Hollinger Rankings: C's #11

[Fafnir]

The guy is the one who setup the Maverick's team adjusted +/- system.

http://waynewinston.com/wordpress/?p=1194

He actually did predict the Mavericks woudl beat the heat, largely based on how said system rated their playoff performances in the first three rounds.

http://waynewinston.com/wordpress/?p=1102

The C's were actually [dang] close to the Heat and other contenders, we had a legit shot at a run till Wade's takedown. We were as good a team as the Mavericks when you cut our bench down and played playoff ball.

[byennie]

Now, THAT is a strawman argument.  I haven't dismissed stats.  All I asked was "Where's the statistical analysis that predicted the Mavs over the Heat" and pointed out stat head Hollinger's incorrect forecast of last year's playoffs.  What, exactly, is unfair or objectionable about pointing out the failure of one stat head in particular and statistical analysis in general when it came to last year's NBA playoffs?

YOU'RE the one who took issue with those statements and if it wasn't out of typical stat head sensitivity, why was it?

Mike

I actually am at least partially confusing you with other poster(s), so you're right - you haven't been as dismissive as I claimed. Others here have claimed that opinion and scouting are not just important, but better measures when left alone. So I apologize if I lumped you in with others.

The problem with pointing out one particular instance where stats failed is that it almost perfectly misses the whole point of the stats. The "stat heads" freak out because you are giving exactly the wrong argument against a stat-based prediction.

Suppose you roll a die. My prediction is that it will land on a "3" or higher, based on statistical odds that I have calculated. It comes up "2" and you claim my stats failed... but I'm still pretty darn sure that my prediction was a good one. I've got some math to back it up, i.e. 4/6 = 67% chance, and you have anecdotal evidence.

[byennie]

Also worth adding, suppose you had THE best system which would give you perfect "odds" for each team winning the championship and it came out something like this:


Celtics - 10%
Clippers - 10%
Sixers - 10%
Lakers - 10%
Thunder - 15%
Bulls - 20%
Heat - 25%

Now my prediction is Heat over Thunder in the finals, and let's say the Celtics lose in the 1st round because while they have a shot at going all the way, they are still most likely to lose over 50% of the time to the Bulls or Heat.

I may have made the best possible predictions available, nobody could ever name a more likely scenario. And yet, I will still be wrong to some degree about 90% of the time.

[guava_wrench]

It is completely unreasonable to expect anyone or anything to predict everything correctly. It is preposterous to ask 'why didn't method X predict even Y' in sports.

Statistical methods are used to get advantages that improve outcomes over time and multiple decisions. The idea is to improve your success rate from, for example, 53% to 63%.

It does not make sense to judge Oakland on just championships. They need to be judged on bang for the buck. The succeeded more than they should of, but even that success can only last so long until other teams adopt the same methods. Oakland could not just throw money at the problem like the Yankees or Red Sox do.

Understanding the point of the methods should help show why Hollinger's rankings are interesting. The methods are used to identify undervalued assets, NOT necessarily the best players. By attempting to move from passionate observations to dispassionate math, you can find neglected value. In basketball, for example, overvalued assets like scorers can be identified and you can then bring in the undervalued assets to have a stronger roster while staying under the cap. With the Hollinger rankings, we might get insight into a team with a good record being overrated due to a home-game heavy early season with narrow victories over bad teams.

[guava_wrench]

Yeah, you're wrong about that.

He's wrong that Hollinger doesn't watch basketball, but he IS right in that Hollinger's stats are not very sound.

Hollinger bases most of his work off of PER and +/-, which are notoriously fickle numbers that don't in any way reflect performance. For example, if you JUST look at the season PER and +/- averages, and use it to extrapolate the record of a team, you will usually be WAY off. Therefore it isn't a good indicator of performance, because it can't be used retroactively to EXPLAIN performance. This is also what leads to players like Carmelo Anthony being rated as superstars, because of high PERs, despite being incredibly inefficient players.

There are much better metrics out there (though none are perfect). WP/WP48, WS/WS48 (there is a different between Wins Shared and Wins Produced), efficiency differentials, and others can all be used to explain performance.

You're referring to the wrong numbers. Hollinger's rankings are not based on PER (the validity of which is a different discussion). They're based on margins of victory, strength of schedule, amongst others.

Also, for all his faults, calling Carmelo Anthony an incredibly inefficient player isn't accurate.

