Professional and College Basketball Forums banner

1 - 20 of 21 Posts

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #1 ·
Here's a website that has +/- numbers for 228 players from the '02-'03 season. Garnett, Nowitzki, and Duncan are 1,2, and 3 respectively.

NBA Roland Ratings

Surprising players at the top (remember this is last year):

7.) Ostertag!
8.) Kukoc
13.) Battie
22.) Crawford (he is/was good for his team afterall)
32.) Glover
37.) Declercq
40.) Mike James

Surprising players in the very middle:

128.) Ginobili (was the only player to record an identical rating: +5.3 in and out of the game for a difference of 0.0. So does this mean he had absolutely zero impact on his team last year? :D)

Surprising players at the bottom:

228.) McInnis (dead last for all rated players)
227.) Turkoglu
226.) Mike Curry (I guess his defense isn't that valuable)
216.) Jay Williams
215.) Cassell
201.) Jalen Rose
200.) Rashard Lewis
169.) Amare
165.) Matrix (Amare and Matrix each had an over 2.0 negative impact while on the court!)

These ratings represent a player's value to a particular team and are not intended to be an accurate gauge of the ability and talent of the player away from the specific team.
I really like this stat, and am glad hockey uses it. I wish the NBA would take the time to make this an official stat, because isn't the margin of points the best indicator of a player's value?
 

·
Top Of The Pops
Joined
·
27,472 Posts
Originally posted by <b>RebelSun</b>!

I really like this stat, and am glad hockey uses it. I wish the NBA would take the time to make this an official stat, because isn't the margin of points the best indicator of a player's value?
It's a fun toy, but it's too potentially influenced by others.

For example, suppose mediocre player A plays for the Spurs and only plays when Duncan is on the floor, and only is on the bench when Duncan is on the bench. Don't you think his +/- is going to be pretty hot, since I imagine the Spurs are going to do great when he's on the floor and not so well when he's not. Not because of him, but because the Spurs are far better with Duncan than without.

Is Ostertag really the seventh most important player to his team? Or was he on the court a lot with Stockton and Malone and off the court when they were, making his +/- very high?

Pippen ranked pretty highly last season, as I recall, and that's an example of the system actually showing something, in my opinion. While Pippen's numbers weren't exciting, he clearly focused the team and ran their offense well. The Blazers demise this year has been largely due to no offensive leader. So Pippen was very valuable to his team while not being a great player anymore.

But there are just too many potentially noisy outside influences for +/- to necessarily be a great measure for the basketball. It might be a better measure for hockey because, as far as I know (I'm not a huge hockey fan), players are subbed for so often, players tend to play with all sorts of combinations of players...there's not a "first team" that plays a ton of minutes together.

The more random the group of players you play with from instance to instance, the more meaningful +/- is, because it's less likely you're just piggy-backing being with the best or worst players all the time. That cancels that effect and shows more truely what impact you, alone, have.
 

·
Registered
Joined
·
1,099 Posts
Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!
Is Ostertag really the seventh most important player to his team? Or was he on the court a lot with Stockton and Malone and off the court when they were, making his +/- very high?
Ostertag takes a lot of criticism as not being a good player... but here are the facts from last year, which explains his +/- rating.

Without Ostertag on the floor, the Jazz gave up 98.8 points per 48 minutes. With Ostertag? 86.6 Meanwhile, the Jazz gave up more points with Malone/Stockton in the game, than when they weren't.

The +/- can be very misleading, but with Ostertag I don't believe it's the case. He's a very underrated player.
 

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #4 ·
Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!
It's a fun toy, but it's too potentially influenced by others.
That may be true, there are obviously a lot of influences. However, there are all-stars w/ big negative ratios; if they were really that good a player, why would the team have a better ratio when they were out of the game?

I think these numbers help separate the good players from the winning players.I think it's especially good for defensive purposes. There are so many offensive stats, but this one shows how balanced a player is.

There are many variables, but over a full season, I think it's a valid indicator.
 

·
Top Of The Pops
Joined
·
27,472 Posts
Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>RebelSun</b>!

There are many variables, but over a full season, I think it's a valid indicator.
Only if the varables keep changing over the season, which I don't think they do (significantly) for many players in the NBA.

Again, refer to my player A on San Antonio example. While it would be extreme to say he's only on when Duncan is on and never on when Duncan isn't, I think many players are, say, 80% to 90% correlated with other players' playing time. And that's through the season, with no factor to even things out.

It may tell you something, but I suspect most people will use it to tell them what they want to hear, which is the problem with such abstract measures.
 

