Ben Alamar of ESPN is a Hack (new ESPN power rankings!)

[Pterinochilus]

I'll just use my eyes instead of these ridiculous made up stats. The Hawks haven't just been winning against the top teams in the league they have been destroying them. It's so preposterous that Chuck just said he thinks the Bulls and Wizards are the best teams in the East even though the Bulls and Wizards just got dismantled by us.

[rd79]

 

You may be familiar with the Basketball Power Index from college basketball. ESPN's NBA version, which debuts Monday and will appear each week, is similar to the College BPI and the Hollinger Power Rankings but with some sophisticated tweaks so that it accounts for overall strength of schedule, pace, number of days of rest, game location and preseason expectations.

 

I found the hidden secret which allows to make the ranking to show whatever you want...

 

The Hawks have the best record vs above .500 teams and are 10-2(?) vs the West? How should they climb the rankings? Since the East is the weaker conference and we have to play vs Eastern teams the majority of our games there is no chance to climb in such rankings. But nevertheless at the end there will only be one Eastern and one Western team in the finals and then nobody cares about this BS-rankings...

[DrReality]

This dude sounds like some of our posters!  Who does he think he is? Hotlanta?

ha ha ha ha

[Guest]

I found the hidden secret which allows to make the ranking to show whatever you want...

The Hawks have the best record vs above .500 teams and are 10-2(?) vs the West? How should they climb the rankings? Since the East is the weaker conference and we have to play vs Eastern teams the majority of our games there is no chance to climb in such rankings. But nevertheless at the end there will only be one Eastern and one Western team in the finals and then nobody cares about this BS-rankings...

Hehehe, yup. This essentially means Alamar is tossing in a subjective measure, then hoping no one notices in his explanation of a new analytic. Analytics are supposed to be objective. Not so much here.

[Guest]

Now that's the Hotlanta I know!

Oh God, what a moron. I'm a homer now. This is so stupid and clearly Hots is too dumb to realize I have statistical issues with the guy and not homer tendencies. I wouldn't expect him to realize that though because he's f***ing stupid and can only be negative.

I'd love to hear him try and explain Bayes theorem, which I imagine he'd question why that is even important here, thus illustrating what I just said above.

[High5]

This is the dumbest article I've seen posted since that Wages of Win site called Marvin our best player and a "star forward." 

[TheNorthCydeRises]

The only metric you need to trust in ... is the NTR ( Northcyde Team Rankings )

ORtg + DRtg* / 2 = NTR

* Defensive rating is adjusted, since the strength of most good to great teams is predicated on how well they play defense. So I'll add a multiplier to the overall defensive rating, depending where they are ranked.

Defensive multipliers:

Top 5: ( DRtg x .700 ) .. elite defense

6 - 10: ( DRtg x .850 ) .. good defense

11 - 20: ( DRtg x 1.00 ) .. avg defense

20 - 25: ( DRtg x 1.15 ) .. bad defense

Bottom 5: ( DRtg x 1.30 ) .. horrible defense

Top 10 teams ( lower the number the better )

Golden St ( 2.35 )

Atlanta ( 4.55 )

Portland ( 6.2 )

San Antonio ( 7.75 )

Houston ( 7.9 )

LA Clippers ( 8.5 )

Dallas ( 10.0 )

Memphis ( 10.5 )

Washington ( 10.9 )

Toronto ( 11.0 )

Note: Most NBA Finals teams have an NTR rating of 8 and below. Most NBA Champions are at 6 or below.

At 2.35, Golden State is the strongest NTR team since the 1995 - 96 Chicago Bulls ... who posted a perfect 0.85 rating ( seeing that they were #1 in offense and defensive rating that season.

[Bahamut]

F'in moron.

[TheNorthCydeRises]

The only metric you need to trust in ... is the NTR ( Northcyde Team Rankings )

ORtg + DRtg* / 2 = NTR

* Defensive rating is adjusted, since the strength of most good to great teams is predicated on how well they play defense. So I'll add a multiplier to the overall defensive rating, depending where they are ranked.

Defensive multipliers:

Top 5: ( DRtg x .700 ) .. elite defense

6 - 10: ( DRtg x .850 ) .. good defense

11 - 20: ( DRtg x 1.00 ) .. avg defense

20 - 25: ( DRtg x 1.15 ) .. bad defense

Bottom 5: ( DRtg x 1.30 ) .. horrible defense

Top 10 teams ( lower the number the better )

Golden St ( 2.35 )

Atlanta ( 4.55 )

Portland ( 6.2 )

San Antonio ( 7.75 )

Houston ( 7.9 )

LA Clippers ( 8.5 )

Dallas ( 10.0 )

Memphis ( 10.5 )

Washington ( 10.9 )

Toronto ( 11.0 )

Note: Most NBA Finals teams have an NTR rating of 8 and below. Most NBA Champions are at 6 or below.

At 2.35, Golden State is the strongest NTR team since the 1995 - 96 Chicago Bulls ... who posted a perfect 0.85 rating ( seeing that they were #1 in offense and defensive rating that season.

When looking at NBA Finals matchups, the best NTR rated team only won 5 of the last 11 championships. But the better defensive rated team of the 2 have won 8 of the last 11 titles. And it would've been 9 out of 11, had Popovich not taken his best defensive player out of the game, enabling Miami to get that rebound and for Ray Allen making that 3. Edited January 20, 2015 by TheNorthCydeRises

[Guest]

This is the dumbest article I've seen posted since that Wages of Win site called Marvin our best player and a "star forward."

Ah yes. Another group that gives analytics a bad rep. At least they've never consulted for a team while Alamar has. He is also a frequent guest at the Sloan Sports Analytics Conference.

