Jamesian blog

[huevonkiller]

My question is: Does either metric have a tendency to widely miss a player's actual value, under extreme conditions -- as in, great team success or failure?
If so, are there reasonable "corrections" that could be made?

I agree with you, and I call this the Mo Williams effect. He has more in common with Mario Chalmers than a legit all-star. This happens in certain "extreme" circumstances.

I like to use your earlier equation "xyz", but using more metrics. Personally I like to use, WS/48, D.O.'s defensive rating, defensive ASPM, regular PER, usage-rate-adjusted PER, and usage-efficiency tradeoffs using D.O.'s offensive rating. All of these figures have value in their own way. Oh and on/off court impact, I use that sometimes.

That's how I adjust for discrepancies at least.

[agentkirb]

many consider jordan to be the greatest SG to ever play the game, possibly the greatest player period. others may consider magic, bird, chamberlain, or russell the greatest player ever, but they played different positions than jordan did...

paul certainly has a chance to be one of the greatest PGs of all time, maybe the greatest ever once his career is done. but his career isn't done yet - he's only 27 years old. just because he hasn't won any titles doesn't preclude him from the discussion. karl malone and charles barkley never won titles, but each is considered an all-time great...

I try not to turn all GOAT discussions into a "well, this guy has more titles so he's obviously better" debate as well, and I want to extend the same courtesy to Chris Paul. But we're not going to put Jordan and Paul in the same GOAT conversation. It seems like among the casual fan its "Kobe vs Jordan" and before Kobe won his first title without Shaq a few years ago, people were calling that crazy.

Now, Chris Paul definitely has the opportunity to get to that point when all is said and done, but the two statistics that were posted suggests he's in that conversation NOW. And we're forgetting about the fact that Stoudimire and Camby are also in the top 15 all time according to the same statistic. I'm all about using statistics, I wouldn't be here if it weren't for that fact. But at some point we need to have a filter that says "ok, something isn't right about that" instead of blindly accepting that because he's one of the best according to this statistic that it must be true.

[huevonkiller]

I try not to turn all GOAT discussions into a "well, this guy has more titles so he's obviously better" debate as well, and I want to extend the same courtesy to Chris Paul. But we're not going to put Jordan and Paul in the same GOAT conversation. It seems like among the casual fan its "Kobe vs Jordan" and before Kobe won his first title without Shaq a few years ago, people were calling that crazy.

Your posts are built on false assumptions, because CP3 was not in the second link I posted. In lieu of this your argument seems silly . It is clear you didn't pay attention at all.

Further, some of Jordan's seasons are better than 09 Paul but most of them aren't. Jordan was a shell of himself in those last few years with the Bulls.

Jordan is a ballhog sometimes, it seems logical that his efficiency is overrated in certain years. Taking long two-point jumpers isn't always going to cut it against efficient players like Chris Paul.

Now, Chris Paul definitely has the opportunity to get to that point when all is said and done, but the two statistics that were posted suggests he's in that conversation NOW.

No, the second link did not say that at all.

I can re-post it for you too, so your argument continues to puzzle me.

And we're forgetting about the fact that Stoudimire and Camby are also in the top 15 all time according to the same statistic. I'm all about using statistics, I wouldn't be here if it weren't for that fact. But at some point we need to have a filter that says "ok, something isn't right about that" instead of blindly accepting that because he's one of the best according to this statistic that it must be true.

You've yet to convince me that Amare's Shaq-like production is unworthy.

As I said earlier, you already made a large error about the results of the second link, you should have read that more carefully.

[agentkirb]

I'm sorry, because of the time delay between my posts I probably forgot a lot of the context of the debate. I was responding mostly to some chart that had CP3, Amare and Camby listed among the all-time win shares for individual seasons (I think). I believe it was Win Shares per 48 minutes.

I don't know about this 2nd link you are talking about. I went back to your first post in the thread, and the second link in that post leads to a dead link. Whatever was in that link isn't there anymore. But what I was referring to was the all-time rankings for win-shares which had Chris Paul, Amare and Camby high in the all-time rankings. You say they aren't in the second link... if that's true then in a sense you are agreeing with me aren't you? Basically my whole argument is that I don't trust a ranking that has those three guys high in the all-time list. Or a better way to put it would be to say that I trust the statistic, but I have an inner filter for when I see guys like Camby and Amare ranking high among all-time players that automatically goes "ok, they are obviously not all-time greats, but the rest of this is right".

