PDA

View Full Version : Pacer +/- with regards to offensive and defensive rating so far this year



Hicks
11-08-2010, 10:26 PM
Kind of interesting. I have to admit, over the past several years, the more I've looked at/thought about plus-minus, the less skeptical I've become.

I'm no longer surprised that I tend to agree with what it bears out in this adjusted format in the link:

http://www.indycornrows.com/2010/11/8/1801244/11-8-10-pacers-stat-of-the-week-plus-minus

Offensive and defensive rating are a team's points per game and points allowed per game, only they're both adjusted to account for pace, allowing you to use these numbers to fairly compare all 30 teams, regardless of how fast/slow their tempos are from one another.

Here, we're seeing the Pacers' Off and Def ratings while each individual player is on or off the floor.

flox
11-08-2010, 10:28 PM
My only problem with this blog, and this was something that I was going to comment on before I got distracted by the bobcats game, is that it doesn't focus on certain units. Of course certain players will have worse +/- because they are more often paired with other players (looking at Posey and Solo)

OakMoses
11-09-2010, 12:59 AM
Don't let the anti-Dunleavy crowd look at these numbers...

Who am I kidding? They'll just write them off anyhow. Just imagine how good they'd be if he were shooting well...

Really, other than Dunleavy, it's almost exactly what you'd expect. Everyone knows that the drop off from Roy to Solo - in every facet of the game - is unbearable. Posey and TJ's numbers are a bit surprising. However, I'd expect Posey's to rise and TJ's to stay relatively the same as the season progresses.

Like Flox, I'd like to see a bit more of a breakdown of the units. I wonder what TJ and Posey's numbers look like during their non-Solo time.

IUfan4life
11-09-2010, 01:08 AM
Kind of interesting. I have to admit, over the past several years, the more I've looked at/thought about plus-minus, the less skeptical I've become.

I'm no longer surprised that I tend to agree with what it bears out in this adjusted format in the link:

http://www.indycornrows.com/2010/11/8/1801244/11-8-10-pacers-stat-of-the-week-plus-minus

Offensive and defensive rating are a team's points per game and points allowed per game, only they're both adjusted to account for pace, allowing you to use these numbers to fairly compare all 30 teams, regardless of how fast/slow their tempos are from one another.

Here, we're seeing the Pacers' Off and Def ratings while each individual player is on or off the floor.

not too be so off topic, but Breaking Bad is the best show in TV

pwee31
11-09-2010, 01:09 AM
Wow Solo..Wow

graphic-er
11-09-2010, 01:35 AM
I can't believe these numbers have any sort of relevancy for one second. No way Dunleavy has a more positive impact on the game than Granger. The Dunleavy who goes 1 for 7 from 3pt line, who constantly gets beat off the dribble and is slow to close out on shooters, who scores less than 10 pts in 38 minutes of play. IF anything his rating is take away from others rating. Does this rating system take into account when a player get scored on because mike dunleavy lost his assignment seconds earlier?

Hicks
11-09-2010, 01:45 AM
not too be so off topic, but Breaking Bad is the best show in TV

I'm only four episodes in via Netflix blu-ray. I really like it.

kester99
11-09-2010, 02:32 AM
Does this rating system take into account when a player get scored on because mike dunleavy lost his assignment seconds earlier?

Yes. It takes into account all points scored for/against while a player is on the floor.

Here's a link to the stats site in question. There is an adjusted +/- that takes into account what players you were on the floor with (so the claim goes). That column lists Hibbert, Granger, Mike, George...in that order, as the 4 Pacers with a positive adjusted plus/minus.

http://basketballvalue.com/teamplayers.php?year=2010-2011&mode=summary&sortnumber=94&sortorder=DESC&team=IND

cdash
11-09-2010, 02:38 AM
I'm only four episodes in via Netflix blu-ray. I really like it.

