Espn rankings

[Fan from VT]

*shrug* I give up, heaven forbid anyone make a wrong prediction.

Apparently it means they can no longer be taken seriously. Just look at win loss records and be content that no one in the world knows basketball, because everyone makes incorrect predictions.

It is just another way at looking at performance in the NBA. Whenever people get upset about it I think about the Cetlics in 08-09 when they were 27-2, but had played pretty poorly over the last 5 or 6 games of that torrid stretch.

The whole freakin' point of Hollinger's stats is that they give him a better understanding of the game and what's going to happen when two teams play each other.  THAT'S THE WHOLE FREAKIN' POINT.  If they don't do that (and I don't see any evidence they do), they're not worth anything more than killing time in BS internet discussions.  Which is perfectly fine.

Mike

But you can't pick one game that didn't work out as evidence of a flaw. You need to look at the whole body of every prediction he has made compared to what your predictions are for every game of an entire season (no exceptions) and compared to other experts' predictions of all games for an entire season.

[Assassin70]

Looks like the BCS to me...pure foolishness.   

[BballTim]

*shrug* I give up, heaven forbid anyone make a wrong prediction.

Apparently it means they can no longer be taken seriously. Just look at win loss records and be content that no one in the world knows basketball, because everyone makes incorrect predictions.

It is just another way at looking at performance in the NBA. Whenever people get upset about it I think about the Cetlics in 08-09 when they were 27-2, but had played pretty poorly over the last 5 or 6 games of that torrid stretch.

The whole freakin' point of Hollinger's stats is that they give him a better understanding of the game and what's going to happen when two teams play each other.  THAT'S THE WHOLE FREAKIN' POINT.  If they don't do that (and I don't see any evidence they do), they're not worth anything more than killing time in BS internet discussions.  Which is perfectly fine.

Mike

But you can't pick one game that didn't work out as evidence of a flaw. You need to look at the whole body of every prediction he has made compared to what your predictions are for every game of an entire season (no exceptions) and compared to other experts' predictions of all games for an entire season.

  Here, though (from Stein's power rankings) seems to be a fairly serious flaw:

  "The Heat's six wins, in case you missed it, came against teams (Sixers, Magic, Nets twice, Wolves and Raps) that are a combined 16-32. The teams (Celtics twice, Hornets and Jazz) that inflicted Miami's four losses: 22-5."

  Yet this 6-4 team sits 2nd in Hollinger's rankings, a healthy margin over 3rd place. This isn't a situation like last year's Celts that had injury issues that screwed up their stats. It's not a "one game thing" where the lower ranked Celts beat the higher ranked Heat so the formula's wrong. It's a team that has a decent record and beats up on weak teams while struggling against good teams. Hollinger can see their record and he can see their strength of schedule. But he can't see that how they play against good teams is also a very important indicator and he doesn't factor it into his rankings.
 

[Fan from VT]

*shrug* I give up, heaven forbid anyone make a wrong prediction.

Apparently it means they can no longer be taken seriously. Just look at win loss records and be content that no one in the world knows basketball, because everyone makes incorrect predictions.

It is just another way at looking at performance in the NBA. Whenever people get upset about it I think about the Cetlics in 08-09 when they were 27-2, but had played pretty poorly over the last 5 or 6 games of that torrid stretch.

The whole freakin' point of Hollinger's stats is that they give him a better understanding of the game and what's going to happen when two teams play each other.  THAT'S THE WHOLE FREAKIN' POINT.  If they don't do that (and I don't see any evidence they do), they're not worth anything more than killing time in BS internet discussions.  Which is perfectly fine.

Mike

But you can't pick one game that didn't work out as evidence of a flaw. You need to look at the whole body of every prediction he has made compared to what your predictions are for every game of an entire season (no exceptions) and compared to other experts' predictions of all games for an entire season.

  Here, though (from Stein's power rankings) seems to be a fairly serious flaw:

  "The Heat's six wins, in case you missed it, came against teams (Sixers, Magic, Nets twice, Wolves and Raps) that are a combined 16-32. The teams (Celtics twice, Hornets and Jazz) that inflicted Miami's four losses: 22-5."

  Yet this 6-4 team sits 2nd in Hollinger's rankings, a healthy margin over 3rd place. This isn't a situation like last year's Celts that had injury issues that screwed up their stats. It's not a "one game thing" where the lower ranked Celts beat the higher ranked Heat so the formula's wrong. It's a team that has a decent record and beats up on weak teams while struggling against good teams. Hollinger can see their record and he can see their strength of schedule. But he can't see that how they play against good teams is also a very important indicator and he doesn't factor it into his rankings.
 

well its simply the fact that, despite how this sounds, there is a huge body of evidence that going forwards, close games tend toward being 50-50 in the long run while consistent blowouts, regardless of opponent, better predict future wins against any future opponent. fact is, relying on outcomes in close games as a predictor of future outcomes has repeatedly been found to be incredibly unreliable and inconsistent despite the appealing logic of such thinking.