[Mike G]

All the links to b-r.com Play Index results seem to be dead.

Jordan is a ballhog sometimes, it seems logical that his efficiency is overrated in certain years. Taking long two-point jumpers isn't always going to cut it against efficient players like Chris Paul.

A player's efficiency is "overrated"?
What have "long two-point jumpers" to do with anything?

The Marcus Camby anomaly was, as I recall, that his 1999 postseason made a list of top playoff performances since 1980-something; not that he was an all-time great.

[huevonkiller]

Basketball-reference has all the links. I measured single season playoff greatness with a 300 minute cutoff (3-point era).

Then I measured playoff production from ages 24-35 with a 1000 minute cutoff (all eras). CP3 doesn't even qualify, yet you decided to make these odd criticisms.

I'm sorry, because of the time delay between my posts I probably forgot a lot of the context of the debate. I was responding mostly to some chart that had CP3, Amare and Camby listed among the all-time win shares for individual seasons (I think). I believe it was Win Shares per 48 minutes.

I don't know about this 2nd link you are talking about. I went back to your first post in the thread, and the second link in that post leads to a dead link. Whatever was in that link isn't there anymore. But what I was referring to was the all-time rankings for win-shares which had Chris Paul, Amare and Camby high in the all-time rankings. You say they aren't in the second link... if that's true then in a sense you are agreeing with me aren't you? Basically my whole argument is that I don't trust a ranking that has those three guys high in the all-time list. Or a better way to put it would be to say that I trust the statistic, but I have an inner filter for when I see guys like Camby and Amare ranking high among all-time players that automatically goes "ok, they are obviously not all-time greats, but the rest of this is right".

The point is you confused yourself from the very beginning, not just recently. Link 1 measured X, Link 2 measured Y.

I'll give you an example; T-Mac isn't in the GOAT conversation but his peak season is better than some of Jordan's seasons. Criticizing a metric that thinks highly of a young T-Mac, doesn't make sense.

A player's efficiency is "overrated"?

Yes, a player's PER can be "overrated" or "underrated".

What have "long two-point jumpers" to do with anything?

It seems obvious to me Jordan takes more of these kinds of shots. While this may not hurt his PER, it does hurt his real efficiency.

Jordan chucks his way into 36%, 37%, 38% usage, CP3 does not. Sometimes a high usage strategy will backfire.

The Marcus Camby anomaly was, as I recall, that his 1999 postseason made a list of top playoff performances since 1980-something; not that he was an all-time great.

I didn't really dispute this, I had a problem with other comments.

[kjb]

What have "long two-point jumpers" to do with anything?

It seems obvious to me Jordan takes more of these kinds of shots. While this may not hurt his PER, it does hurt his real efficiency.

Jordan chucks his way into 36%, 37%, 38% usage, CP3 does not. Sometimes a high usage strategy will backfire.

Are you talking about Michael Jordan? The guy who had on ortg of 118 on a usage rate of 33.3 for his career -- including two inefficient seasons at ages 38 and 39? Wow. I think most teams would be happy for their star players to "chuck" their way to Jordan-like results.

[Mike G]

In 1993, before Jordan's baseball days, the Bulls' ORtg was 4.9 better than league avg, and their DRtg was 1.9 better than avg.
In 1996, the revamped Bulls' ORtg was 7.6 above avg, and their DRtg was 5.8 better. Both were league best.
In 1998, their ORtg was just 2.7 above avg, and DRtg was 5.2 better.

In '98, Jordan's ORtg was 114, while his team was 107.7; the other players on the team were below NBA average in ORtg, and that's with him drawing the D.
He had his career-low (to that time, for a full season) TS%, eFG%, Ast%, Stl%, and Blk%; well below his career ORb%, DRb%, and ORtg. He'd become a "scoring specialist", more than at any other time in his Bulls' career.

