someone else king then crown them. It's not going to make a difference come June