Where Al, Josh and Joe stand statiscally

[thecampster]

I thought I'd do some digging....light digging.

Checking Josh, Al and Joe per their FIC40 and Reina values (FIC is Floor Impact per 40 minutes. Reina is value for their contract.)

Al has an FIC40 of 18.2 and a Reina Value of 217% above Salary.

Josh has an FIC40 of 16.2 and a Reina Value of 40% above Salary.

Joe has an FIC 40 of 12.6 and a Reina Value of 37% below Salary.

What this means.

Al makes 5,444,857 but is performing at $17,260,197

Josh makes 11,600,000 but is performing at $16,240,000

Joe makes 16,324,500 but is performing at $10,284,435

[macdaddy]

Wow. I have no idea how they come up with those numbers but they seem a little crazy. I read that as Al and Josh are contributing like some of the best players in the league as the salary figure is close to the max. Not sure how that can be. Joe's $10 million sounds dead on.

[niremetal]

Yeah. Because stats tell the whole story.

[thecampster]

Yeah. Because stats tell the whole story.

Reina is contrived directly off of FIC.

The author tries to not skew for rebounds but it's hard not to. The ratio still skews against volume shooters.

I am sorry I didn't post a link to values:

http://basketball.realgm.com/src_playerrankings.php

Joe's number is down because the formula accurately weighs against volume shooting and not getting to the free throw line.

Edited March 28, 2011 by thecampster

[thecampster]

Posted from http://basketball.realgm.com/src_fromtherafters/138/20080109/the_reina_value/

System provided by Christopher Reina

FIC

Here is the formula I use to determine the FIC:

(Points

+ Offensive Rebounds

+ .75 Defensive Rebounds

+ Assists

+ Steals

+ Blocks

- .75 Field Goal Attempts

- .375 Free Throw Attempts

- Turnovers

- .5 Personal Fouls)

Now, here is how Reina Value is calculated:

With some context given to the statistic used to rank players, here is the crux of what the Reina Value actually is:

Players are ranked from highest to lowest by the total FIC for the season, not per 40 minutes or per game, which I decided to do because players are only valuable when they are on the floor.

Beside each player?s actual salary, we slide in raw salary figures, ranked top to bottom, which determines their ?deserved? salary. We use the actual salaries because players deserve whatever their agents can negotiate and GMs/owners can afford to pay.

The player who has the highest FIC receives the highest ?deserved? salary. The player with the second highest FIC receives the second highest salary. The player with the hundredth highest FIC receives the hundredth highest salary.

We then calculate the percentage increase or decrease from the actual and deserved.

If Dwight Howard, who is making a little over $6 million in 2007-2008, is ranked first using the FIC, his ?Reina Value? would be Kevin Garnett?s $23.7 million salary, making his difference 291.9%.

Conversely, if Stephon Marbury is ranked 139th and his actual salary is just over $19 million, he?d be in line to make about $5 million. His percentage difference, therefore, would be -73%.

By using the Reina Value to examine the rosters of all 30 teams, it soon becomes clear which GM's are getting favorable returns and which players are out/under performing their contracts.

I find it to be a fairly solid metric.

[Randy]

Interesting, but I think the salary part is a bit questionable. Guys salaries are determined by the salary cap and free angency rules. A guy who has been around a long time and becomes an unrestricted free agent gets paid a lot more than a guy perfroming at the same level who is in rookie contract or into his restricted free agency period. This salary ratio is probably a rough measure of how screwed up the salaries are in the NBA due to free agency rules and the salary caps.

[benhillboy]

No metric is perfect, but numbers don't lie and you can take a respectable amount of info from them and disregard those figures that can be skewed and manipulated as you see fit. Bottom line, even last year, most metrics showed both Al and Josh to be statistically more valuable and efficient than Joe before he and Al re-upped. He made like 120,000 more than them both combined. In summary, I guess we should focus on the great value we get from those two instead of harping on Joe's 120 that we unfortunately have to shut up and live with until a silly GM takes it off our hands.

[GameTime]

Joe makes 16,324,500 but is performing at $10,284,435

[terrell]

I thought I'd do some digging....light digging.

Checking Josh, Al and Joe per their FIC40 and Reina values (FIC is Floor Impact per 40 minutes. Reina is value for their contract.)

Al has an FIC40 of 18.2 and a Reina Value of 217% above Salary.

Josh has an FIC40 of 16.2 and a Reina Value of 40% above Salary.

Joe has an FIC 40 of 12.6 and a Reina Value of 37% below Salary.

What this means.

Al makes 5,444,857 but is performing at $17,260,197

Josh makes 11,600,000 but is performing at $16,240,000

Joe makes 16,324,500 but is performing at $10,284,435

Looks about right to me. lol

[AHF]

Under this system:

Kevin Love (#1) > Dwight Howard (#3) & Lebron James (#2)

Zach Randolph (#8) > Dwayne Wade (#9) > Derrick Rose (#10)

Josh Smith (#17) > Kobe Bryant (#18)

Kyle Lowry > Ray Allen, Stephen Curry, Danny Granger, Chauncey Billups, Jrue Holliday, etc.