I wish I could go back and watch it all for the first time. Really a great TV show.

croz24
11-09-2010, 03:10 AM
it's all about roy. when he's in the game and getting touches, the pacers are scoring efficiently. when he's in the game and patrolling the paint, the pacers are doing a better job of denying penetration. when he's in the game period, the pacers are winning.

Mackey_Rose
11-09-2010, 06:22 AM
Don't let the anti-Dunleavy crowd look at these numbers...

Who am I kidding? They'll just write them off anyhow. Just imagine how good they'd be if he were shooting well...

Really, other than Dunleavy, it's almost exactly what you'd expect. Everyone knows that the drop off from Roy to Solo - in every facet of the game - is unbearable. Posey and TJ's numbers are a bit surprising. However, I'd expect Posey's to rise and TJ's to stay relatively the same as the season progresses.

Like Flox, I'd like to see a bit more of a breakdown of the units. I wonder what TJ and Posey's numbers look like during their non-Solo time.

I suppose I would fall in the "anti-Dunleavy crowd," and I looked at the numbers.

And you are right, I am going to just write these off because any statistical system that concludes Dunleavy's defense is a net-positive is inherently flawed.

I tend to go by what I see, and not a computer or statistical based model, and I see Dunleavy as the biggest defensive flaw on the team right now. Granger is a very close second, but at least his offense is enough to make up for it.

wintermute
11-09-2010, 06:58 AM
I don't like looking at plus/minus so early in the season. With only 5 games played, the results are easily skewed. But now that we're discussing it, here's the link for 5-man units:

http://basketballvalue.com/teamunits.php?year=2010-2011&team=IND

The starters have generally been decent, but the bench has been another story.



Most effective units (in terms of minutes and overall rating):

Collison-Dunleavy-Granger-McRoberts-Hibbert 67.53 min +10.70
Collison-Dunleavy-Granger-Posey-Hibbert 24.80 min +3.74
Collison-Dunleavy-Granger-Hansbrough-Hibbert 14.12 min +46.42
Ford-Dunleavy-Granger-Hansbrough-Hibbert 8.68 min +21.18
Ford-George-Granger-Hansbrough-Hibbert 7.95 min +59.49

Least effective units (in terms of minutes and overall rating):

Ford-George-Granger-Posey-S.Jones 17.47 min -38.97
Ford-George-Dunleavy-Posey-S.Jones 9.37 min -39.41
Ford-Dunleavy-Granger-Hansbrough-McRoberts 6.98 min -28.02
Ford-D.Jones-Granger-Posey-Hibbert 6.42 min -16.67
Collison-Dunleavy-Granger-Posey-S.Jones 5.83 min -41.67


A few things stand out to me. The Ford-George-Posey-S.Jones quartet has been absolutely terrible as a unit. Rush's return from suspension hopefully adds something to the bench mix. If I were Obie, I'd be trying to work Price in there too. Not many options at reserve center, but I'd like McBob to get more minutes there to see if he could do better than Solo. In fairness to Solo though, his individual +/- probably suffers from not playing much with the starters.

On a brighter note, the Hibbert-Hansbrough-Granger trio has been very productive in limited minutes.

Brad8888
11-09-2010, 10:56 AM
The plus / minus throws everybody's cumulative impact statistics against the wall, they predictably go splat, and then the plus / minus becomes useful in validating personal viewpoints, in the eyes of whoever chooses to use them, for whatever purpose they choose to use them for at a given time.

Look at Troy Murphy. I hated the game he was told to play, and, as a result, I hated the game the team played while he was on the floor. His plus / minus validated my viewpoint, and those of others, but I think it was more of an indictment against the O'Brien system than Murphy as a player, and he was the worst on the team by a wide margin IIRC.

So, in my view, plus / minus partially measures player effectiveness and coaching effectiveness, while not being terribly effective at either because it cannot quantify game situations that contributed to the overall performance, such as how much time was spent against teams who weren't playing at full force due to blowouts, play against other second string units, etc.

vnzla81
11-09-2010, 11:06 AM
So according to this Dunleavy is the best player?

