Behold, the idiocy that is John Hollinger

[j0Shi]

I really can't say I am a hater, because as a statistical freak myself I really like what he is doing. But I don't like whenever he praises his own numbers as the only source of value. PER has been accurate three times in the past 8 seasons to determinate the league MVP: James (08/09 & 09/10) and Garnett (03/04).

Other years and the PER-ranks of MVPs:
Kobe 07/08: 24.09 (7th). No MVP-candidate using PER
Dirk 06/07: 27.70 (2nd). Strong MPV-candidate
Nash 05/06: 23.29 (14th). No MVP-candidate
Nash 04/05: 22.04 (18th). No MVP-candidate
Duncan 02/03: 26.46 (t-4th). Weak MVP Candidate

Even if you give the model only two seasons (04/05 & 05/06) where it was completely wrong in case of MVP-rating, it is clearly not the only source of estimating. So an open-minded opinion should consider more than that. Everything else is just lopsided.

[FINtastic]

Originally Posted by j0Shi

I really can't say I am a hater, because as a statistical freak myself I really like what he is doing. But I don't like whenever he praises his own numbers as the only source of value. PER has been accurate three times in the past 8 seasons to determinate the league MVP: James (08/09 & 09/10) and Garnett (03/04).

Other years and the PER-ranks of MVPs:
Kobe 07/08: 24.09 (7th). No MVP-candidate using PER
Dirk 06/07: 27.70 (2nd). Strong MPV-candidate
Nash 05/06: 23.29 (14th). No MVP-candidate
Nash 04/05: 22.04 (18th). No MVP-candidate
Duncan 02/03: 26.46 (t-4th). Weak MVP Candidate

Even if you give the model only two seasons (04/05 & 05/06) where it was completely wrong in case of MVP-rating, it is clearly not the only source of estimating. So an open-minded opinion should consider more than that. Everything else is just lopsided.

But he didn't say that PER was the only criterion to judge an MVP on. He threw out several of them, and none really gave you a clear-cut MVP. Personally, I don't know how anyone can say there is a true MVP candidate at this point. You can pick holes in any choice that gets brought up.

And I don't think it's right to criticize PER for not selecting the correct MVP. When you have a bunch of journalists coming up with an MVP, it's hard to create a statistical model that tells you how they vote. Especially with the screwy reasons some of these guys come up with to justify the particular ballots they selected. So it's not really the purpose of PER to predict how the MVP vote is going to turn out.

[fluid.forty.one]

Originally Posted by Marionsucks

Who the hell is John Hollinger???

That's a fairly legitimate question.

[alexamenos]

Results after 2 weeks and 104 games:

Stein -- 74 - 30
Hollinger -- 79 - 25
HCAMARM (tm) - 80 - 24

DirNo was all over it, but I underestimated the significance of not having a means of considering home court advantage for Stein -- basically the difference between Stein and Hollinger is that Hollinger gets the home team in games between any two *close* teams, so not too surprisingly Hollinger (and the HCAMARM for that matter) are a bit ahead of Stein. Much more on this later.

[alexamenos]

Results after 3 weeks and 143 games:

Stein 100-43
Hollinger 107-36
HCAMARM (tm) 108-35

I initially surmised that Stein would trump Hollinger for the simple reason that Stein can take Hollinger's rankings into consideration as he makes his rankings but Hollinger can't consider softer, qualitative (but significant) factors. The problem (as DirNo suggested) was not making any allowance for home court advantage with Stein's rankings. In fact, in 17 games over three weeks where Hollinger predicted the winner and Stein missed, Hollinger picked the home team every single time.

Consider:

As of today, the 8th ranked team (OKC) has a Hollinger rating of 103.022 and the 17th ranked team (Mil) has a ranking of 99.554. According to Hollinger, the predicted winner of a game between OKC and Milwaukee, at Milwaukee, would be Milwaukee --

OKC -- 103.022
Mil -- 99.554 + 3.5 = 103.054

Point being....for all the (false) empiricism of Hollinger's model, he's basically predicting the home team whenever two teams are *close*, close in this case being the difference between the 9th and 17th ranked teams. (And this is before we take into consideration any reasonable 'confidence boundaries' around Hollinger's point predictions.)

Returning to a point I've made a few dozen hundred times, Hollinger's model (and others of its kind) tends to be quite good at telling us things which are obvious (the best teams are better than the worst), but lousy at telling us things which we'd like to know (who's the better between the mavs and the spurs).