Shooting "only" .533 while taking some 30% of Bulls' shots brought the team TS% up to .516.
Only Steve Kerr took as many as 3 (true) shots per game while making a higher % (.581)
Jordan averaged 27 shots and Kerr 6.

In the '98 playoffs, the Bulls retained their 5.2 better-than-avg DRtg, and they were also +5.1 in ORtg.
They got their TS% up to .522, with Jordan hitting .545.
Going another 3 MPG, he upped his Usg% from 33.7 in the RS, to 36.6 in PO.

[Mike G]

Did Jordan regularly demolish opposing backcourts in the Finals? Or did it just seem like it?
Payton gave him a good battle in '96. Anyone else?
Here's a summary, in points per game, for regular season and Finals, for the six Bulls' opponents' starting backcourts vs Michael Jordan.

Code: Select all

NBA Finals                Regular Season    RS   Finals     MJ
guard 1   guard 2    tm    PPG1    PPG2   total   total    gain

Jordan     1991     Chi    31.5            31.5    31.2    10.5
Magic     Scott      LA    19.4    14.5    33.9    23.1    
                            
Jordan     1992     Chi    30.1            30.1    35.8     8.2
Drexler   Porter    Por    25.0    18.1    43.1    40.6    
                            
Jordan     1993     Chi    32.6            32.6    41.0     7.0
KJ        Majerle   Phx    16.1    16.9    33.0    34.4    
     

NBA Finals                Regular Season    RS   Finals     MJ
guard 1   guard 2    tm    PPG1    PPG2   total   total    gain
                       
Jordan     1996     Chi    30.4            30.4    27.3     0.5
Payton    Hawkins   Sea    19.3    15.6    34.9    31.3    
                            
Jordan     1997     Chi    29.6            29.6    32.3     4.6
Stockton  Hornacek  Uta    14.4    14.5    28.9    27.0    
                            
Jordan     1998     Chi    28.7            28.7    33.5    10.6
Stockton  Hornacek  Uta    12.0    14.2    26.2    20.4    

If regular season averages had been maintained in the Finals -- or if both MJ and his opponents had equally depressed averages -- then Jordan would have outscored (Hornacek + Stockton) by some 2.5 PPG in 1998; yet he outscored them by 13 PPG.
He thus "gained" 10.6 PPG vs them, relative to the regular season.

[huevonkiller]

Oh and Marcus Camby had 25 PER on 18.4% usage, and led the playoffs in offensive rating.

He just played too few minutes per game for me to care, but his rate stat is being dismissed surreptitiously, I'd say.

Are you talking about Michael Jordan? The guy who had on ortg of 118 on a usage rate of 33.3 for his career -- including two inefficient seasons at ages 38 and 39? Wow. I think most teams would be happy for their star players to "chuck" their way to Jordan-like results.

If this were a Paul George vs Michael Jordan debate, I would love and demand that Jordan take that many shots.

But were talking about 2009 (or 2008) Chris Paul vs an off-peak Jordan, so Chris scoffs at that kind of ballhogging. That's not going to cut it against the very best individual seasons.

[huevonkiller]

In 1993, before Jordan's baseball days, the Bulls' ORtg was 4.9 better than league avg, and their DRtg was 1.9 better than avg.
In 1996, the revamped Bulls' ORtg was 7.6 above avg, and their DRtg was 5.8 better. Both were league best.
In 1998, their ORtg was just 2.7 above avg, and DRtg was 5.2 better.

In '98, Jordan's ORtg was 114, while his team was 107.7; the other players on the team were below NBA average in ORtg, and that's with him drawing the D.
He had his career-low (to that time, for a full season) TS%, eFG%, Ast%, Stl%, and Blk%; well below his career ORb%, DRb%, and ORtg. He'd become a "scoring specialist", more than at any other time in his Bulls' career.

Shooting "only" .533 while taking some 30% of Bulls' shots brought the team TS% up to .516.
Only Steve Kerr took as many as 3 (true) shots per game while making a higher % (.581)
Jordan averaged 27 shots and Kerr 6.

