RealGM

Many on this board seem to think David West is trash.

Is the net rating just a fluke?

I don't think he's trash, but I think he's an abysmal fit for an uptempo team. And I have no idea how his offensive rating is 123.1, it defies logic. All that said, he's played 77 minutes for us. The stat that really jumps out is his OREB%, which is double his career #s and seems largely unsustainable. Also, why is he +29 here and -5 on bbref?

Weirdly enough, the same question should be asked about the #2 NetRtg guy, Andre Iguodala, who has drawn a ton of ire himself.

FNQ wrote:I don't think he's trash, but I think he's an abysmal fit for an uptempo team. And I have no idea how his offensive rating is 123.1, it defies logic. All that said, he's played 77 minutes for us. The stat that really jumps out is his OREB%, which is double his career #s and seems largely unsustainable. Also, why is he +29 here and -5 on bbref?

Weirdly enough, the same question should be asked about the #2 NetRtg guy, Andre Iguodala, who has drawn a ton of ire himself.

The ratings are the team ratings when they are on the court.

It benefits the players that do the "little things".

Ok that's vague though. I know what ORTG is on bbref, what is this sourced from, and why is there is 34 point difference from a stat that's relatively accurate?

FNQ wrote:Ok that's vague though. I know what ORTG is on bbref, what is this sourced from, and why is there is 34 point difference from a stat that's relatively accurate?

NBA.com. I don't think it is vague though. bbref's "personal" off/def rating is more useless.

http://stats.nba.com/team/#!/1610612744/players/advanced/?sort=NET_RATING&dir=1

The goal is to win a game. It is a team game. So if your team's net rating is good when you are on the court, you are doing something right.

It is like ESPN's team rating. You don't care how the team reaches the top 3 off or def rating. It just shows you the end result without showing you how you get there. bbref's personal rating I mostly use for entertainment purpose.

I am interested in your view if it is different.

That is still vague. I'll look it up though

The validity of stats is that it's formula reflects its purpose. Bbref's ratings aren't perfect, but if you consider it useless, you likely are using it incorrectly. It's is a quality marker in looking at players on the same team and comparing impacts. Not flawless, but extremely accurate most of the time

FNQ wrote:That is still vague. I'll look it up though

The validity of stats is that it's formula reflects its purpose. Bbref's ratings aren't perfect, but if you consider it useless, you likely are using it incorrectly. It's is a quality marker in looking at players on the same team and comparing impacts. Not flawless, but extremely accurate most of the time

I said it was "more" useless. It has value, like any other stat. One important thing it doesn't do, it doesn't tell you directly whether his team is winning or losing with him on the court.

I am surprised you have no idea what the team rating is about. You should look it up.

You want lineups that have a positive net ratings. A player with a good team net rating means he is on a lot of positive net rating lineups. If you want to win, that's what is important.

Re: formula - the team net rating has no vague formula. It is team points/points allowed per 100 possession. For people who know, there is some difference in terms of how a "possession" is counted, but it's the only "vague" thing, if any about team net rating.

Ok so I just read it.. Offensive rating is the number of points scored by a team per 100 possessions, defensive rating is the opposite. It's raw data that has nothing to regularize... well, anything. So how can it track the little things when it really tracks nothing other than basic raw data?

BBref actually accounts for the individual player's average contributions when factoring in their %. That's why it's a much better and easily far more accurate tool.

all that said, West is due regression to the mean with offensive rebounding and the fact that Iguodala is right up there with him. In such a limited amount of time, it can be attributed to many things - playing with one of the starters who is hot, playing when Ian Clark went buckwild.. or being held out of garbage time when teams usually catch up with us (the reason why guys like McCaw, McGee, and McAdoo are just so brutal) while the young backups take the statistical beating.

So overall, yeah, I think its definitely flukey, but its worth monitoring. It points to us either staggering our lineups in an effective way, or our 2nd units annihilating other teams' 2nd units. Because both Iguodala and West have are -5 in terms of Net Rating on bbref, and the advanced stats back that up.

likashing wrote:The goal is to win a game. It is a team game. So if your team's net rating is good when you are on the court, you are doing something right.

This was the part I was referring to as vague. It doesn't explain how the stat works, and so you know, the link doesn't help on mobile devices, so that's why I was asking for what the formula was. But its essentially raw data, so the only formula is pts per possession x 100, and pts scored against x 100.

FNQ wrote:

likashing wrote:The goal is to win a game. It is a team game. So if your team's net rating is good when you are on the court, you are doing something right.

This was the part I was referring to as vague. It doesn't explain how the stat works, and so you know, the link doesn't help on mobile devices, so that's why I was asking for what the formula was. But its essentially raw data, so the only formula is pts per possession x 100, and pts scored against x 100.

Yup, that's the beauty about it. There's no formula.

The thing about bbref's ratings are it takes into rebounds/steals etc etc...

The players that get empty rebounds (not-in-traffic) / steals while giving up driving lanes... it inflates the bbref rating.