That is, Hollinger's model, shorn of all the statistical pretense, basically predicts that the winner of any game in which we might have any interest will most likely be the home team.

How much of a fancy schmancy statistical model do you need to tell you that?

Note that the HCAMARM (the Home Court Advantage Modified Alexamenos Ranking Methodology) performed comparably to Hollinger's model over a significant number of games. The HCAMARM is rankings based on win % combined with an assumption that the hometeam will win any game between two teams ranked within 5 spots of each other.

Why 5 spots? It's a number I pulled out of my ass, that's why....and my model with ass-pulled numbers holds it own quite well versus Hollinger.

anyhoo...once again I'll tout the superiority of the ARM -- the ARM may not be any better than other models at predicting the outcome of any single game, but it is quite useful in predicting playoff seedings (he said with tongue in cheek), therefore it can be used to predict who's gonna get homecourt...and that's worth knowing.

[LonghornDub]

So just to clarify alexa--your model ranks solely on the basis of win %, and always takes the higher ranked team unless they are within 5 spots, in which case it takes the home team?

[alexamenos]

Originally Posted by LonghornDub

So just to clarify alexa--your model ranks solely on the basis of win %, and always takes the higher ranked team unless they are within 5 spots, in which case it takes the home team?

yep

[LonghornDub]

Originally Posted by alexamenos

yep

And despite it's simplicity, it's still as good as Hollinger's whacky formula. Figures.

[alexamenos]

Originally Posted by LonghornDub

And despite it's simplicity, it's still as good as Hollinger's whacky formula. Figures.

yeah, hence a point I made earlier:

I'm not claiming that Hollinger's forumula has less statistical power than a simple win%. I'm saying that statistical upgrade plus 50 cents will get you a coke.

[iella]

Apparently I must spread rep, but you deserve a medal. Or a coke at the very least.

[nowhereman]

Alexamenos, that's freakin' awesome.

[DTown All Day]

who cares what a journalist writes?

real talk

[alexamenos]

Originally Posted by DTown All Day

who cares what a journalist writes?

People who read journals?

[muzak]

Trollinger strikes again.

http://sports.espn.go.com/nba/playof...ormances-31-40

Dwayne Wade's 2011 finals performance ranks 20th of all time; and Dirk's clutch performances rank 33rd.

[markus1234]

Best single-game performances: No. 6

http://sports.espn.go.com/nba/playof...e-Individual-6

Tim Duncan: Going up against two of the better defensive centers of the decade in Jason Collins and Dikembe Mutombo, Duncan ripped off 32 points, 20 rebounds, seven blocks, six assists and three steals. He even shot 10-for-14 from the free throw line and he had only one turnover.

...

http://sports.espn.go.com/nba/boxscore?gameId=230604024

Dikembe Mutombo MIN 6

He played SIX FREAKIN MINUTES in that game !

Duncan=Kevin Love vs. the undersized NJN
Trollinger=liar

[bernardos70]

Clutch statistics anyone? Does Hollinger know about that?

[markus1234]

too bad Trollingr does not get it that the mavs D concentrated on LBJ and let Wade beat us na pad his meaningless stats...

[purplefrog]

Originally Posted by muzak

Trollinger strikes again.

http://sports.espn.go.com/nba/playof...ormances-31-40

Dwayne Wade's 2011 finals performance ranks 20th of all time; and Dirk's clutch performances rank 33rd.

Then I guess D-Weasel should have been the MVP. Hollinger has gone from being just confused to signs of severe dementia on this one.

BTW, D-Wade 2006 is ranked #1.

[sefant77]

i gave a crap about Trollinger before and now with a ring i gonna dedicate my next toilet dump to him

[TripleDipping]

Hollinger knows cluth like miss South Carolina knows geography.

[Windmill360]

[dude1394]

Mavs shoved a very large gold object up hollingers ass, ouchie....must hurt...

In the emergency room the conversation goes...

"It was an accident, I just sat......"

[MascisMan]

Originally Posted by Windmill360

Well...at least he got #2 right...

[chumdawg]

Originally Posted by MascisMan

Well...at least he got #2 right...

*He* didn't get anything right or wrong. The respective teams did or didn't. In retrospect, that looks about right. The Bulls and the Heat did play for the Eastern Conference, and each had a fairly convincing path to get there. The West was a little murky, with so many teams packed so close together, so it's not surprising how it played out. The Spurs, for example, were a stride above in the W/L record, but you will note that Hollinger didn't have them so high. You could damn near throw a blanket over all the rest, but do note that the Mavs rated out over Portland and OKC. The biggest surprises were the Spurs backing out so early and the Lakers getting their ass hung up on now! Other than that, looks like pretty solid methodology for Holly.