[Fafnir]

Is losing to good teams an important indicator? Everyone is assuming it has a lot of predictive value, but I'm not so sure.

We'll have to see how the season plays out. I think this factor skews things because its currently an artifact of small sample size. I don't think a team will consistently have a point differential bordering on elite but continue to lose to other contenders everytime.

In the Heat's next 7 games they play 5 decent teams, I guess we'll see.

Phoenix, Charlotte, @Memphis, Indiana, @Orlando, Philly, @Dallas

[BballTim]

*shrug* I give up, heaven forbid anyone make a wrong prediction.

Apparently it means they can no longer be taken seriously. Just look at win loss records and be content that no one in the world knows basketball, because everyone makes incorrect predictions.

It is just another way at looking at performance in the NBA. Whenever people get upset about it I think about the Cetlics in 08-09 when they were 27-2, but had played pretty poorly over the last 5 or 6 games of that torrid stretch.

The whole freakin' point of Hollinger's stats is that they give him a better understanding of the game and what's going to happen when two teams play each other.  THAT'S THE WHOLE FREAKIN' POINT.  If they don't do that (and I don't see any evidence they do), they're not worth anything more than killing time in BS internet discussions.  Which is perfectly fine.

Mike

But you can't pick one game that didn't work out as evidence of a flaw. You need to look at the whole body of every prediction he has made compared to what your predictions are for every game of an entire season (no exceptions) and compared to other experts' predictions of all games for an entire season.

  Here, though (from Stein's power rankings) seems to be a fairly serious flaw:

  "The Heat's six wins, in case you missed it, came against teams (Sixers, Magic, Nets twice, Wolves and Raps) that are a combined 16-32. The teams (Celtics twice, Hornets and Jazz) that inflicted Miami's four losses: 22-5."

  Yet this 6-4 team sits 2nd in Hollinger's rankings, a healthy margin over 3rd place. This isn't a situation like last year's Celts that had injury issues that screwed up their stats. It's not a "one game thing" where the lower ranked Celts beat the higher ranked Heat so the formula's wrong. It's a team that has a decent record and beats up on weak teams while struggling against good teams. Hollinger can see their record and he can see their strength of schedule. But he can't see that how they play against good teams is also a very important indicator and he doesn't factor it into his rankings.
 

well its simply the fact that, despite how this sounds, there is a huge body of evidence that going forwards, close games tend toward being 50-50 in the long run while consistent blowouts, regardless of opponent, better predict future wins against any future opponent. fact is, relying on outcomes in close games as a predictor of future outcomes has repeatedly been found to be incredibly unreliable and inconsistent despite the appealing logic of such thinking.

  I'm not talking about blowouts vs close games, I'm talking about performance vs good teams as opposed to performance against bad teams. If you're going to try and tell me that teams that can't win vs top competition but pour it on vs inferior teams fare well in the post season I'll disagree.

[LakersFan_33]

It's amazing how much discussion Hollinger still gets. I personally don't take him seriously anymore, just simply because of how many times he was wrong...If you like his system, more power to you.

[Greenbean]

ESPN has two sets of rankings (Hollinger's math-based model with daily updates and Stein's subjective one), and the Celtics are on top of the Heat in Stein's (2nd overall to the Lakers (last week - they haven't updated the page yet but will today):

http://espn.go.com/nba/powerrankings

It just proves Hollinger's model is seriously flawed (SOS is just raw W-L, with no factors such as a 2nd level W-L).

Yes very flawed. I think the first couple of years he had this, it did a pretty good job of predicting the best team in the NBA. However, there might have been some luck involved there. He needs to relook at his method.

The first couple of years it also had a decent sample size. Things will level out once SoS and the sample gets sufficient. Even now you can see its slowly getting the top teams in the league ranked in some order.

Well last year he had alot of teams over the Lakers and Celtics if i remember correctly.

It would be nice for his model to go a layer deeper to analyze teams performances against good competition. In other words factor in some kind of relation to margin of victory over teams over .500.

Since when is the "model" supposed to predict the finals participants everytime?

It already factors in strength of schedule and margin of victory, to add yet another factor in attempt to "conform" to people's expectations is poor model making. Similar to what the BCS has tried to do.

Well the model is supposed to show who is playing the best basketball at the time. Going into the playoffs, his model had quite a few teams above the C's and I think the Lakers were 7th or 8th.

All I am saying is it could go a bit deeper. I understand how adding variables could completely skew the results. However i would be curious to see how adding another layer would affect things thats all. Keep in mind the model maker can weight it however he/she wants.

