BC: Plus-minus statistic is misleading and not very useful
Old 02-10-2010, 09:14 AM   #1 (permalink)
landry fields forever

Banned
 
Acie's Avatar
 
Join Date: Dec 2007
Location: Living in a van down by the river
Posts: 19,089
Representing:
Exclamation BC: Plus-minus statistic is misleading and not very useful

From Doug Smith's Blog.

Quote:
“For example, the whole plus-minus statistic is very misleading and not very useful.

“It’s far too generic … now, you can look at a grouping of players and give them a plus-minus but to go further, you’ve got to look at certain things in the flow of a game. There might be one player out there playing against a second unit so again, those numbers can be misleading.

“I looked at a boxscore the other day and somehow Kris Humphries was 6-for-8 from the field, he had whatever number of free throws, he ends up with 15 points, eight rebounds and he was minus-26. It made zero sense to me.”

LINK - Toronto Star
Acie is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 09:18 AM   #2 (permalink)
mooing

Senior Member
 
babyface's Avatar
 
Join Date: Dec 2008
Posts: 5,088
Representing:
Default

funny how some people here always use that statistic to justify their arguments
babyface is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 09:23 AM   #3 (permalink)
adoring his baby girl.

Senior Member
 
Sick, Wicked and Nasty!'s Avatar
 
Join Date: Dec 2007
Location: Canada eh?
Posts: 3,278
Representing:
Default

Quote:
Originally Posted by babyface View Post
funny how some people here always use that statistic to justify their arguments
There's nothing wrong with that if it's done responsibly. 60% of the time, it works everytime!
Sick, Wicked and Nasty! is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 09:36 AM   #4 (permalink)
is back baby

Large and in charge
 
Snooch's Avatar
 
Join Date: Jul 2008
Location: All up in there
Posts: 8,248
Representing:
Default

no kidding.

Seems to me I have talked against plus/minus alot here.
Snooch is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 09:53 AM   #5 (permalink)
is pounding the rock!

Senior Member
 
Superjudge's Avatar
 
Join Date: Dec 2007
Posts: 17,371
Representing:
Default

I use whatever stat I can

Superjudge is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 10:18 AM   #6 (permalink)
is back baby

Large and in charge
 
Snooch's Avatar
 
Join Date: Jul 2008
Location: All up in there
Posts: 8,248
Representing:
Default

Quote:
Originally Posted by Superjudge View Post
I use whatever stat I can

or whatever one you can understand.

Snooch is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 10:27 AM   #7 (permalink)
is the baby faced assassin

Banned
 
Join Date: May 2008
Location: YO MAMMA
Posts: 81,104
Representing:
Default

Quote:
Originally Posted by babyface View Post
funny how some people here always use that statistic to justify their arguments
+1

ie:Raptor123
jeffb is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 04:19 PM   #8 (permalink)
whatever

Senior Member
 
Join Date: Dec 2007
Location: toronto
Posts: 10,417
Representing:
Default

All stats are meaningless when you take one game into account
any scrub has a day when he scores 20 points, that doesn't make him a 20ppg player
moremilk is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 04:31 PM   #9 (permalink)
ball

Member
 
HeDoes's Avatar
 
Join Date: Jul 2009
Posts: 425
Representing:
Default

but a players' 20ppg does not "always" means that he is more useful than a 10ppg player...
HeDoes is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 04:31 PM   #10 (permalink)
looking to shed a few bad contracts

Junior Member
 
Join Date: Dec 2009
Posts: 89
Representing:
Default

I would hope that a GM would find most stats useless. They should be making judgments based on their qualitative analysis of a player, that's why they get paid hundreds of thousands or millions of dollars. Stats are for the media and fans.
WinnipegThunder is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 04:35 PM   #11 (permalink)
censored

Senior Member
 
Join Date: Apr 2008
Posts: 5,356
Representing:
Default

Quote:
Originally Posted by WinnipegThunder View Post
I would hope that a GM would find most stats useless. They should be making judgments based on their qualitative analysis of a player, that's why they get paid hundreds of thousands or millions of dollars. Stats are for the media and fans.
yup. if i were the president of any basketball team and i found out that my general manager was using PERs to make decisions, i'd fire him. I honestly find it sad that self proclaimed hardcore fans use them.
dfunkie1 is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 05:25 PM   #12 (permalink)
whatever

Senior Member
 
Join Date: Dec 2007
Location: toronto
Posts: 10,417
Representing:
Default

Quote:
Originally Posted by HeDoes View Post
but a players' 20ppg does not "always" means that he is more useful than a 10ppg player...
of course not, ppg is by far the most misleading stats of all to begin with

20 ppg on 25 possesions is horrible, 20ppg on 15 possesions is great. However, among players with a similar TS%, PPG is more reliable. But there are other factors, who you're playing with, if you're the only scorer on the team, you will get more shots, and therefore more points, on a team loaded with scorers, shots don't come as easy. So even if pierce scores 20ppg, he could easily score 26ppg (and did) with similar percentages. While Ariza increased his PPG in houston, at the dramatic expense of TS%.