This looks like an arbitrary and flawed metric by my eyeball test.

[niremetal]

No metric is perfect, but numbers don't lie

Anyone can find or invent statistics to support their position. To quote Twain, there are three kinds of lies in the world: Lies, damn lies, and statistics.

[thecampster]

Under this system:

Kevin Love (#1) > Dwight Howard (#3) & Lebron James (#2)

Zach Randolph (#8) > Dwayne Wade (#9) > Derrick Rose (#10)

Josh Smith (#17) > Kobe Bryant (#18)

Kyle Lowry > Ray Allen, Stephen Curry, Danny Granger, Chauncey Billups, Jrue Holliday, etc.

This looks like an arbitrary and flawed metric by my eyeball test.

Actually, I would submit to you that although the metric is flawed (all are), you could easily say the metric is a sampling system. +/- error kind of a thing.

As I said earlier, rebounds are always going to be weighed too heavily. But you can't discount them. Where Howard gets hurt and why he isn't 1st is his missed free throws.

Lebron gets hurt on FG% because he's fouled going to the basket a lot which generates FG misses and FT misses. Same with Kobe.

But if you consider it as slightly flawed, you find Al and Josh in the top 20 and then you +/- however you want. But you find Joe far down in the 60's. It just goes to show that he is not anywhere near where his contract says he is.

and by the way, it doesn't rate best/worst player. It rates floor impact. Kevin Love's floor impact for Minnesota is beyond measure.

[Guest]

Reina is that dope from RealGM, he talks out of his *ss the whole time and doesn't know a damn thing about statistics. He is just pulling different metrics together and is almost always baseless with his justifications.

Stats don't lie (unless they are incorrectly calculated), but you need to know how to interpret them. Whenever a dope tells you his statistic and it includes their name (like "Reina Value" or whatever it is) then you can safely conclude its worthless. I recall going through Reina's articles a few years back and from what I recall it was pure garbage. He understands that there is a cap along with restrictions on contracts, yet he completely ignores it when calculating a "Reina Value". As for the FIC, its an ad hoc formula much like John Hollinger's PER.

[Seano]

Where do our guys stand statistically against *good* teams? That would be my question, because by that metric, I would say they'd all rank much lower than that given by how they perform against quality competition.

[AHF]

Reina is that dope from RealGM, he talks out of his *ss the whole time and doesn't know a damn thing about statistics. He is just pulling different metrics together and is almost always baseless with his justifications.

Stats don't lie (unless they are incorrectly calculated), but you need to know how to interpret them. Whenever a dope tells you his statistic and it includes their name (like "Reina Value" or whatever it is) then you can safely conclude its worthless. I recall going through Reina's articles a few years back and from what I recall it was pure garbage. He understands that there is a cap along with restrictions on contracts, yet he completely ignores it when calculating a "Reina Value". As for the FIC, its an ad hoc formula much like John Hollinger's PER.

There are a lot of junk formula that basically add up all the countable stats and use that to evaluate player performance and this looks like one of them. To me, it looks like a very raw (and thereby even less valuable) PER.

points + offensive rebounds + .75 defensive rebounds + steals + blocks + assists....

That is junk without meaningful explanations as to why those numbers predict good performance going forward, why they are weighted like they are, etc.

Then layering in the salary numbers while treating them like free market economy numbers without taking into account the fact that there are contract minimums and maximums, etc. just compounds the error.

Looking year over year, the variation is pretty stark:

Rajon Rondo 2009-10 #3, 2010-11 #22

David Lee 2009-10 #10, 2010-11 #37

Monta Ellis 2009-10 #77, 2010-11 #21

Joe Johnson 2009-10 #28, 2010-11 #60

Al Jefferson 2009-10 #42, 2010-11 #16

Are these guys that different from last year? Elton Brand or Michael Beasley I get. But these guys haven't changed thaat much so I question the worth of the statistic just like I would question ranking guys by total point scored to determine how good a player they are. The number tells you something but what does it tell you? Reina seems to make no effort to describe the limitations of his "statistics."

Let's single out an example.

Kyle Lowry

#138

http://basketball.realgm.com/src_playerrankings.php?season=2009-2010&filter=all&highlight=

2009-10

15.5 PER

.126 WS/40

13.5 pp36

.536 ts%

5.4 rp36

6.7 ap36

1.3 sp36

2010-11

#42

http://basketball.realgm.com/src_playerrankings.php?season=2010-2011&filter=all&highlight=

16.6 PER

.128 WS/40

16.6 pp36

.553 ts%

4.4 rpg36

7.1 ap36

1.5 sp36

Which of these numbers tells you the most about how much a player has improved or regressed over the two seasons? I'll give you a hint as the number that tells you the least.