I wouldn't even call him inefficient.  I mean if you are looking at purely rates, Anthony and Pierce are nearly identical players in TS%, eFG%, REB%, AST%, TOV%, BLK%, STL%, USG%, etc.  Anthony is better in some and Pierce is better in some, but none are really significantly in favor of one over the other.  Pierce is a better defender (but even the gap there is inflated by respective teams and systems). 

Yes, exactly. The composition of the Knicks is inefficient. Anthony, himself, isn't. In fact, PP's career path is probably a good barometer for where Anthony will go.

The composition of the Knicks has nothing to do with it.

The complaints that Melo is extremely inefficient date back to his Denver days when he had more talent around him.

He is inefficient because he takes a lot of bad shots and does fancy moves that he can't finish. He waits for defenses to dig in and then tries to go 1 on 3, finishing with short-armed layups.

He some some great skills, but is far from the total package.

[LooseCannon]

I think that some people just fundamentally understand the usefulness of statistical analysis.

One thing that people don't understand is that numbers like margin of victory are a better predictor of future wins and losses than previous wins and losses.  That's a bit counter-intuitive for the mindset that wins and losses are the only stat that matters.  People sometimes overestimate the degree of control you have over one's fate.  Whether its wins or how much money is made from financial investments, they underestimate how much luck and variance can play a role.

Statistical analysis also does not lead to completely deterministic results.  A firm understanding of how things work means that predictions should be expressed as probabilities.  This doesn't fit into the yes/no binary thinking of punditry, whether sports or politics.  Statistical analysis didn't fail if it didn't predict the Mavs beating the Heat.  That an outcome is less likely doesn't mean it necessarily is completely unlikely.  A stat-minded person would have dismissed those sports talking heads who make hyperbolic claims such as the Mavericks having zero chance of beating the Heat last season.

[FallGuy]

Yeah, you're wrong about that.

He's wrong that Hollinger doesn't watch basketball, but he IS right in that Hollinger's stats are not very sound.

Hollinger bases most of his work off of PER and +/-, which are notoriously fickle numbers that don't in any way reflect performance. For example, if you JUST look at the season PER and +/- averages, and use it to extrapolate the record of a team, you will usually be WAY off. Therefore it isn't a good indicator of performance, because it can't be used retroactively to EXPLAIN performance. This is also what leads to players like Carmelo Anthony being rated as superstars, because of high PERs, despite being incredibly inefficient players.

There are much better metrics out there (though none are perfect). WP/WP48, WS/WS48 (there is a different between Wins Shared and Wins Produced), efficiency differentials, and others can all be used to explain performance.

You're referring to the wrong numbers. Hollinger's rankings are not based on PER (the validity of which is a different discussion). They're based on margins of victory, strength of schedule, amongst others.

Also, for all his faults, calling Carmelo Anthony an incredibly inefficient player isn't accurate.

I wouldn't even call him inefficient.  I mean if you are looking at purely rates, Anthony and Pierce are nearly identical players in TS%, eFG%, REB%, AST%, TOV%, BLK%, STL%, USG%, etc.  Anthony is better in some and Pierce is better in some, but none are really significantly in favor of one over the other.  Pierce is a better defender (but even the gap there is inflated by respective teams and systems). 

Yes, exactly. The composition of the Knicks is inefficient. Anthony, himself, isn't. In fact, PP's career path is probably a good barometer for where Anthony will go.

The composition of the Knicks has nothing to do with it.

The complaints that Melo is extremely inefficient date back to his Denver days when he had more talent around him.

He is inefficient because he takes a lot of bad shots and does fancy moves that he can't finish. He waits for defenses to dig in and then tries to go 1 on 3, finishing with short-armed layups.

He some some great skills, but is far from the total package.

Agreed he's not the total package.

Completely disagree that he's extremely inefficient. Efficiency includes things beyond scoring and Melo has some serious skills in other areas too.

[byennie]

Yes, exactly. The composition of the Knicks is inefficient. Anthony, himself, isn't. In fact, PP's career path is probably a good barometer for where Anthony will go.

I still think there is "hope" for Anthony (and even with his flaws he's pretty good); however, Paul Pierce has a TS% of 56.9% for his career, Anthony's is 54.4%. For 7 straight years, Pierce has had a higher TS% than ANY year in Anthony's career.