[LonghornDub]

Originally Posted by chumdawg

*He* didn't get anything right or wrong. The respective teams did or didn't.

Pretty strange take there, unless you're also prepared to say that nobody ever gets any sports prediction, ranking, or other comparative construct right or wrong because the teams are the ones who decide the games.

Of course he got stuff wrong. The real defense isn't that he didn't get anything wrong, it's that everybody gets stuff wrong, no matter their system.

[chumdawg]

Originally Posted by LonghornDub

Pretty strange take there, unless you're also prepared to say that nobody ever gets any sports prediction, ranking, or other comparative construct right or wrong because the teams are the ones who decide the games.

Of course he got stuff wrong. The real defense isn't that he didn't get anything wrong, it's that everybody gets stuff wrong, no matter their system.

Not quite. I don't believe that "he" gets anything wrong, in that "he" doesn't ever inject his own opinion into his rankings. His rankings are entirely earned by the teams themselves. He has his own methodology, sure--even if it is extraordinarily simple--but it's the teams that do what they do.

Of course the teams decide the games, as you and Hollinger know.

[Thespiralgoeson]

I still respect Hollinger way more than just about anyone else at ESPN. Although, granted, that probably isn't saying a whole lot.

[sike]

haven't we discussed/cussed this myopic little man enough yet? I refuse to allow his love affair with his flawed system to rain on this.

[LonghornDub]

Originally Posted by chumdawg

Not quite. I don't believe that "he" gets anything wrong, in that "he" doesn't ever inject his own opinion into his rankings.

No way. Each iteration of a ranking system based on a completely subjective formula ultimately reflects the opinion of the person who created the completely subjective formula. He can't hide behind the numbers just because he doesn't tweak each iteration with his beliefs at that immediate moment. His opinion is rooted in the very foundation of the rankings.

[chumdawg]

Originally Posted by LonghornDub

No way. Each iteration of a ranking system based on a completely subjective formula ultimately reflects the opinion of the person who created the completely subjective formula. He can't hide behind the numbers just because he doesn't tweak each iteration with his beliefs at that immediate moment. His opinion is rooted in the very foundation of the rankings.

Oh, come on, TDub. "Completely subjective?" There is no way. The dude lives and dies by his objective rankings. Do you like his objective rankings? I don't know. But it doesn't matter. They are objective, through and through.

[Murphy3]

Originally Posted by chumdawg

Not quite. I don't believe that "he" gets anything wrong, in that "he" doesn't ever inject his own opinion into his rankings. His rankings are entirely earned by the teams themselves. He has his own methodology, sure--even if it is extraordinarily simple--but it's the teams that do what they do.

Of course the teams decide the games, as you and Hollinger know.

It's his ranking system.. So the term "he" fits perfectly.

[xrobx]

then explain why he picked portland to beat us in the first round when he had us higher than portland in his ranking system? did he get that one wrong?

[joemoeschmoe]

Its not so much that his ranking system is wrong. He puts himself forward as a very by the numbers analytical guy when he's not. Just because he uses numbers doesn't make him analytical. He cherry picks his numbers that support what he sees.

As someone pointed out, he picked Portland despite ALL the numbers pointing to Dallas.

I remember him basically chalking up Dallas' run to the title as "getting hot." You know what an analytical numbers guy would do? Look up the clutch time stats that people on this board found.

Not only that, but he kept saying our defense was not that great and, while that is generally true, he failed to look at the defense of our crunch time unit. Again, cherry picking the numbers.

You can find numbers to support just about any argument if you cut and slice it just right. That is what he does.

He picked against us in 3/4 rounds and he picked us in 7 vs OKC, so I think he just wants to find an escape for him being flat out wrong.

[joemoeschmoe]

Furthermore, the Mavs winning is kind of a knock on his system along with 3 of the top 6 going out in the first round. Yet another reason for him to be defensive.

I still contend that the basic assumption -- that MOV tells us (almost) all about how good a team is -- is fundamentally flawed in basketball.

You tell me this: Mavs vs some scrub team with 2 min to go and tied. You're telling me that this is supposed to be a coin flip? Pick any elite team and any scrub team. I call BS.

[Underdog]

Originally Posted by chumdawg

Oh, come on, TDub. "Completely subjective?" There is no way. The dude lives and dies by his objective rankings. Do you like his objective rankings? I don't know. But it doesn't matter. They are objective, through and through.