On the flip side, the players that do the little things - box out / hockey assist / screen for open shots - they have a low bbref rating.

Points per 100 (or allowed) is not vague, it is cold-blooded-hard results staring in your face. It answers the question:

Does your team win when you are on the court?

FNQ wrote:It points to us either staggering our lineups in an effective way, or our 2nd units annihilating other teams' 2nd units.

bbref ratings don't adjust for playing against 2nd units either.

Staggering lineups effectively is the whole key obviously, it's not a game of one-on-one.

likashing wrote:The thing about bbref's ratings are it takes into rebounds/steals etc etc...

The players that get empty rebounds (not-in-traffic) / steals while giving up driving lanes... it inflates the bbref rating.

On the flip side, the players that do the little things - box out / hockey assist / screen for open shots - they have a low bbref rating.

Points per 100 (or allowed) is not vague, it is cold-blooded-hard results staring in your face. It answers the question:

Does your team win when you are on the court?

Yup, you've nailed the imperfections in it. It's not weighted. That's why I considered RAPM a significant step up because it closes the gap on things like this. But to be fair, the BBREF ORTG weighs scoring more than it does rebounds and steals.. and of course, PPP. Things like rebound rate and assist rate do change the formula a bit, but not significantly enough to skew the stat to unreliability.

I'd prefer it's flaws to that of raw data, simply because with raw data you essentially have to go through everything all over again - scoring, rebounding, defense, intangibles - before coming to something accurate. With BBREF's, you really only have to account for intangibles, and that to me makes it more accurate.

As for empty rebounds and steals while giving up lanes.. I think that is mostly outlier stuff. Especially the steals - if you get a high # of steals but in doing so cause your team to suffer, you get a slightly higher defensive rating than your court-mates but its baked into your defensive rating as well. But still, the rating gives you a good starting place of where to go and what to look for. It's great when paired with 82games.com's stats.. which are MIA this year

if you watch every game..why the heck do you need stats?

West has been awful to start the season..but he looked a little more comfortable in his role last night..His legs are still dead but maybe he can be useful just by being skilled and savvy.

FNQ wrote:BBref actually accounts for the individual player's average contributions when factoring in their %. That's why it's a much better and easily far more accurate tool.

Individual ORTG/DRTG on basketball-reference don't 'factor in' a player's individual contributions in the box score (I assume this is what you mean with % here) - it's exclusively based on it. Well, for defense it's a little trickier but it basically just spreads out equal value to all five players on the court for defensive events not tracked by the boxscore.

Simply put, it's a possession-based boxscore metric - i.e. how efficient would a player be if every possessions was ended by him per 100 possessions. Hence it's entirely different from the +/- approach towards which on/off NetRtg, on-court NetRtg, off/court NetRtg, APM, RAPM or RPM (in part) would be counted. So they aren't comparable without comparing boxscore-based metrics and PM-metrics in general. PM-metrics are much, much more valid in their approach as they try to measure actual impact - in fact it's the only valid approach generally speaking (of course adjustments are made in different metrics to address specific problems at the expense of pure validity). Their problem is reliability in small samples and arguably a lack of information about the portability of impact as it measures impact in team-specific contexts.

Therefore I'm not sure what you mean with more accurate. More accurate in factoring in boxscore events? Sure.

Re: OP's question. No, he hasn't been. Having the highest NetRtg obviously doesn't mean he's been the best or most impactful player on the Warriors. The sample size is a major concern but this isn't what the data tells us anyway. What it does indicate, however, is that West has been in extremely successful line-ups when he was on the court and that the Warriors, with him on the floor, did very well.

Was West the main reason for it? Most likely not. But it's hard to be 'trash' or a detriment to the team when the line-ups which include you have been highly successful. Sample size is certainly too small to draw any definite conclusion but a truly negative impact player - note: to this date, we don't know anything about the impact going forward - most likely would have been spotted already even in these small samples.

Little Digger wrote:if you watch every game..why the heck do you need stats?

Because your eyes are not able to obeserve every move all 10 players on the court do at any time and assess what exact effects they have in specific moments and the broader context? It's impossible. Stats exist as tools to work with in addition to what you perceive on the court while watching the game.

The-Power wrote:

FNQ wrote:BBref actually accounts for the individual player's average contributions when factoring in their %. That's why it's a much better and easily far more accurate tool.

Individual ORTG/DRTG on basketball-reference don't 'factor in' a player's individual contributions in the box score (I assume this is what you mean with % here) - it's exclusively based on it. Well, for defense it's a little trickier but it basically just spreads out equal value to all five players on the court for defensive events not tracked by the boxscore.

Simply put, it's a possession-based boxscore metric - i.e. how efficient would a player be if every possessions was ended by him per 100 possessions. Hence it's entirely different from the +/- approach towards which on/off NetRtg, on-court NetRtg, off/court NetRtg, APM, RAPM or RPM (in part) would be counted. So they aren't comparable without comparing boxscore-based metrics and PM-metrics in general. PM-metrics are much, much more valid in their approach as they try to measure actual impact - in fact it's the only valid approach generally speaking (of course adjustments are made in different metrics to address specific problems at the expense of pure validity). Their problem is reliability in small samples and arguably a lack of information about the portability of impact as it measures impact in team-specific contexts.

