This is depressing!!!!

[playmaker0017]

+/- and per48 have a lot in common. They're a crap shoot that sometimes coincide with reality.

+/- and per48's have a LOT of basis in reality - when you are comparing similar things.

To equate the difference between a 37MPG guy and a 35MPG guy ... per48 is the best estimator of equality. It creates a benchmark and a baseline. It also has a large enough sample size (per minute) than comparing a 10MPG guy and a 37MPG guy.

Plus - just so we are on the same page - when I discuss +/-, I am discussing the player's efficiency against an opponent ... NOT the +/- of him being on the court. That's a pretty difficult stat to track. You have to track unit to unit, with only 1 player difference to see the true +/- effect.

You have to UNDERSTAND statistics in order to compare the correct numbers and sample sizes. But, to dismiss per48s because "you don't like them" is silly.

To show you what I mean:

Brand's Net +/-: +2.4

Brand's Net +/-: 14.4

Medvedenko's Net +/-: +106.5 (not a typo)

Medvedenko's Net +/-: 0

Mobley's Net +/-: 11.4
Mobley's Net +/- last year: -3.0

Mobley This Year: +1.0
Mobley Last Year: -1.4

Webber's Net +/- this year: +0.5
Webber's Net +/- in 2002-03 (pre-injury): +2.0

Webber's Net +/- When I talk about +/- it's:
THIS YEAR: -1.5
02-03: +9.8

When examining +/-, you can come up with absurd numbers if you are either ignorant or stupid. You cannot statistically compare a player that nets 10 minutes in 3 games to a player that plays 40 minutes in 30 games. It's an effort in futility to get any meaningful analysis. But, when the sample size grows the results are quite significant.

 

[Zyphen]

+/- and per48's have a LOT of basis in reality - when you are comparing similar things.



To equate the difference between a 37MPG guy and a 35MPG guy ... per48 is the best estimator of equality. It creates a benchmark and a baseline. It also has a large enough sample size (per minute) than comparing a 10MPG guy and a 37MPG guy.

If you're saying a minute of play is a single sample, then even a 10MPG will have enough samples over 82 games. The confounding variables here are things like opposing player quality, diminishing returns, and correlation between time on the floor and production. Do you really think if the 37 minute guy played 2 less minutes, he'd lose production proportionally? It could happen but just as likely not. His production may very well hold steady or go up if he's less tired because of it (gets a longer rest going into 3rd quarter). If 2 players' mpg are that close together, why not just compare their numbers directly?

+/Plus - just so we are on the same page - when I discuss +/-, I am discussing the player's efficiency against an opponent ... NOT the +/- of him being on the court. That's a pretty difficult stat to track. You have to track unit to unit, with only 1 player difference to see the true +/- effect.

Ok, I thought it was the other +/-. I wasn't even aware of this kind of +/-. How do they determine this statistic? Or are you just "fiddling" around with stats like you did last time? Sure proved your understanding of statistics then.

+/You have to UNDERSTAND statistics in order to compare the correct numbers and sample sizes. But, to dismiss per48s because "you don't like them" is silly.

I'm pretty sure I understand statistics. I'm an AI major and I've taken 3 courses in stats theory and application at UVA. I also have a minor in Econ. I dismiss per48s not because I don't like them. It's because they are invalid due to their implicit inclusion of confounding variables.

+/To show you what I mean:





Brand's Net +/-: 14.4





Medvedenko's Net +/-: 0







Mobley This Year: +1.0

Mobley Last Year: -1.4







Webber's Net +/- When I talk about +/- it's:

THIS YEAR: -1.5

02-03: +9.8



When examining +/-, you can come up with absurd numbers if you are either ignorant or stupid. You cannot statistically compare a player that nets 10 minutes in 3 games to a player that plays 40 minutes in 30 games. It's an effort in futility to get any meaningful analysis. But, when the sample size grows the results are quite significant.

First, how are you getting these numbers?