Edited March 28, 2011 by AHF

[Bawse]

Yeah. Because stats tell the whole story.

Well..

Statistically, this is the worst season Joe Johnson has had as an Atlanta Hawk. A little disturbing, getting 18.5 PPG from a 126mill player, am I right?

Edited March 28, 2011 by Boss

[AHF]

Well..

Statistically, this is the worst season Joe Johnson has had as an Atlanta Hawk. A little disturbing, getting 18.5 PPG from a 126mill player, am I right?

Nobody is arguing that this has been a poor season for Joe. I don't think these stats are good ones for telling that story compared to something like the WS/40 metric which at least ties out to reality better.

According to this metric:

2010-11 Jose Calderon > Joe Johnson, Tyson Chandler, Danny Granger, Gerald Wallace, Brook Lopez, Chauncey Billups, etc.

[thecampster]

There are a lot of junk formula that basically add up all the countable stats and use that to evaluate player performance and this looks like one of them. To me, it looks like a very raw (and thereby even less valuable) PER.

points + offensive rebounds + .75 defensive rebounds + steals + blocks + assists....

That is junk without meaningful explanations as to why those numbers predict good performance going forward, why they are weighted like they are, etc.

Then layering in the salary numbers while treating them like free market economy numbers without taking into account the fact that there are contract minimums and maximums, etc. just compounds the error.

Looking year over year, the variation is pretty stark:

Rajon Rondo 2009-10 #3, 2010-11 #22

David Lee 2009-10 #10, 2010-11 #37

Monta Ellis 2009-10 #77, 2010-11 #21

Joe Johnson 2009-10 #28, 2010-11 #60

Al Jefferson 2009-10 #42, 2010-11 #16

Are these guys that different from last year? Elton Brand or Michael Beasley I get. But these guys haven't changed thaat much so I question the worth of the statistic just like I would question ranking guys by total point scored to determine how good a player they are. The number tells you something but what does it tell you? Reina seems to make no effort to describe the limitations of his "statistics."

Let's single out an example.

Kyle Lowry

#138

http://basketball.realgm.com/src_playerrankings.php?season=2009-2010&filter=all&highlight=

2009-10

15.5 PER

.126 WS/40

13.5 pp36

.536 ts%

5.4 rp36

6.7 ap36

1.3 sp36

2010-11

#42

http://basketball.realgm.com/src_playerrankings.php?season=2010-2011&filter=all&highlight=

16.6 PER

.128 WS/40

16.6 pp36

.553 ts%

4.4 rpg36

7.1 ap36

1.5 sp36

Which of these numbers tells you the most about how much a player has improved or regressed over the two seasons? I'll give you a hint as the number that tells you the least.

But the numbers bear out similarly when looking at PER

AL 19th

Josh 36

Joe 73

As far as are the players that much different - Joe has had a reduction in FG%, 3PT%, FT%, Rebounds, Points, steals and .1 more T/O a game so yes...it is that much different.

[Guest]

Looking year over year, the variation is pretty stark:

Rajon Rondo 2009-10 #3, 2010-11 #22

That I can explain by a stupid RealGM sorting mechanism. They are sorting based on the raw FIC over the year, which includes the playoffs. Rondo played 105 games last year, the most out of anyone. I don't understand your Lowry example, it appears the "rankings" are just based on the sorting mechanism, but also Lowry is a marginal player whose ranking will undoubtedly go down based on other marginal players simply playing more games once the playoffs trigger the sorting mechanism's effect.

But a lot of your points are valid, there is no justification for the weights. Its just a slightly tweaked version of PER. I still like the WOW approach best because they actually run regressions to figure out weights. Hollinger and Reina are dopes who just pluck numbers out of their *ss. And if you are pompous enough to put your name on a statistic you are an assclown, especially if its a site like RealGM.

[CBAreject]

I don't think anybody is suprised with those rankings for our particular players. We'd almost unanimously say Al is our best player and Josh is our 2nd best player, and most of us would say that Joe is overpaid but still a good player.

The difficulty with basketball metrics is, to quote a hackneyed phrase, "there is only one ball". In short, somebody has to take the shots, and somebody has to get the rebounds. Any field goal attempt takes one away from a teammate. Some portion of the time, a rebound takes one away from a teammate. Subbing a 15 RPG player for an 8 RPG player doesn't give you 7 more team rebounds, for example; it might just give you 3 or so.

Compare this with baseball, where the number of run scoring opportunities are unlimited. In baseball, the runs created and offensive win shares formulas are very good and can be tested with years of cumulative statistics and computer simulations. Basketball is a much more dynamic game. I think it might be a useful starting point to take years of team statistics and see how assists, FTA, and offensive rebounds translate into greater efficiency in points per 100 possessions and so on. I'm sure someone has done this, and I'm betting Hollinger has used something similar. But as AHF mentioned, any arbitrary weighting of statistics is prone to copious error.

Edited March 29, 2011 by CBAreject