:wtf2:

jhondog28
11-09-2010, 11:08 AM
So according to this Dunleavy is the best player?

:wtf2:

hahaha that never gets old

Unclebuck
11-09-2010, 11:27 AM
I do sometimes look at stats, but I still go by what I see while watching the games. And through 5 games whenver Roy and or Danny go to the bench the team falls apart. Also in some games when Mike goes to the bench, IMO things don't work as well offensively. George might come in and score, he might make a steal, but the team is not workong as well because he is a rookie and most rookies play on an island - meaning they aren't able to play the team part of the game.

I can only determine that by watching the game and I get a better feel when I'm at the game.

ksuttonjr76
11-09-2010, 11:49 AM
Basically, we need another player, ahem Dunleavy, to give the bench some type of chemistry. IMHO, Rush or George should be the starting SG. I would prefer George, since the wouldn't have the "pressure" of scoring with Collison, Granger, and Hibbert, but he could do all the "little" things to help the Starting 5. Rush and Dunleavy would provide some type of offense for the bench, and could help keep the game close against the other team's bench.

DaveP63
11-09-2010, 11:50 AM
Let's revisit these staticstics in another month or two, but I'm thinkin' they're not too far off. UB is right on the money.

Hicks
11-09-2010, 12:02 PM
Technically they're not off at all. You just have to interpret them while keeping in mind what they are and what they are not telling you.

flox
11-09-2010, 12:19 PM
Wow Dunleavy is on more poor units than I thought. And Ford's effectiveness in certain units is exactly on par as to what i'd suspect.

Infinite MAN_force
11-09-2010, 12:22 PM
I think Mike's positive numbers can be explained mostly by two factors. One is that he spends most of his time on the floor with Roy, Danny, and the starters, clearly the unit that will produce the best +/-.

Two is that when Mike subs out Paul George subs in for him, who then proceeds to take everyone's numbers down with his mistakes. The offensive flow and team defense suffer when Paul subs in... not because Mike subs out.

Not a knock on Paul, he is just a rookie... but I still don't think Mike is some kind of "secret ingredient" to our offensive success. Roy Hibbert is the one that can make that claim. I also truly believe that the numbers would be even better with Rush in Dunleavy's place, because he would shoot at a higher percentage and play better defense.

OakMoses
11-09-2010, 12:34 PM
I suppose I would fall in the "anti-Dunleavy crowd," and I looked at the numbers.

And you are right, I am going to just write these off because any statistical system that concludes Dunleavy's defense is a net-positive is inherently flawed.

I tend to go by what I see, and not a computer or statistical based model, and I see Dunleavy as the biggest defensive flaw on the team right now. Granger is a very close second, but at least his offense is enough to make up for it.

Sorry. My 'anti-Dunleavy crowd' comment was a bit hasty and was the sort of unproductive comment I generally hate.

I agree with you that Dunleavy's defense is not good and is by far the worst in the starting 5.

However, I do think that these numbers prove that the Pacers as a team, have been better with Dunleavy on the floor than with him off the floor.

All the other arguments that have been brought up stand: The sample size is too small, Dunleavy usually plays with Hibbert, etc.

One thing the numbers confirm for me is that Dunleavy and Hibbert are really the only 2 guys who consistently run the offense the way O'Brien wants it to be run. Dunleavy is the only guy who consistently moves well without the ball. His movement, however, is a shadow of what it was pre-surgery.

The bottom line for me is that, if he were shooting well, Dunleavy's offensive numbers would be fantastic if he were making shots, and I don't think 5 games is enough of a sample to pull a guy who's historically been a very good shooter. His FG% will rise substantially over the rest of the season. His defensive deficiencies will still exist, but right now it's the offense, not the defense, that is losing us games.

beast23
11-09-2010, 12:46 PM
I do sometimes look at stats, but I still go by what I see while watching the games. And through 5 games whenver Roy and or Danny go to the bench the team falls apart. Also in some games when Mike goes to the bench, IMO things don't work as well offensively. George might come in and score, he might make a steal, but the team is not workong as well because he is a rookie and most rookies play on an island - meaning they aren't able to play the team part of the game.