Therefore I'm not sure what you mean with more accurate. More accurate in factoring in boxscore events? Sure.

Re: OP's question. No, he hasn't been. Having the highest NetRtg obviously doesn't mean he's been the best or most impactful player on the Warriors. The sample size is a major concern but this isn't what the data tells us anyway. What it does indicate, however, is that West has been in extremely successful line-ups when he was on the court and that the Warriors, with him on the floor, did very well.

Was West the main reason for it? Most likely not. But it's hard to be 'trash' or a detriment to the team when the line-ups which include you have been highly successful. Sample size is certainly too small to draw any definite conclusion but a truly negative impact player - note: to this date, we don't know anything about the impact going forward - most likely would have been spotted already even in these small samples.

I was under the impression that BBREF's version uses the baseline that NBA.com uses as its starting point.. am I thinking of another site?

FNQ wrote:

The-Power wrote:

FNQ wrote:BBref actually accounts for the individual player's average contributions when factoring in their %. That's why it's a much better and easily far more accurate tool.

Individual ORTG/DRTG on basketball-reference don't 'factor in' a player's individual contributions in the box score (I assume this is what you mean with % here) - it's exclusively based on it. Well, for defense it's a little trickier but it basically just spreads out equal value to all five players on the court for defensive events not tracked by the boxscore.

Simply put, it's a possession-based boxscore metric - i.e. how efficient would a player be if every possessions was ended by him per 100 possessions. Hence it's entirely different from the +/- approach towards which on/off NetRtg, on-court NetRtg, off/court NetRtg, APM, RAPM or RPM (in part) would be counted. So they aren't comparable without comparing boxscore-based metrics and PM-metrics in general. PM-metrics are much, much more valid in their approach as they try to measure actual impact - in fact it's the only valid approach generally speaking (of course adjustments are made in different metrics to address specific problems at the expense of pure validity). Their problem is reliability in small samples and arguably a lack of information about the portability of impact as it measures impact in team-specific contexts.

Therefore I'm not sure what you mean with more accurate. More accurate in factoring in boxscore events? Sure.

Re: OP's question. No, he hasn't been. Having the highest NetRtg obviously doesn't mean he's been the best or most impactful player on the Warriors. The sample size is a major concern but this isn't what the data tells us anyway. What it does indicate, however, is that West has been in extremely successful line-ups when he was on the court and that the Warriors, with him on the floor, did very well.

Was West the main reason for it? Most likely not. But it's hard to be 'trash' or a detriment to the team when the line-ups which include you have been highly successful. Sample size is certainly too small to draw any definite conclusion but a truly negative impact player - note: to this date, we don't know anything about the impact going forward - most likely would have been spotted already even in these small samples.

I was under the impression that BBREF's version uses the baseline that NBA.com uses as its starting point.. am I thinking of another site?

If you're referring to the Individual ORTG/DRTG - which is the numbers you see next to the players in the 'Per 100 Possessions' section - then this is a metric invented by basketball-reference itself and isn't used anywhere else. ORTG/DRTG on stats.nba.com is always the team's ORTG/DRTG when a player is on and off the court respectively. BBRef has on/off-court ORTG/DRTG on their site as well, under the on/off section (can be found in the header of each player's page). Their numbers and the NBA's official numbers aren't congruent because they define possessions differently but the logic behind is exactly the same. Many people confuse Individual ORTG/DRTG (boxscore-based metric invented by BBRef) and actual on/off-court ORTG/DRTG, though, because of the misleading labels.

edit: It was actually invented by Dean Oliver of course (made popular by bbref, though). Sorry for any confusion. Here's the link how it's calculated: http://www.basketball-reference.com/about/ratings.html

The three most important sentences:

"In Dean's words, "Individual offensive rating is the number of points produced by a player per hundred total individual possessions. In other words, 'How many points is a player likely to generate when he tries?'""

"Out of necessity (owing to a lack of defensive data in the basic boxscore), individual Defensive Ratings are heavily influenced by the team's defensive efficiency."

"Introducing a concept he called "Skill Curves", he acknowledged that a player's ORtg needed to be judged in conjunction with his Usage Rate, a measure of how big a role the player fills in his team's offense."

ORTG/DRTG's are used in a lot of places and seem to mean different things on different sites. I'd have to look to remember where my favorite one was (EvanZ? 82games?).. lol haven't looked at this stuff in a couple years now. Anyways my favorite was a composite of the raw #s and then factors in the player's individual contributions and spits out an adjusted version of the traditional Net Rating that I found extremely accurate. Could have swore it was BBREF but I just saw their glossary, and yeah, individual stats based. I'm very meh on that

Not great....It's not that he's garbage...it just feels like he offers nothing defensively....and I would much rather have Looney in there.

I think if he focuses on the pick and pop game while he is on the floor...he can contribute.