Second, the validity of a sample size is not determined relative to larger samples. It's determined relative to the expected margin of error. I'm glad you agree that it's stupid to compare a 10MPG player to a 40MPG player but we seem to have different reasons for the same conclusion. It IS a statistically valid comparison given its limited parameters. The things it doesn't take into account such as player matchups, fitness, diminishing returns, etc. make the resulting statistic suspect. That's why per48 is no good. It doesn't take into account a lot of important information. Comparing Webber and Reef is already incorrect by your own admission due to their large gap in minutes, even without considering confounding factors. Or don't you think they play significantly different minutes?

 

[Zyphen]

Consider this:

If the only time per48 is valid is when 2 player's minutes are almost the same. Then why even bother with per48? If they're playing the same minutes, why not just compare their raw numbers (and then go on to discuss teammates, coaching, etc...)? Do you think if one player played 1 more or 1 less minute, their production proportionally shifts along with that time change? Because that's EXACTLY what per48 is doing. The statistic is made on the assumption that a player's production is related linearly with time. That's the absurdity and makes the premise for the entire statistic rather naive. I think a gaussian would be more fitting though probably still not very good. I imagine production will trail off after 40 minutes and there will also be a spike for a lot of players at 20 minutes and above (though not always the case). Some players might work better with limited minutes though.

 

[Kings241]

aaw there is too much math in this forum!!

 

[Zyphen]

Sorry, I just find it slightly erking to be challenged in math by some football jock (or say he claims). I tried to keep the explanation simple. It's not so much a math issue as a logic issue and also an invalid application of simplistic statistics.

 

[Bonzinator]

what do you want somebody to say? that webber came back too soon? perhaps. that he made an attempt to return to his previous role in the team's strategy? perhaps. that he stole peja's thunder, however brief it was? perhaps. what did anybody expect? chris webber was the leader of the sacramento kings. he wanted to come back and make a difference. not for personal glory, but to win a championship. you can't fault him for that. he's a superstar, one of the best in the league at his position, even now. the idea (and this is no secret) was to bring webber back when he was expected, work him back into the rotation, and get him as ready as he could be for the playoffs. it didn't pan out the way most of us would have liked, but he's not the devil who brought the kings down, as voison and many others would have you think.

working him back in the rotation is one thing but staring him and playing him 35 to 40 minutes is another. if he didnt start and would have come off the bench like duncan did for the spurs the kings would have a ring right now. but now they are 12-18 and will struggle to make the playoffs. i still think they will but its still not good enough.

 

[BMiller52]

aaw there is too much math in this forum!!

I agree...

 

[6th]

::6th stands to applaud Zyphen::

Bravo!!

 

[KP]

Thank God somebody with some intelligence finally pointed out how ridiculous some of these stats have been, talk about obscure and meaningless. Nice posts Zyphen.

 

[vj9999]

There is too much english too. I hate having to deal with all these letters and words when I am trying to read up on my favorite team

 

[love_them_kings]

what do you want somebody to say? that webber came back too soon? perhaps. that he made an attempt to return to his previous role in the team's strategy? perhaps. that he stole peja's thunder, however brief it was? perhaps. what did anybody expect? chris webber was the leader of the sacramento kings. he wanted to come back and make a difference. not for personal glory, but to win a championship. you can't fault him for that. he's a superstar, one of the best in the league at his position, even now. the idea (and this is no secret) was to bring webber back when he was expected, work him back into the rotation, and get him as ready as he could be for the playoffs. it didn't pan out the way most of us would have liked, but he's not the devil who brought the kings down, as voison and many others would have you think.

This is a good post. I agree completely. Hindsight is 20/20, and yes, maybe looking back on it things could have been handled differently, but I don't think it was a horrible plan. Webb is a warrior, and a competitor to the end, and of course he is going to start if given the chance. He wanted to win a championship so badly, and it's hard to blame him after his injury robbed the Kings of the chance the year before. Should Adelman have handled it differently despite what Webb wanted? Maybe. But again, hindsight is 20/20. If Adelman had made no effort to bring Webb up to speed and we lost in the first round everyone would be on his case for that too.