Here is a partial list of players who have consistently higher TS% than Anthony, most with several years higher than Anthony's career best:

Pierce, Allen, James, Wade, Bryant, Dirk, Deron, Granger, Stoudemire, Nash, Durant

Every one of those players has flirted with or passed 60% in a season - Anthony tops out around 56%.

Here are a couple about the same or worse:
Johnson, Igoudala

I'm not saying Anthony is bad player, but compared to his assumed peers, he doesn't score very well on efficiency.

[byennie]

Agreed he's not the total package.

Completely disagree that he's extremely inefficient. Efficiency includes things beyond scoring and Melo has some serious skills in other areas too.

I agree with this in principle. I do think his scoring is pretty inefficient, more so than most people think, but he is also a very strong rebounder, a good post-up SF, and one of a handful of players that can score 30+ on any given night. He's just overrated until he becomes more efficient at scoring, because while the other stuff matters, scoring IS his primary value.

This coming from someone that started a Pierce from Anthony thread, less you think I don't value him...

[guava_wrench]

Yes, exactly. The composition of the Knicks is inefficient. Anthony, himself, isn't. In fact, PP's career path is probably a good barometer for where Anthony will go.

I still think there is "hope" for Anthony (and even with his flaws he's pretty good); however, Paul Pierce has a TS% of 56.9% for his career, Anthony's is 54.4%. For 7 straight years, Pierce has had a higher TS% than ANY year in Anthony's career.

Here is a partial list of players who have consistently higher TS% than Anthony, most with several years higher than Anthony's career best:

Pierce, Allen, James, Wade, Bryant, Dirk, Deron, Granger, Stoudemire, Nash, Durant

Every one of those players has flirted with or passed 60% in a season - Anthony tops out around 56%.

Here are a couple about the same or worse:
Johnson, Igoudala

I'm not saying Anthony is bad player, but compared to his assumed peers, he doesn't score very well on efficiency.

I think you have made the important point.

Critics aren't saying that Anthony is not an all star caliber player. They are saying that he is not in the same league as Lebron, Wade, Kobe or Dirk. He is not a transcendent player.

Melo likes 'hero ball' too much and is another 'feast or famine' player whose feasts make sportscenter while only those watching him daily know about all the games lost due to him shooting his team out of a victory.

[FallGuy]

Yes, exactly. The composition of the Knicks is inefficient. Anthony, himself, isn't. In fact, PP's career path is probably a good barometer for where Anthony will go.

I still think there is "hope" for Anthony (and even with his flaws he's pretty good); however, Paul Pierce has a TS% of 56.9% for his career, Anthony's is 54.4%. For 7 straight years, Pierce has had a higher TS% than ANY year in Anthony's career.

Here is a partial list of players who have consistently higher TS% than Anthony, most with several years higher than Anthony's career best:

Pierce, Allen, James, Wade, Bryant, Dirk, Deron, Granger, Stoudemire, Nash, Durant

Every one of those players has flirted with or passed 60% in a season - Anthony tops out around 56%.

Here are a couple about the same or worse:
Johnson, Igoudala

I'm not saying Anthony is bad player, but compared to his assumed peers, he doesn't score very well on efficiency.

I think you have made the important point.

Critics aren't saying that Anthony is not an all star caliber player. They are saying that he is not in the same league as Lebron, Wade, Kobe or Dirk. He is not a transcendent player.

Melo likes 'hero ball' too much and is another 'feast or famine' player whose feasts make sportscenter while only those watching him daily know about all the games lost due to him shooting his team out of a victory.

This is all true, but across the league it's tilting almost too much in favor of the new narrative, where 'Melo is some team cancer, with whom you cannot excel. That's not true, either.

[guava_wrench]

Mentioning Melo's rebounding brings up another important point.

How good of a rebounder was KG in his prime and how good is Howard? It is hard to say, but Rebound numbers can be misleading. In another thread, someone pointed out that Orlando is not a good rebounding team. This can easily happen if Howard is stealing rebounds from teammates. Look at KG's style of rebounding in the past. Instead of boxing out, he goes to the hoop and tries to grab the ball off of the rim (which he can no longer do well with less ups). This means if his 4 teammates boxed out their man and he didn't box out his, he might have increased the chances of the opponent getting an offensive board. The reason is if he missing the rebound, his man is running free. If he just boxed out his man and missed the board, all of the opponents are boxed out and one of his teammates would grab the ball.

