Give it up for Marquis Daniels

[Roy H.]

Marquis only played 6 minutes of SF in the first game, and 10 minutes at SF in the second game.

For whatever it's worth, I had Marquis at about 8 non-Pierce minutes in Game 1, and 13.5 minutes with Paul out of the game in Game 2.  

If that's accurate, that means Marquis gave up 17 points to Lebron in 21.5 minutes (although, again, some of these points were off turnovers and techs).  Without Marquis at SF, Lebron scored 49 points in 65.5 minutes.

Lebron vs. Marquis: .79 points per minute, 44% shooting

Lebron vs. everyone else:  .75 points per minute, 45% shooting

Whatever the difference is there, it's pretty negligible, especially when you take into account that Marquis played many of those minutes beside Nate, Semih, and BBD, while Pierce got more time with the starters.

Points per minute is a strange measure to use Roy. In fact I don't think anyone really uses it. Points per possession and/or TS% are much better ones. They show a bigger difference than using per minute numbers do.

Marquis in: 1.161 PPP with TS% .679
Everyone else: .91 PPP with TS% .551

His assist rate does go down, I think he tries to attack Daniels more. 3/40 minutes versus 6.1/40 minutes.

Can I ask where you're getting your numbers?  I manually went through each possession when Marquis was in the game without Pierce, and came up with 38 total offensive possessions for Miami.  That's .447 points for Lebron per Miami offensive possession with Marquis in the game.

For your number to be accurate, Marquis would have had to be in there for only 14.6 of Miami's offensive possessions over 2 games.  Also, for Lebron to account for over a point for every one of his team's possessions would be unheard of.  

As for TS%, what relevance does Lebron's FT% (a factor in TS%) have to an assessment of defense?  It's not like Marquis or Paul is defending Marquis at the FT line.

I just used the numbers you tabulated for LeBron while Marquis was in and wasn't. I'm not talking about team efficiency.

EDIT:  To recap:

Games 1 & 2, Lebron's stats with Marquis playing SF:  17 points on 4-9 shooting from the field, with 2 assists and 3 turnovers 6-8 FTA

Games 1 & 2, Lebron's stats with anybody put Marquis playing SF: 49 points, 10 assists, 15-33 FG, 19-26 FTs, 9 turnovers

Italics adding in the free throw numbers

Quickie Points per possession (not Miami just LeBron)
Total points / (FGA + FTA * .44 + Turnovers)

I used TS% just to give another measure of scoring efficiency. Its not about LeBron’s free throw percentage, rather his free throw attempts. FT/FGA might be a better thing to look at as it doesn’t factor in how many he made.

FT/FGA and FT%
Marquis:    .88 FT per FGA  .67% FT
Everyone else: .78 FT per FGA 73% FT

He got a single tech while marquis was in, so we could remove that and they'd be almost the same. (though he probably got some tech free throws against others too)

I don't think that's how you calculate "per possession" stats (at least at an individual level).  You either need to look at all of Miami's possessions (actual or estimated), or at Lebron's individual possessions.  Unfortunately, there's not a mechanism for looking at LBJ's actual possessions, and what you proposed as a quicky method isn't legit at the individual level, because there are plenty of ways a "Lebron possession" can end other than with a shot attempt (made or missed), free throw attempt, or a turnover.  For instance, he can just pass the ball to Wade.

Regardless, I think that the difference between Pierce's teammates and Marquis' would make up for any difference.  

[Fafnir]

I don't think that's how you calculate "per possession" stats (at least at an individual level).  You either need to look at all of Miami's possessions (actual or estimated), or at Lebron's individual possessions.  Unfortunately, there's not a mechanism for looking at LBJ's actual possessions, and what you proposed as a quicky method isn't legit at the individual level, because there are plenty of ways a "Lebron possession" can end other than with a shot attempt (made or missed), free throw attempt, or a turnover.  For instance, he can just pass the ball to Wade.

This is exactly how you calculate possessions for an individual. A possession is only used up by a team with a made shot, a missed shot with a defensive rebound, a trip to the line, or a turnover. So if LeBron just passes to Wade, he hasn't used up one of his team's possessions. Instead he's just given the ball for someone else to use. He can easily get the ball right back from Wade and then dunk or turn it over.

Now this quickie method does not account for assists, so Rondo ranks poorly for example. And it also doesn't deal with offensive rebounds extending a possession, but it is valid.

Now you could say that he gets up less shot attempts due to effective defense.

He used up 15.52 poss.in 21 minutes, .73 uses per minute

He used up 53.44 poss. in 65 minutes, .82 uses per minute

Regardless, I think that the difference between Pierce's teammates and Marquis' would make up for any difference.  