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #6 ·
Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!
For example, suppose mediocre player A plays for the Spurs and only plays when Duncan is on the floor, and only is on the bench when Duncan is on the bench. Don't you think his +/- is going to be pretty hot, since I imagine the Spurs are going to do great when he's on the floor and not so well when he's not. Not because of him, but because the Spurs are far better with Duncan than without.
That is a good point. I believe the more minutes a person plays the more valid the +/- is and vice-versa. Role players could have very skewed numbers.
 

·
Top Of The Pops
Joined
·
27,472 Posts
Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>RebelSun</b>!


That is a good point. I believe the more minutes a person plays the more valid the +/- is and vice-versa. Role players could have very skewed numbers.
Except in that example, the player is playing Duncan's minutes, which hardly gives him a role-player's minutes. Duncan plays major minutes.

Remember, I said this player is on when Duncan is on and off when Duncan is off. That means he's playing the same number of minutes Duncan is.
 

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #8 ·
Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!
It may tell you something, but I suspect most people will use it to tell them what they want to hear, which is the problem with such abstract measures.
This number is not for measuring a player's talent level, just their value to that particular team that year.

Who would've thought that Rashard, Sam, and Jalen would be that detrimental to their team's success?

Obviously Ostertag isn't the 7th best player in the league, but he may have been the 7th most valuable player to his team in one year, and that's all the statistic can suggest.

I'm not saying this is the end-all rating for a player's value, but it offers another look into one's contribution to a team's success.
 

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #9 ·
Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!
Remember, I said this player is on when Duncan is on and off when Duncan is off. That means he's playing the same number of minutes Duncan is.
That could definitely skew a player's number. However, 230 players aren't all going to only be playing when superstars are on the court. There will always be some that will be skewed, but over the course of 230 players, I think the accuracy would balance out.
 

·
Top Of The Pops
Joined
·
27,472 Posts
Re: Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>RebelSun</b>!


This number is not for measuring a player's talent level, just their value to that particular team that year.
I know that. I'm just doubtful that Ostertag was the seventh most valuable player to his team (I realize that that's different from the seventh best player in the league).

So the Sonics would have been a stronger team, last year, without Rashard Lewis?

In the end, it's possible that all these things are true. However, I simply believe that there's too much statistical noise for this measure in basketball, whereas there may be less in hockey, for this measure (for reasons I gave earlier).

If I were going to make a quick-and-dirty way to judge defense, I might look at opposing player +/- for a particular position.

How did the other team's center do when Center A (on your team) was on the floor and when he was on the bench?

How did the other team's shooting guard do when Guard A (on your team) was on the floor and when he was on the bench?

That would be less influenced by who's on the floor with you (though there are other factors that you may have to adjust for...haven't thought this through totally, yet).
 

·
Top Of The Pops
Joined
·
27,472 Posts
Re: Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>RebelSun</b>!


That could definitely skew a player's number. However, 230 players aren't all going to only be playing when superstars are on the court. There will always be some that will be skewed, but over the course of 230 players, I think the accuracy would balance out.
Maybe, but it's the outliers that you're looking at. You're marvelling at the people who finished at the top and bottom of the list mostly (and most people would)...and those may be the ones most "skewed."
 

·
just foul
Joined
·
6,913 Posts
Discussion Starter · #12 ·
Regardless of all the possible outside factors that may taint the rating, if we are to assume that the job of a basketball player is to help his team score more points than the opposition, then wouldn't a +/- number determine one's value to a team?
 

·
Top Of The Pops
Joined
·
27,472 Posts
Originally posted by <b>RebelSun</b>!
Regardless of all the possible outside factors that may taint the rating, if we are to assume that the job of a basketball player is to help his team score more points than the opposition, then wouldn't a +/- number determine one's value to a team?
An untainted +/- would be an awesome measure of a player's value to a team. Even a slightly tainted measure would be of some value.

I just don't know how "tainted" +/- is. Still, I look at it. It's one more thing to consider, like efficiency rating which is another measure that I think is flawed...but not useless.

+/- is definitely a cool idea. No arguments with the general concept.
 