 

[1kingzfan]

MINUTES!!!! Zyphen got it right...

It's not so much a math issue as a logic issue and also an invalid application of simplistic statistics.

Per 48 stats have only somewhere between limited and zero usage, and as Zyphen has stated, it doesn't take a bunch of explanation of statistical theory to understand why.

Very simply, it usually doesn't work because comparing Minutes for Player A are not the same as Minutes for Player B. It's apples and oranges, an invalid comparison, because the minutes themselves are not the same:

- starting player minutes are usually different from bench player minutes in duration (typical duration thus can become a factor, for many reasons, in what a player produces, for reasons already mentioned here)

- "prime time" minutes are different from "garbage time" minutes

- basketball is a team sport, so minutes with different personnel will affect personal production

- "rhythm" minutes (remember KMart's interview a while back?) can be different from limited appearance minutes (check out Garcia, Martin, Thomas when they play starter's minutes versus playing in limited cameo appearances)

In the end, player comparisons using Per48 stats try to do what is really almost always quite impossible to do.

 

[BigWaxer]

There is too much english too. I hate having to deal with all these letters and words when I am trying to read up on my favorite team

HAHAH Ya no kidding can't you guys just use pictures for us dum folk.

 

[VF21]

This is a good post. I agree completely. Hindsight is 20/20, and yes, maybe looking back on it things could have been handled differently, but I don't think it was a horrible plan. Webb is a warrior, and a competitor to the end, and of course he is going to start if given the chance. He wanted to win a championship so badly, and it's hard to blame him after his injury robbed the Kings of the chance the year before. Should Adelman have handled it differently despite what Webb wanted? Maybe. But again, hindsight is 20/20. If Adelman had made no effort to bring Webb up to speed and we lost in the first round everyone would be on his case for that too.

word.

 

[playmaker0017]

Sorry, I just find it slightly erking to be challenged in math by some football jock (or say he claims).

First, playing football does NOT preclude one from getting an education.

In fact, it paid for 5 years ... during which I obtained 2 degrees. After graduating, I couldn't get a job due to the .com drop, I went to Boston (MIT) to get an MBA and a Masters of Accounting.

Please do not make the rash assumption that:

Football = Stupidity.

That's a bad math equation in and of itself.

Second, why would I have the need to lie about playing ball at my alma mater? I'm not too sure football experience at UA gets me a lot of headway on a message board with fans mostly based in Cal.

I tried to keep the explanation simple. It's not so much a math issue as a logic issue and also an invalid application of simplistic statistics.

I'll discuss that in a bit. I've got to go to PT right now though.... to rehab one of those injuries I got playing that football thing ... or so I say. I just had to reply because this was a VERY condescending post - in which you tried to play the intellectual superiority BS card ... and it's not appreciated.

 

[Zyphen]

I was being flippant. But I'm definitely superior to you in terms of knowledge of statistics. You've got quite a loose grasp of basic concepts for a guy who went to MIT. I got wait-listed by those guys for their CS program.

Anyways, you can talk all you want about your supposed real life accomplishments of football and MIT degrees, but you still haven't provided me the methodology for your +/- numbers. I suspect you just made them up. Much like you made up the last set of numbers with Reef's rebounding. It's not so much I don't believe a jock can be smart (I love sports and play tennis competitively, though not good enough to make a career out of it). It's that I find anything you claim about yourself rather shady judging from what you say and how little you back it up. I thought they at least teach a little math up in MIT.

Oh, and why did you need to find a job? Didn't you play for the Houston Texans? I'm sure that post is still in the archives...

P.S.: I guess even this might go over your head, so let me explain. I'm waving the LIAR card at you. Having to rehab a football injury? So it's not been that long since the Texans dropped you? I'm sure if I background checked all the players on the Texans from the last 5 years, I can prove your story to be BS, but why bother? Does anyone else believe this ****?

 

[Bricklayer]

This has gotten unproductive.