Hollinger considers any small margin of victory to be "lucky" - please explain where luck falls into an objective model? There's no such thing as "luck" in mathematics, EVERY variable is accounted for with a rational explanation.

He might as well be using ancient prophecies or fairy dust to back up his rankings...

[DirkFTW]

Simple model for overly complex system... check
Select datapoints to plug into said model... check
Results point to uncertainty... check
Job requires conclusive prediction... check
Make s#%* up... check



"I’m Brick Hollinger. It’s playoffs now. We are here today to witness the miracle of MOV. Soon my top power rankings team will prevail. Let’s take a look… I DON’T UNDERSTAND!"

[LonghornDub]

Originally Posted by chumdawg

Oh, come on, TDub. "Completely subjective?" There is no way. The dude lives and dies by his objective rankings. Do you like his objective rankings? I don't know. But it doesn't matter. They are objective, through and through.

You're right. He discovered this objective formula existing naturally in the wilderness, slapped the "Hollinger" name on it, and threw it on ESPN.com. There was no subjective professional judgment involved in its creation.

Please.

[endrity]

I am one of the biggest Hollinger fans around here, and I really appreciate his work because out of all the writers out there he's probably the only ones that constantly gives me food for thought with his ideas. Add to that the fact that he has been one of Dirk's biggest and most vocal supporters in the national media. I certainly don't think he's "hater" at all. In fact if there is anything stupid in all of this debate is calling Hollinger a hater because he has no reason at all to be a Dallas "hater

However, it hasn't been a good year for his system. Also, that Mavs ending 5-man unit is putting some holes to his theory that in tight games victory is often lucky. These guys constantly win tight games, way beyond what should be expected in a normal distribution. And I think he's starting to believe that as well, because the last couple of articles he wrote during the Finals he wrote a lot about the amazing unit the Mavs put on the floor in the end.

[endrity]

Originally Posted by LonghornDub

You're right. He discovered this objective formula existing naturally in the wilderness, slapped the "Hollinger" name on it, and threw it on ESPN.com. There was no subjective professional judgment involved in its creation.

Please.

Please you!

Of course the formula is raw, if there could be such a thing as a formula to predict winners to begin with. His work is groundbreaking, in that none of this stuff has been done before. The only few people who can understand his work are:
1) other statisticians who work for specific teams and are not at liberty to discuss their ideas freely given the secrecy regarding the best statistics
2) obscure economists at universities who have a passion for basketball and the background to understand the technicality of this stuff, and once in a while write smth when they are bored with the zero lower bound monetary policy or unbiased econometric estimators and the other crazy stuff they think about. Those guys will publish their findings in professional journals that 99% of the people in the world have no access or no background to understand.

So, who is going to criticize Hollinger's work for you to feel satisfied? Adande, Mariotti, Skip, Stephen A., Ric Bucher???? Please. I am sure the majority of the people on this board now more about basic statistic distribution theory and regression analysis than those guys do.

[LonghornDub]

Originally Posted by endrity

Please you!

Of course the formula is raw, if there could be such a thing as a formula to predict winners to begin with. His work is groundbreaking, in that none of this stuff has been done before. The only few people who can understand his work are:
1) other statisticians who work for specific teams and are not at liberty to discuss their ideas freely given the secrecy regarding the best statistics
2) obscure economists at universities who have a passion for basketball and the background to understand the technicality of this stuff, and once in a while write smth when they are bored with the zero lower bound monetary policy or unbiased econometric estimators and the other crazy stuff they think about. Those guys will publish their findings in professional journals that 99% of the people in the world have no access or no background to understand.

So, who is going to criticize Hollinger's work for you to feel satisfied? Adande, Mariotti, Skip, Stephen A., Ric Bucher???? Please. I am sure the majority of the people on this board now more about basic statistic distribution theory and regression analysis than those guys do.

Hollinger adds something new to the discussion. He gets credit for that. He's also a very smart guy. He gets credit for that too.

But that's not the point. The point is whether it's fair to give the guy a pass when the formula he created spits out bad predictions because, hey, it's the formula! Hollinger didn't do it! Please.

Like I said, the appropriate defense for Hollinger is that everybody is wrong frequently. I happen to come at Hollinger more than other talking heads because I don't think his model is particularly worthwhile, he writes a bunch of self-serving ex post facto rationalizations, and he is generally a douche. But he's certainly not the only one who makes bad predictions.