·
Registered
Joined
·
28 Posts
As of Jan. 18 this season, roland ratings leaders:

Rank Player Team On Court +/- Off Court +/- Roland Rating
1 Kidd NJN +7.8 -15.5 +23.3
2 Garnett MIN +8.9 -11.3 +20.2
3 Davis NOH +3.4 -13.7 +17.1
4 Jefferson NJN +5.8 -11.2 +17.0
5 Collins NJN +8.9 -7.3 +16.2
6 Martin NJN +7.7 -6.2 +13.9
7 O'Neal IND +8.4 -5.0 +13.4
8 Sprewell MIN +7.7 -5.4 +13.1
9 Cassell MIN +8.4 -4.4 +12.7
10 O'Neal LAL +10.9 -0.7 +11.6
11 Payton LAL +8.9 -2.6 +11.5
12 Miller SAC +10.8 -0.3 +11.0
13 Kirilenko UTA +3.2 -7.6 +10.8
14 Hinrich CHI -2.5 -12.5 +10.0
15 Nowitzki DAL +5.5 -4.3 +9.8
16 Miller DEN +4.6 -5.1 +9.8
17 Kittles NJN +5.1 -4.7 +9.8
18 Howard ORL -3.1 -12.4 +9.3
19 Artest IND +7.3 -1.8 +9.0
20 Foster IND +9.5 +0.8 +8.7
21 Christie SAC +11.5 +2.9 +8.6
22 Glover ATL -3.5 -12.0 +8.4
23 Battier MEM +5.5 -2.7 +8.2
24 Divac SAC +11.7 +3.6 +8.1
25 Alston MIA -1.3 -9.3 +8.0
26 Hassell MIN +8.8 +1.1 +7.7
27 Jackson HOU +5.7 -2.0 +7.6
28 Brown NOH +2.8 -4.8 +7.5
29 Ostertag UTA +3.9 -3.7 +7.5
30 Abdur-Rahim ATL -5.0 -12.5 +7.5
 

·
Registered
Joined
·
343 Posts
I don't believe that +/- is a terribly useful tool to determine a player's talent. It may be useful in determining how valuable a player is to their team, but that really doesn't have a lot of correlation to talent. The reason being, that +/- is highly dependant not only on the player in question, but also on the team's bench (or starters). For example, I'm quite sure Mike Bibby is a pretty good pg, yet, I'd wager his +/- isn't particularly hot and is probably much less than it should be, due mainly to the fact the Kings are a fairly deep team overall, and Bobby Jackson is backing up Bibby. The numbers regarding +/- this year I think prove that NJ literally has no bench. All 5 starters are in the top 17.. that's pretty scary when you think about it. But I'm pretty sure it doesn't mean that Jason Collins or Kerry Kittles are particularly talented.
 

·
Welcome to the Grindhouse!
Joined
·
19,803 Posts
Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>Minstrel</b>!


Only if the varables keep changing over the season, which I don't think they do (significantly) for many players in the NBA.

Again, refer to my player A on San Antonio example. While it would be extreme to say he's only on when Duncan is on and never on when Duncan isn't, I think many players are, say, 80% to 90% correlated with other players' playing time. And that's through the season, with no factor to even things out.

It may tell you something, but I suspect most people will use it to tell them what they want to hear, which is the problem with such abstract measures.
This sounds reasonable but it has no basis in reality. If so you would see a trend that the top players on the scale would have teamates who had a +/- that was unjustly high. However, that is not what you see. Now for a team to team comparison you might want to subtract scoring margin from the +/-.
 

·
Registered
Joined
·
9,527 Posts
it's probably best used by coaches to really understand what the best units are for them out on the floor. and that's not using ratings, but situational ratings.
 

·
Registered
Joined
·
2,725 Posts
Like Minstrel said, this stat is full of confounding variables. I know people love to try to reason this out, but real stat people find this sort of stat extremely misleading. More than just the players on the court with another, players could be involved in extreme games a few times and that could help one over the other. I'm sure the people that do this for a living could come up with more confounding that we cant even think of.
 

·
Top Of The Pops
Joined
·
27,472 Posts
Re: Re: Re: Re: Re: NBA +/- Ratings (Last Year)

Originally posted by <b>MemphisX</b>!


This sounds reasonable but it has no basis in reality. If so you would see a trend that the top players on the scale would have teamates who had a +/- that was unjustly high. However, that is not what you see.
Actually, you do see things like that, like with Kidd and the rest of his teammates.

You would not see it in every case because there are other factors, like whether a player's backup is so good or so terrible that he drives up or down the player's +/- just by not being the extreme guy.

Basically, there's so much noise, you cannot really expect to see any trends. Obviously, some top players are going to be near the top and some bad players are going to be near the bottom to make it seem like maybe there's some validity to it...and, of course, team fans are only to happy to create stories for why Mr. Mediocre or Mr. Horrific is actually super-important. I even did that with Pippen which, while I think this season bears my "story" out, might just be BS.
 
1 - 20 of 21 Posts
Top