[lethalweapon3]

Meh, the dude's just talking out of his posterior distribution...

 

~lw3

[Admin]

I think that the senior members here need to stage an @hawksfanatic anger intervention! 

[Guest]

I think that the senior members here need to stage an @hawksfanatic anger intervention!

Very true - he's been on a tear since the year started.

[Guest]

This is too funny to pass up:

Tweet Added - Refresh Page to see it live!

If you don't see the Tweet here AFTER REFRESHING PAGE, then it wasn't entered correctly. Please edit and try again.

Important: If the Tweet doesn't load, please refresh the page before editing the Tweet as it may need to refresh to parse properly

When you click the link and go to his website....it's empty. Like his brain.

How in the world is this guy able to convince people he is intelligent? He is part of the reason why people think sports analytics is not helpful. But he is just bad.

[RandomFan]

@hawksfanatic

 

Since you seem to be the resident analytics guru, I wondered what you thought about the breakdowns in this video done by David Locke, the radio announcer of the Utah Jazz. He talks a good bit about the warriors and also the Hawks and some of the reasons we are having good seasons. Also very high on Korver. You can skip the first 2 minutes of the video.

 

[Guest]

@hawksfanatic

 

Since you seem to be the resident analytics guru, I wondered what you thought about the breakdowns in this video done by David Locke, the radio announcer of the Utah Jazz. He talks a good bit about the warriors and also the Hawks and some of the reasons we are having good seasons. Also very high on Korver. You can skip the first 2 minutes of the video.

I'm not entirely sure what I was supposed to look at or what level Locke was aiming for. I was able to get to about 11 minutes in before I shut it off. He described how good teams have many players with above league average points per possession. Maybe he discusses his metric a bit more in previous episodes? But describing points per possession in these contexts is NBA Analytics 101. That was first discussed with Dean Oliver in Basketball on Paper and then further pushed forward by Dave Berri's Wages of Wins.

A spreadsheet does not imply analytics is being done. As far as I could tell, Locke is a somewhat Excel guru (which is un-intuitively a bad sign for analytics) who developed his own metric, which is never known here. The reason why it's counter intuitive is that your typical excel user is more business oriented and for analytics you want someone who is primarily data oriented. That would mean using a statistical software program that was originally built for analytics, Excel was not built for that. Excel is a spreadsheet manager for maintaining data, but then Microsoft added on more functionality. So if Excel is your go-to, you're either really old, have more of a business side of background, or for some strange f***ing reason enjoy having lots of Microsoft products.

I'm going to take a big guess that Locke doesn't have a firm grasp on what is going on, but thinks he does, because I saw that he holds the concept of replacement level at the top of his spreadsheet. That is a sabermetrics trait and as far as I can tell it has never been completely hashed out in terms of what replacement level is supposed to represent and why we care. But there is this big fad in analytics communities for the NBA to drop in replacement level. But then they don't explain it. But it's not likely to work in the NBA because of salary cap restrictions and other market distortions. I've tried to engage in discussions with sabermetricians as to what "Replacement Level Theory" is, but they never give a satisfying answer. They simply talk in terms of analogies and how MLB can be thought of like a fantasy baseball league with a fixed number of roster spots and a fixed amount of money to be spent on players.........that is true for fantasy sports but not true in the real world. Which makes the whole discussion frustrating.

Edited January 27, 2015 by hawksfanatic

[Guest]

Oh and @RandomFan, thanks for having an interest in analytics and having found a video by the Jazz radio broadcaster. I like watching/reading different takes on analytics in the NBA. But I'm a really blunt person (aka asshole) so that's just my opinion of Locke. It's not bad, it's just that he has a few things that irks me about analytics and it was too low of a level of analysis.

Take a look at Basketball on Paper or Wages of Wins and you'll see Locke is essentially describing the first few chapters in those books. I'd suggest Basketball on Paper as a place to get caught up to speed with 75% of the people out there who talk about themselves as an NBA analytics person.

[RandomFan]

Thanks. While I'm interested in some of the things analytics can tell us, I'll be honest and say I'm not overly interested to the point of delving that deeply into it. That's why I figured I'd ask the expert if what this guy was spewing was hogwash or not. =D Thanks for the clarification.

 

And yeah, he spent the first half of the video setting things up too much; the 2nd half is more talking about individual players and teams. The points-per-possession stat he kept hammering on did make me think it is something that ignores defensive efficiency, so I suspected it was fairly inept at being much of a predictor.

[Guest]

Thanks. While I'm interested in some of the things analytics can tell us, I'll be honest and say I'm not overly interested to the point of delving that deeply into it. That's why I figured I'd ask the expert if what this guy was spewing was hogwash or not. =D Thanks for the clarification.

 

And yeah, he spent the first half of the video setting things up too much; the 2nd half is more talking about individual players and teams. The points-per-possession stat he kept hammering on did make me think it is something that ignores defensive efficiency, so I suspected it was fairly inept at being much of a predictor.

He's no expert, but he was not completely full of it. The first known proponent of per possession stats was Dean Smith in the 50s (although I think Dean Smith attributed this to someone else at UNC). Dean Oliver wrote on it in Basketball on Paper which brought this to a larger audience.

I don't know if PPP is inept, but yeah in this light it ignores defensive efficiency. This is generally fine if you are comparing players across an 82 game season since the competition that the each player faces will be roughly equivalent. But if you want to be predicting game to game fluctuations in performance, using PPP for players varies greatly and is affected by defensive efficiency. Locke is not accounting for that at all from what I could tell.