This could very well be true.

[housecall]

Marquis is a decent backup defefnder...if you remember in the playoffs last season aganist the cavs it was a team effort to keep Lebron out of the paint as much as possible.I rarely see one player ever gaurding him throughout an entire game.The Mavericks did the best job on him so far this season of any team ive seen and their defense resembled what we did in last yrs.playoffs to slow him down,a collective effort.This seems to be a copycat league,especially if you want to be successful at slowing Lebron down.

[Roy H.]

I don't think that's how you calculate "per possession" stats (at least at an individual level).  You either need to look at all of Miami's possessions (actual or estimated), or at Lebron's individual possessions.  Unfortunately, there's not a mechanism for looking at LBJ's actual possessions, and what you proposed as a quicky method isn't legit at the individual level, because there are plenty of ways a "Lebron possession" can end other than with a shot attempt (made or missed), free throw attempt, or a turnover.  For instance, he can just pass the ball to Wade.

This is exactly how you calculate possessions for an individual. A possession is only used up by a team with a made shot, a missed shot with a defensive rebound, a trip to the line, or a turnover. So if LeBron just passes to Wade, he hasn't used up one of his team's possessions. Instead he's just given the ball for someone else to use. He can easily get the ball right back from Wade and then dunk or turn it over.

Now this quickie method does not account for assists, so Rondo ranks poorly for example, but it is valid.

Which sites / statisticians use this method?  I looked on Basketball Prospectus, Basketball Reference, and Wages of Wins, and didn't see anything like what you're suggesting. 

[Fafnir]

I don't think that's how you calculate "per possession" stats (at least at an individual level).  You either need to look at all of Miami's possessions (actual or estimated), or at Lebron's individual possessions.  Unfortunately, there's not a mechanism for looking at LBJ's actual possessions, and what you proposed as a quicky method isn't legit at the individual level, because there are plenty of ways a "Lebron possession" can end other than with a shot attempt (made or missed), free throw attempt, or a turnover.  For instance, he can just pass the ball to Wade.

This is exactly how you calculate possessions for an individual. A possession is only used up by a team with a made shot, a missed shot with a defensive rebound, a trip to the line, or a turnover. So if LeBron just passes to Wade, he hasn't used up one of his team's possessions. Instead he's just given the ball for someone else to use. He can easily get the ball right back from Wade and then dunk or turn it over.

Now this quickie method does not account for assists, so Rondo ranks poorly for example, but it is valid.

Which sites / statisticians use this method?  I looked on Basketball Prospectus, Basketball Reference, and Wages of Wins, and didn't see anything like what you're suggesting. 

They all use true shooting don't they? Points per possession as I've outlined is just a modified TS% (it considers turnovers the same as a missed shot)

Wage of wins does use something similar all the time by using points per shot as part of his comparison of two players. FT/FGA is used by Dean Oliver for teams and lots of sites measure free throw rate as well.

Hollinger uses something similar that actually acounts for assists to calculate usage.

Usage Rate - the number of possessions a player uses per 40 minutes. Usage Rate = {[FGA + (FT Att. x 0.44) + (Ast x 0.33) + TO] x 40 x League Pace} divided by (Minutes x Team Pace)

He just adds assists with a .33 weight and then adjusts for pace. So if you want we could account for assists in that way. It is not pace adjusted, but I don't think you're objecting that it doesn't account for the pace of the game.

Basketball reference:

Usg%
Usage Percentage (available since the 1977-78 season in the NBA); the formula is 100 * ((FGA + 0.44 * FTA + TOV) * (Tm MP / 5)) / (MP * (Tm FGA + 0.44 * Tm FTA + Tm TOV)). Usage percentage is an estimate of the percentage of team plays used by a player while he was on the floor.

This uses my exact formulation but for a estimate of possesions used in total by a player as a percentage of his teams. (it uses minutes played to weight things properly)

FGA + FTA * .44 + TOV is a well accepted formula for possessions used.

[gar]

People are trying to use numbers to hide the fact that they have been bad mouthing a perfectly good player for no good reason. His performance this year should not surprise anyone.

[chicagogreen]

i really like that lineup with no pg.  let marquis and pp, deal with the other teams attempts at pressure and lets see if they can deal with a lineup of that size, wont work every night, but if they can't disrupt our ball handling I'd love to see that lineup for 5 or 10 minutes a night, reminds me of the old bulls playing with no point gaurd for long stretches, adds huge length to the back court, and often really messed up the opposing pg play. 

[Roy H.]