In the '98 playoffs, the Bulls retained their 5.2 better-than-avg DRtg, and they were also +5.1 in ORtg.
They got their TS% up to .522, with Jordan hitting .545.
Going another 3 MPG, he upped his Usg% from 33.7 in the RS, to 36.6 in PO.

On the other hand, in 1988 the Bulls offensive rating was only 1 point above the league average in Jordan's second best season. In 1994 the Bulls defensive rating improved from 106.1 (7th in the league) in 1993 to 102.7 (6th in the league), without the "All-NBA" defender.


Honestly I don't think these types of team stats say much. The individual figures are far more important.

[Mike G]

That's right. And individual stats in the context of team and opponent stats say even more.

[huevonkiller]

You can add all the context you want, either way the reality is that old Jordan was more like Kobe instead of a peak Chris Paul. A ballhog in comparison, inferior as a player but on a great team.

[agentkirb]

@huevon: I think I can see where you are coming from. All I'm really saying is that the stat where Camby, Amare, and Paul were in the top 15 (I believe it was WS/48), isn't a reflection of the players greatness. And you don't seem to necessarily refute that point. Now I suppose you can throw the stat out there and say that said players put up pretty good performances for those years specifically.

I did a little more thinking about it, and I realize a lot of what's bothering me about applying the statistic to the GOAT debate is the inconsistency in how WS is among players. Marcus Camby is a good example because apparently he had pretty much an all-time great season WS-wise and the rest of his career was just above average.

http://www.basketball-reference.com/pla ... yma01.html

If you scroll down a bit you find WS and WS/48 numbers and his best WS season he was 10.3 (.233 /48), and the season is a good 25% higher than any other season he posted up. The next highest WS/48 he posted was .182 the year before and the next highest WS was 8.5 in one of his seasons in Denver, where he averaged a career high in minutes that year coincidentally. And all of the other years are around the 5-6 range for WS or .140 for WS/48. And note, that if you just look at normal stats, stuff like blocks, points, rebounds don't have the same noticeable spikes to correspond with having two great seasons and being just good for the rest of his career.

My point is, when you can have fluctuations in those kinds of statistics without a way to explain them... you start to wonder what the statistic is really measuring. Maybe that team he was on in New York was the perfect fit to the type of player he is and as a result he was able to post his two best WS/48 seasons.

[Mike G]

Camby's career has some anomalous highs and lows, as measured by Win Shares. He's also gone from really bad to really good teams, and WS is apparently hypersensitive to this.

A possible correction would be to apply a factor that mitigates the 14th-power Pythagorean effect that WS applies to players' average possession used. The straightforward thing is to halve the difference between an exceptionally good (or bad) team and an average team, with the factor:
(W% + .5)/(2*W%)
Call this Equivalent W%. Players for a .200 W% team then don't average .040 WS/48, but .070
That is, they total .35 per game, not .20

Here's Camby's career WS/48, his teams' W% and their "equivalent W%", and his "adjusted" WS/48; along with PER for corroboration.

Code: Select all

.yr.     tm     W%    ws/48   eqW%     Adj     PER
1997    Tor   .366    .095    .433    .112    17.8
1998    Tor   .195    .022    .348    .039    15.9

1999     NY   .540    .176    .520    .169    17.6
2000     NY   .610    .182    .555    .166    19.0
2001     NY   .585    .233    .543    .216    20.9
2002     NY   .366    .102    .433    .121    15.1

2003    Den   .207    .119    .354    .203    17.5
2004    Den   .524    .154    .512    .150    17.8
2005    Den   .598    .138    .549    .127    18.0
2006    Den   .537    .141    .519    .136    19.9
2007    Den   .549    .155    .525    .148    19.1
2008    Den   .610    .148    .555    .135    17.2

2009    LAC   .232    .129    .366    .204    18.7
2010    LAC   .354    .137    .427    .165    18.3

2010    Por   .610    .152    .555    .138    17.0
2011    Por   .585    .116    .543    .108    14.8
2012    Por   .424    .100    .462    .109    14.5

2012    Hou   .515    .179    .508    .176    19.6

According to Adj, it seems Camby has played especially well in his first year with a new team, slacking off toward the end of each stint.