Hollinger on the Three Pointer

[TripTango]

Interesting piece on the correlation between three-point attempts and overall NBA success. Obviously, correlation doesn't equal causation -- there are a few different ways one can view this statistical bit of evidence, but at the very least it does suggest that complaints about the Blazers taking too many threes might be missing the mark. (Of course, taking too many BAD threes is always fair!) Having a strong inside game is the single best way to complement a good outside game, and balance keeps defenses honest.

http://sports.espn.go.com/nba/columns/story?columnist=hollinger_john&page=PERDiem-090327

 

[BIG Q]

I read the article, but it seems to be typical blather from Hollinger. In the article he called Portland an overachiever based on all of his three point hooey. So I then looked at his RPI crap also on ESPN, and it showed that Portland was expected to be 47-25, not the 45-27 they actually are. That would point to Portland being underachievers. Which is it Hollinger, are the Blazers over or under achievers? Or do you have another bit of hooey to explain this statistical oddity away?

 

[TripTango]

Instead, the top 10 is a list of the league's overachievers; Orlando, New York, New Jersey, Cleveland, Atlanta and Portland all have won more games than most expected, while one could argue that Indiana has, too, given all its injuries. And the other three teams up there -- San Antonio, Houston and New Orleans -- aren't exactly chopped liver.

I don't think he is stating that Portland won more games than he expected.

I also don't quite understand your animosity towards the guy. He works numbers, and has some creative and interesting formulas for doing so. Every pundit has a biased ranking system -- at least Hollinger makes his blatantly obvious.

 

[Draco]

Good team offense will penalize the defense and force them to leave somebody open for a 3.

The top teams shoot a lot of open threes because the defense has no other choice.

 

[BengalDuck]

I read the article, but it seems to be typical blather from Hollinger. In the article he called Portland an overachiever based on all of his three point hooey. So I then looked at his RPI crap also on ESPN, and it showed that Portland was expected to be 47-25, not the 45-27 they actually are. That would point to Portland being underachievers. Which is it Hollinger, are the Blazers over or under achievers? Or do you have another bit of hooey to explain this statistical oddity away?

Actually he called Portland an overachieving team based on the initial predictions of their record. I think the "three point hooey" went to say potentially it's a bigger indicator of success than prognostics.

 

[drexlersdad]

hollinger had most of the western playoff teams with 40 somomething wins, and portland at 42 or so at the start of the season.

his season prediction = fail.

 

[PapaG]

I read the article, but it seems to be typical blather from Hollinger. In the article he called Portland an overachiever based on all of his three point hooey. So I then looked at his RPI crap also on ESPN, and it showed that Portland was expected to be 47-25, not the 45-27 they actually are. That would point to Portland being underachievers. Which is it Hollinger, are the Blazers over or under achievers? Or do you have another bit of hooey to explain this statistical oddity away?

Hollinger lost me when he factored in "luck" as a net negative in his pre-season assessment of the Blazers and how they would finish.

His data seems to still need some refining, given its subjective nature and how it contradicts his own opinions at times.

I also wonder how a team that is 2nd in the West in point differential yet 5th in the standings is "overachieving", but that's Hollinger for you.

 

[BIG Q]

Hollinger lost me when he factored in "luck" as a net negative in his pre-season assessment of the Blazers and how they would finish.

His data seems to still need some refining, given its subjective nature and how it contradicts his own opinions at times.

I also wonder how a team that is 2nd in the West in point differential yet 5th in the standings is "overachieving", but that's Hollinger for you.

Agreed. He seems to miss the boat often on player and team assesment. He tries so hard to make everything fit into an equation as if he is Charlie Epps. At least Charlie knows that you must include every bit of data, and it must be accurately factored in as a positive or negative. As you mentioned, he factored in luck as a negative. He also factored youth as a negative, which to me is a positive with this team. A team full of vets tends to give up more when down 15 points in the third quarter. Portlands youth usually responds to make the game close and win their fair share.

Hollinger also did not factor chemistry and team play into his model. I guess that so few teams play team ball that it never occurred to him chemistry would be a major factor in Portlands success. He also missed the boat on Roy being even better, say +25% per over last season better in his predictions. This is why I am not a fan of his. His projections are more off than on. And I am not claiming to be better, but I am not claiming to be in the know like he is either.

 

[PapaG]

Agreed. He seems to miss the boat often on player and team assesment. He tries so hard to make everything fit into an equation as if he is Charlie Epps. At least Charlie knows that you must include every bit of data, and it must be accurately factored in as a positive or negative. As you mentioned, he factored in luck as a negative. He also factored youth as a negative, which to me is a positive with this team. A team full of vets tends to give up more when down 15 points in the third quarter. Portlands youth usually responds to make the game close and win their fair share.

Hollinger also did not factor chemistry and team play into his model. I guess that so few teams play team ball that it never occurred to him chemistry would be a major factor in Portlands success. He also missed the boat on Roy being even better, say +25% per over last season better in his predictions. This is why I am not a fan of his. His projections are more off than on. And I am not claiming to be better, but I am not claiming to be in the know like he is either.

He does some really good work, and I'm just a voice in the wilderness, but I can't take seriously any "statistician" who adds such subjective elements to his data. What strikes me is the instant defensiveness of the Hollinger fanbois, a group that takes seems to take Hollinger's data as objective and argues the method behind it.

One thing he does well, though, is create discussion and get web hits, so he's doing his job, isn't he?

 

[Draco]

I like Hollingers stats.

They add an extra dimension to judge teams and players by. If you just view them as one additional tool to try and understand the value of a performance they can give you some insight. All measurements have drawbacks, PPG is probably the most viewed stat but it overvalues players like Zach Randolph and Carmelo and undervalues Joel Pryzbilla or Bruce Bowen. You can watch the games and form assessments from hundreds of plays, but we also have internal bias's in how we view a players worth which stats can mitigate.

I'm a big fan of PER and some of the other efficiency stats. They are much better then just looking at a players PPG/RPG/APG. Do they have drawbacks? Sure. But if you just use them as one piece of judging a player worth they work great.

 

[TripTango]

I like Hollingers stats.

They add an extra dimension to judge teams and players by. If you just view them as one additional tool to try and understand the value of a performance they can give you some insight. All measurements have drawbacks, PPG is probably the most viewed stat but it overvalues players like Zach Randolph and Carmelo and undervalues Joel Pryzbilla or Bruce Bowen. You can watch the games and form assessments from hundreds of plays, but we also have internal bias's in how we view a players worth which stats can mitigate.

I'm a big fan of PER and some of the other efficiency stats. They are much better then just looking at a players PPG/RPG/APG. Do they have drawbacks? Sure. But if you just use them as one piece of judging a player worth they work great.

Agreed. Numbers are numbers, and anyone can spin them just about any way they want to, but that doesn't mean it isn't fun to talk about em! Hollinger is a professional extrapolator -- he takes past data and projects what is most likely to happen based on the trends within that data. This can be interesting, occasionally informative stuff, but it certainly isn't meant to be perfect or even perfectly objective -- he must inherently prioritize some traits and demote others. Furthermore, as anyone who's been in the stock market lately can tell you, past performance does not guarantee future results. I just don't see how what he does is any worse (or better, mind you) than the sports pundit who bases conclusions on anecdotal evidence, impressions, and gut feelings. It's just discussion fodder!