My point is that there are in fact many misleading stats. Since high school, I have felt that RBIs are ridiculously overvalued and tell us almost nothing about the skill of a player. They really should only be used for recording game history. I maintained that Runs Scored, though similarly flawed, is actually marginally more correlated to individual ability since a play that is clearly linked to individual ability -- stealing home or getting home on a throwing error off of a steal attempt is strongly correlated to a skill of that player, not just opportunity.

Nevertheless, stats can help us better understand players. Though many hate basketball +/- here, it seems to be a stat well respected by basketball statheads and it is easy to see why. Using +/- is an attempt to do a controlled experiment. It is very noisy, as are all stats, but it does afford us the ability to compare that player's performance to an actual alternative - how teammates perform with similar lineups.

It is easy for skilled players to look good while leading their teams to losses. Guys like Ben Gordon can get grossly overpaid, while guys like Bosh can be mocked (like last season) despite the reality that he has decided to win instead of showing off. Guys with gaudy stats sometimes need to make a decision: (a) give up the gaudy stats to provide what the team needs to win or (b) continue to do my own thing. Clearly the big 3 on the Celtics agreed to do (a). Bosh clearly chose (a) and Lebron seems to me to have such a high bb IQ that he is naturally team-oriented and has always been willing to do the dirty work. Apart from Melo's efficiency, there is also the doubt many have that he can adjust his game over the course of an entire season to make his team and teammates better. Pierce learned to move more without the ball under Doc. Dirk developed a post game and stopped taking threes. This season, Lebron too has stopped taking 3's and is showing a legitimate post game. Yet Baron Davis has never stopped taking horrible shots. Who will Melo end up like?

[Who]

Just a note -- Orlando are not actually a weak rebounding team. That is not correct.

The Magic have been a high level rebounding team for the last couple of years. They were in the top five in rebounding differential (per game numbers, not percentages) in each of the last two seasons. Finishing 2nd last year and 4th the year before.

This season they have dropped back down to 9th place (well above average but no longer high level) where they also were in 2009 when they made the NBA Finals. In 2008, they were a middle of the pack rebounding team. In 2007, 5th. In 2006, 8th. In 2005, 10th.

That makes it only once in the past 8 years that Orlando have finished outside of the top ten and in three of those years (including the previous two) they have finished in the top five.

Orlando are routinely amongst the best defensive rebounding teams in the league since SVG joined the club but their offensive rebounding has suffered as a result of his game plan. The four out one in system + his prioritizing transition defense over offensive rebounds.

The Magic have led the NBA in defensive rebounding percentage for each of the past two seasons and finished 2nd the year before that. They were tied 6th the year prior to that. Their offensive rebounding numbers are usually towards the bottom of the league despite Dwight Howard's efforts except for last season when they finished middle of the pack. Prior to SVG, the Magic finished 4th in offensive rebounding percentage but were only 13th in defensive rebounding percentage.

[guava_wrench]

Just a note -- Orlando are not actually a weak rebounding team. That is not correct.

The Magic have been a high level rebounding team for the last couple of years. They were in the top five in rebounding differential (per game numbers, not percentages) in each of the last two seasons. Finishing 2nd last year and 4th the year before.

This season they have dropped back down to 9th place (well above average but no longer high level) where they also were in 2009 when they made the NBA Finals. In 2008, they were a middle of the pack rebounding team. In 2007, 5th. In 2006, 8th. In 2005, 10th.

That makes it only once in the past 8 years that Orlando have finished outside of the top ten and in three of those years (including the previous two) they have finished in the top five.

Orlando are routinely amongst the best defensive rebounding teams in the league since SVG joined the club but their offensive rebounding has suffered as a result of his game plan. The four out one in system + his prioritizing transition defense over offensive rebounds.

The Magic have led the NBA in defensive rebounding percentage for each of the past two seasons and finished 2nd the year before that. They were tied 6th the year prior to that. Their offensive rebounding numbers are usually towards the bottom of the league despite Dwight Howard's efforts except for last season when they finished middle of the pack. Prior to SVG, the Magic finished 4th in offensive rebounding percentage but were only 13th in defensive rebounding percentage.

Thanks for the clarification. I was wondering about that since someone mentioned it a day or two ago. Perhaps they were thinking about offensive rebounds, which don't mean much to me since I am more concerned with getting back on defense.

[LooseCannon]

Compare Hollinger to the Wages of Win ranking the Celtics at #8 (through 2/2).  Of course, you might find a ranking that puts the Clippers and Lakers at 14th and 15th, respectively, a little bit weird.