Take the C's last year. They put up some stinkers against bad teams but I cant remember them getting blown out by any contenders. They actually had the lead in a lot of those games.

I know stats cant account for those scenarios perfectly but just because the Celtics were not blowing out bad teams every night didnt mean that the Magic were a better team for example.

Either way I know what you mean about the sample size. It would be interesting if we could compare two models throughout the season.

[kozlodoev]

It's amazing how much discussion Hollinger still gets. I personally don't take him seriously anymore, just simply because of how many times he was wrong...If you like his system, more power to you.

He wasn't wrong any more often than your regular run-of-the-mill pundit.

[NoraG1]

*shrug* I give up, heaven forbid anyone make a wrong prediction.

Apparently it means they can no longer be taken seriously. Just look at win loss records and be content that no one in the world knows basketball, because everyone makes incorrect predictions.

It is just another way at looking at performance in the NBA. Whenever people get upset about it I think about the Cetlics in 08-09 when they were 27-2, but had played pretty poorly over the last 5 or 6 games of that torrid stretch.

The whole freakin' point of Hollinger's stats is that they give him a better understanding of the game and what's going to happen when two teams play each other.  THAT'S THE WHOLE FREAKIN' POINT.  If they don't do that (and I don't see any evidence they do), they're not worth anything more than killing time in BS internet discussions.  Which is perfectly fine.

Mike

But you can't pick one game that didn't work out as evidence of a flaw. You need to look at the whole body of every prediction he has made compared to what your predictions are for every game of an entire season (no exceptions) and compared to other experts' predictions of all games for an entire season.

  Here, though (from Stein's power rankings) seems to be a fairly serious flaw:

  "The Heat's six wins, in case you missed it, came against teams (Sixers, Magic, Nets twice, Wolves and Raps) that are a combined 16-32. The teams (Celtics twice, Hornets and Jazz) that inflicted Miami's four losses: 22-5."

  Yet this 6-4 team sits 2nd in Hollinger's rankings, a healthy margin over 3rd place. This isn't a situation like last year's Celts that had injury issues that screwed up their stats. It's not a "one game thing" where the lower ranked Celts beat the higher ranked Heat so the formula's wrong. It's a team that has a decent record and beats up on weak teams while struggling against good teams. Hollinger can see their record and he can see their strength of schedule. But he can't see that how they play against good teams is also a very important indicator and he doesn't factor it into his rankings.
 

well its simply the fact that, despite how this sounds, there is a huge body of evidence that going forwards, close games tend toward being 50-50 in the long run while consistent blowouts, regardless of opponent, better predict future wins against any future opponent. fact is, relying on outcomes in close games as a predictor of future outcomes has repeatedly been found to be incredibly unreliable and inconsistent despite the appealing logic of such thinking.

Bad teams usually do not know how to close games out. 50-50 in close games is inaccurate I think.

[reelbig4]

So this system is only valuable only when the playoffs aproach?

[Fafnir]

Bad teams usually do not know how to close games out. 50-50 in close games is inaccurate I think.

I'll have to find a link to the studies, it seems counter-intuitive but its true.

When you're dealing with close wins and losses, one or two possessions, good teams don't perform any better than bad ones.

It makes sense at a certain level. The best offense in the league scores around 112.6 points per 100 possessions. Meanwhile the league worst scores 95.7 points per 100 possessions.

When it comes down to it if the game comes down to 2 or 3 possessions the league best averages 1.1 and the worst averages .96. That's a small difference when you compress the sample size.

[Fafnir]

http://www.nba.com/thunder/news/bulls070201.html

http://www.basketballprospectus.com/article.php?articleid=1067

http://basketballprospectus.com/article.php?articleid=573

These links breakdown how point differential is a more powerful predictor than record and how close games winning percentages are random.

[Greenbean]

http://www.nba.com/thunder/news/bulls070201.html

http://www.basketballprospectus.com/article.php?articleid=1067

http://basketballprospectus.com/article.php?articleid=573

These links breakdown how point differential is a more powerful predictor than record and how close games winning percentages are random.

100% agree with that. Close games are a coin flip. It's nice if you have a team that can consistently win close ones, but it doesnt mean that team is great.

I dont want to completely throw out margain of victory. I think it is a good predictor. However it would not hurt for someone (not me!) to take a look at an adjusted margin to account for competition. I've been looking without success for margain of victory broken down by opponent win pecentage.

[BballTim]

http://www.nba.com/thunder/news/bulls070201.html

http://www.basketballprospectus.com/article.php?articleid=1067

http://basketballprospectus.com/article.php?articleid=573

These links breakdown how point differential is a more powerful predictor than record and how close games winning percentages are random.

  It doesn't really say that close game winning percentage is random. I says that better teams tend to win more close games than worse teams.