I can only determine that by watching the game and I get a better feel when I'm at the game.I agree with your opinion, with one addition. If you have no objective explanation to rationalize what the statistics reveal, I think you have to give some credibility to the "outliers".

I do not consider Granger's stat that is not as positive as we would assume to be an outlier, and it is rationalized by the fact that he spends quite a bit of time on the floor with the second unit.

On the other hand, although I have not been negative towards Posey as a lot of others have been, you have to look at the SJones - Posey dynamic and conclude that there is some pretty compelling evidence to at least never use them together. One might also conclude that their minutes individually should also be minimized, but the sample size, i.e. 5 games is very small.

Looking at the statistics, I would change a few things. I would be inclined to primarily use a 3-man rotation (Roy, Josh, Tyler) in the front court until Foster returns. Since I believe that each of the 3 men are capable of playing 32 minutes to cover the total 96 minutes, I would use SJones sparingly... very sparingly.

I am also surprised by how strong Dunleavy's presence is in the starting unit. I agree with you that my observations lead me to believe he does have a positive influence, but the stats seem to indicate a very positive influence from the Hibbert-Dunleavy dynamic. The strength of that dynamic very much surprises even me, an advocate of Dunleavy remaining in the starting unit.

Rush is returning; I do think it is important to sort out which player has the better influence on the starting unit and also whether either player is able to provide a boost to the second unit. So, I would hope that each player has a lot of playing time with the four other starting players. I would think that would help JOB sort out player combinations regarding Rush and Dunleavy.

The other info that I gain from the stats, although certainly not a "strong" tendancy, is that neither Josh nor Tyler seem to have a huge impact on our defensive success. If a much stronger indication is evident after half a season or so, maybe we should consider that we are not getting the defense and rebounding from either one that we require at the PF position.

Anyone attempting to use these statistics would obviously feel much more confident if all possible combinations of players were presented for equal amounts of minutes, but obviously that is never going to happen. But I do believe the statistics reveal useful information, particularly in cases where there is no reasonable means to explain away what they indicate.

beast23
11-09-2010, 01:10 PM
I think Mike's positive numbers can be explained mostly by two factors. One is that he spends most of his time on the floor with Roy, Danny, and the starters, clearly the unit that will produce the best +/-.

Two is that when Mike subs out Paul George subs in for him, who then proceeds to take everyone's numbers down with his mistakes. The offensive flow and team defense suffer when Paul subs in... not because Mike subs out.

Not a knock on Paul, he is just a rookie... but I still don't think Mike is some kind of "secret ingredient" to our offensive success. Roy Hibbert is the one that can make that claim. I also truly believe that the numbers would be even better with Rush in Dunleavy's place, because he would shoot at a higher percentage and play better defense.Careful... I think your rationalization is a bit of a stretch in what the statistics are intended to reveal.

I don't believe that there is any data that supports the influence that a particular player has in substituting in for another player. The statistics only reveal what happens offensively and defensively when one particular player is either on or off the floor.

If there were an ample amount of data for the four starters with George (minus Mike), then it might be possible that your statement could be supported.

But the data shows such a very strong influence with the combination of Hibbert and Dunleavy that even with only 5 games of data it seems very likely that as a pair they somehow have a positive influence on the starting unit. It is certainly something to watch for awhile to observe what impact a few more games reveals in the +/- statistics.

Infinite MAN_force
11-09-2010, 01:51 PM
Careful... I think your rationalization is a bit of a stretch in what the statistics are intended to reveal.

I don't believe that there is any data that supports the influence that a particular player has in substituting in for another player. The statistics only reveal what happens offensively and defensively when one particular player is either on or off the floor.

If there were an ample amount of data for the four starters with George (minus Mike), then it might be possible that your statement could be supported.