And that's why +/- is the best stat available, in itself, it gives the most information. You can't use it to rank players, especially players on different teams. But, over the course of the season, +/- is the only stat who at one glance can tell you who are the most valuable (not better) players on a team.
moremilk is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 05:28 PM   #13 (permalink)
censored

Senior Member
 
Join Date: Apr 2008
Posts: 5,356
Representing:
Default

Quote:
Originally Posted by moremilk View Post
And that's why +/- is the best stat available, in itself, it gives the most information. You can't use it to rank players, especially players on different teams. But, over the course of the season, +/- is the only stat who at one glance can tell you who are the most valuable (not better) players on a team.
bryan and myself just called you 'not very useful.'
dfunkie1 is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 05:31 PM   #14 (permalink)
whatever

Senior Member
 
Join Date: Dec 2007
Location: toronto
Posts: 10,417
Representing:
Default

Quote:
Originally Posted by Jackson Filth View Post
yup. if i were the president of any basketball team and i found out that my general manager was using PERs to make decisions, i'd fire him. I honestly find it sad that self proclaimed hardcore fans use them.
Houston Rockets are using stats extensively. PER is a great stat in measuring production. If player A and player B are just as good defensively, play fairly similar minutes, in fairly similar roles (not a 5 mpg 10th man vs a 40mpg starter), than PER can be a reliable ranking tool.

In other words, if defense and PT are similar, PER is a difference maker, and a reliable one.

But any GM that solely rely on one stat (or even a collection of stats) to make decisions would not last more than a season. The process is far too complex, and there are too many factors to be sumed in one statistic, regardless of how comprehensive it is. You have to look at all.

That being said, if you need a rebounder, how are you going to start the process of finding one? By looking at the relevant rebounding stats. The alternative is to watch every player in the league for extended minutes and try to judge with your own eyes who is the better rebounder, which is unrealistic.

So there you go - stats are an useful tool, ignoring them OR overrelying on them is a mistake. A deep understanding on how every statistic works (what it means, how it should be interpreted) is indispensable. Many people don't really understand the more complex stats (+/-, PER, EW etc), so they tend to shy away from them.
moremilk is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 05:34 PM   #15 (permalink)
whatever

Senior Member
 
Join Date: Dec 2007
Location: toronto
Posts: 10,417
Representing:
Default

Quote:
Originally Posted by Jackson Filth View Post
bryan and myself just called you 'not very useful.'
I love Bryan, but that doesn't make him foolproof. Statistics are a mathematical field, and many basketball people are not very skilled in that subject. Furthermore, they fear the intrusion, they're not comfortable with the idea that math can ultimately reach a point where it describes the game in such accuracy, that it could eventually be in a position to make management decisions. That day is not here, but they're still afraid of it.
moremilk is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 05:37 PM   #16 (permalink)
is all bout dat life

Can't knock the Hustle
 
Nites's Avatar
 
Join Date: Jul 2008
Location: Vancouver
Posts: 15,083
Representing:
Default

Plus/minus is virtually useless…it makes no effort to sort out teammate effects. so, if i’m playing with a team of lebron, howard, duncan, wade (i’m running the point, suddenly I look great even though i’m a terrible basketball player. Furthermore, an otherwise-strong player can look worse if he’s got a good backup (someone like nicolas batum might look weaker by this measurement than otherwise).
Nites is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 06:00 PM   #17 (permalink)
a Piston Fan !

Senior Member
 
Join Date: Dec 2007
Posts: 2,242
Representing:
Default

Quote:
Originally Posted by jeffb View Post
+1

ie:Raptor123
Jeff lol you had to throw that in. +/- is a useful tool. It illustrates which players or group of players are out scoring or being scored on. It's basic mathematics.

What I really don't understand though is when people talk about good defence and compare it to team shooting percentage. We can play one night and one team are lights against us hitting 50 plus % from the field, however on another night the same shots are taken the team shoots 35 % from the field and they say we played great defence.
raptor123 is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 08:56 PM   #18 (permalink)
looking to shed a few bad contracts

Junior Member
 
Join Date: Dec 2009
Posts: 89
Representing:
Default

Quote:
Originally Posted by moremilk View Post

That being said, if you need a rebounder, how are you going to start the process of finding one? By looking at the relevant rebounding stats. The alternative is to watch every player in the league for extended minutes and try to judge with your own eyes who is the better rebounder, which is unrealistic.
That is exactly what a personnel department in professional sports does. Don't worry, they have all the video and more than enough time to analyze all of it.
WinnipegThunder is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 09:00 PM   #19 (permalink)
LX
effin' ineffable

In the Paint
 
Join Date: Dec 2007
Location: Toronto
Posts: 30,071
Representing:
Default

Yeah - it's hardly unrealistic. Actually - it's just reality. Math is fun, but it isn't reality. It creates models of reality. So do my wet dreams.
LX is offline   Boss Key Wife Key Reply With Quote
Old 02-10-2010, 10:59 PM   #20 (permalink)
whatever

Senior Member
 
Join Date: Dec 2007
Location: toronto
Posts: 10,417
Representing:
Default

unlike your wet dreams, statistic models are fundamental to many non-exact sciences today, and they can model things much more complex than the game of basketball. Of course, very few leading statisticians will devote any time to this area, there are much more important things to focus upon.

All statistical models start by watching a phenomenon and gathering data about it, so you can't go without looking at the games.

Ignoring advanced statistics is going to become a deadly sin in the NBA in the following decade. Already, teams like Houston who rely heavily on it, are having a lot of success drafting, and that creates a following.

Science has always created resistence when it tried to break ground into new areas, but without exception, eventually resistence becomes futile.

As far as the stat itself, I believe most people who oppose +/- are in fact opposing the various interpretations they hear. If a defensive +/- shows that everytime player X is in the game, the team allows 25% more points, that's a very clear indication that for whatever reason, the team doesn't defend as well. It's not necessarily indicating that player X is a bad defensive player, just that in the context of the team, he's having a negative effect. A coach can start at this point, and look carefully at past games and look for clues. It's a lot easier to find something, if you know what it is ..
moremilk is offline   Boss Key Wife Key Reply With Quote
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


All times are GMT -4. The time now is 06:19 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2015, Jelsoft Enterprises Ltd.
Copyright RaptorsForum.com 2005-2011

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24