But the data shows such a very strong influence with the combination of Hibbert and Dunleavy that even with only 5 games of data it seems very likely that as a pair they somehow have a positive influence on the starting unit. It is certainly something to watch for awhile to observe what impact a few more games reveals in the +/- statistics.

Fair enough. I do think Dunleavy's movement has a positive impact. I'm just not convinced its essential. Paul George's impact on team performance is the worst after solo, and that is do to inexperience within the offensive and defensive systems. I think its reasonable to assume he has a higher probability of screwing up team chemistry then most.

We are certainly dealing with a small sample space, there is no doubt about that.

Psycho T
11-09-2010, 03:07 PM
I think Hansbrough should get a chance to start whenever he learns the plays and can execute them well enough.

Mackey_Rose
11-09-2010, 03:28 PM
I think Hansbrough should get a chance to start whenever he learns the plays and can execute them well enough.

I don't think what Hansbrough adds is what we need more of in the starting lineup, but it is exactly what we need more of out of the 2nd unit.

Eleazar
11-09-2010, 03:55 PM
In my opinion +/- is only good for judging how well a group of players play together compared to other groups of players, and nothing more. It has little to do with the individuals, and more to do how well they play together.

beast23
11-09-2010, 04:34 PM
In my opinion +/- is only good for judging how well a group of players play together compared to other groups of players, and nothing more. It has little to do with the individuals, and more to do how well they play together.I think most would agree with that opinion. 5-player +/- are more telling than individual +/-.

The only exception that I have would be the individual +/- that is extreme on either the positive or negative end of the scale. If the extreme statistic cannot be qualified with a logical explanation, then I think you are compelled to take a closer look at it because there may be something very informative there.

Eleazar
11-09-2010, 05:43 PM
I think most would agree with that opinion. 5-player +/- are more telling than individual +/-.

The only exception that I have would be the individual +/- that is extreme on either the positive or negative end of the scale. If the extreme statistic cannot be qualified with a logical explanation, then I think you are compelled to take a closer look at it because there may be something very informative there.

I can agree with that. I also believe there are always exceptions to the rule.

Infinite MAN_force
11-09-2010, 06:57 PM
I think most would agree with that opinion. 5-player +/- are more telling than individual +/-.

The only exception that I have would be the individual +/- that is extreme on either the positive or negative end of the scale. If the extreme statistic cannot be qualified with a logical explanation, then I think you are compelled to take a closer look at it because there may be something very informative there.

Its not always useful in the context of one game (individually), but I think over the course of a long season an extremely low/high plus minus is very telling.

I think Troy Murphy is a very good example of this, I think his consistently last place team plus/minus was very telling last season. :devil:

xBulletproof
11-09-2010, 07:45 PM
Kind of interesting. I have to admit, over the past several years, the more I've looked at/thought about plus-minus, the less skeptical I've become.

I'm in the same boat. I used to ignore it and think it was useless. I've slowly moved towards the point where I look at them now. I do notice they confirm what I'm seeing on the court as well.

This includes Dunleavy. I think people give him crap because his guy is open sometimes, but that's because he's covering someone elses mistake. He doesn't leave his man for no reason, and they like to ignore how often him leaving his man to help is a positive. Of course we're talking about the same group of people who threw a fit about his shooting in a game where he went 3 for 8. Reading that game thread you'd have thought he went Brandon Jennings on us and made 4 out of 20 shots.


not too be so off topic, but Breaking Bad is the best show in TV

I've never even heard of this show, WTF am I missing? What is it about?

OakMoses
11-10-2010, 02:39 AM
The bottom line for me is that, if he were shooting well, Dunleavy's offensive numbers would be fantastic, and I don't think 5 games is enough of a sample to pull a guy who's historically been a very good shooter. His FG% will rise substantially over the rest of the season. His defensive deficiencies will still exist, but right now it's the offense, not the defense, that is losing us games.

Prescience!