I don't think that's how you calculate "per possession" stats (at least at an individual level).  You either need to look at all of Miami's possessions (actual or estimated), or at Lebron's individual possessions.  Unfortunately, there's not a mechanism for looking at LBJ's actual possessions, and what you proposed as a quicky method isn't legit at the individual level, because there are plenty of ways a "Lebron possession" can end other than with a shot attempt (made or missed), free throw attempt, or a turnover.  For instance, he can just pass the ball to Wade.

This is exactly how you calculate possessions for an individual. A possession is only used up by a team with a made shot, a missed shot with a defensive rebound, a trip to the line, or a turnover. So if LeBron just passes to Wade, he hasn't used up one of his team's possessions. Instead he's just given the ball for someone else to use. He can easily get the ball right back from Wade and then dunk or turn it over.

Now this quickie method does not account for assists, so Rondo ranks poorly for example, but it is valid.

Which sites / statisticians use this method?  I looked on Basketball Prospectus, Basketball Reference, and Wages of Wins, and didn't see anything like what you're suggesting. 

They all use true shooting don't they? Points per possession as I've outlined is just a modified TS% (it considers turnovers the same as a missed shot)

Wage of wins does use something similar all the time by using points per shot as part of his comparison of two players. FT/FGA is used by Dean Oliver for teams and lots of sites measure free throw rate as well.

Hollinger uses something similar that actually acounts for assists to calculate usage.

Usage Rate - the number of possessions a player uses per 40 minutes. Usage Rate = {[FGA + (FT Att. x 0.44) + (Ast x 0.33) + TO] x 40 x League Pace} divided by (Minutes x Team Pace)

He just adds assists with a .33 weight and then adjusts for pace. So if you want we could account for assists in that way. It is not pace adjusted, but I don't think you're objecting that it doesn't account for the pace of the game.

Basketball reference:

Usg%
Usage Percentage (available since the 1977-78 season in the NBA); the formula is 100 * ((FGA + 0.44 * FTA + TOV) * (Tm MP / 5)) / (MP * (Tm FGA + 0.44 * Tm FTA + Tm TOV)). Usage percentage is an estimate of the percentage of team plays used by a player while he was on the floor.

This uses my exact formulation but for a estimate of possesions used in total by a player as a percentage of his teams. (it uses minutes played to weight things properly)

FGA + FTA * .44 + TOV is a well accepted formula for possessions used.

I'm sorry, man, but not a single one of those sites use "points per possession" like you did.  I know what usage is -- which isn't what you originally were talking about -- but that's a lot different than saying "Marquis allows 1.1 points per possession", using imaginary numbers.

Anyway, I'm done with this particular conversation because I'm sure I'm right, and I'm sure you equally believe your right.  We can agree to disagree, but I would ask if your methodology is 1) useful and 2) accurate, why does no site use the method you do to prove "points allowed per possession"?   

I mean, this whole conversation started with you telling me that points per minute wasn't a valid stat (despite just about every site, including 82games and basketball-reference, use it).  Now, you're using numbers in a way that nobody else does.  It just doesn't make a lot of sense to me, and I get the feeling that you can type until you're blue in the face, and we're still not going to see eye-to-eye here.

At the end of the day, there's absolutely nothing that suggests that Marquis is demonstrably worse covering Lebron than the rest of the team is, especially when you adjust for quality of teammate.

[Fafnir]

I'm sorry, man, but not a single one of those sites use "points per possession" like you did.  I know what usage is -- which isn't what you originally were talking about -- but that's a lot different than saying "Marquis allows 1.1 points per possession", using imaginary numbers.

Huh?

So basketball reference uses my exact formulation of possessions for usage (which is defined as an estimate of possesions used), weighted for minutes played, and its not used anywhere or accurate?

Its similar to opposing PER, which just looks at your counterparts PER. When LeBron James is at SF against Daniels at SF those are the efficiency numbers. Sites don't break down one on one match ups like that. (synergy does but I don't have access)

I think points per minute is weird way of looking at things. My comment on it being a strange measure was more about your formulation of it in points per minute form. Typically its points/36, points/40, or points per game.

Overall I think points per possession, points per shot, and TS% are better measures to look at. They showed a different side to the story than the scoring per minute(or per 36 minutes). I think efficiency matters more than your rate of scoring.

LeBron is going to take his shots, and he's going to score. They key to effectively defending him is in keeping him from being efficient. (same with Kobe)

Honestly we're talking about 21 minutes and 65 minute sample sizes, none of our "numbers" are useful or accurate. Otherwise we could clearly conclude that the C's have a plan to make Dwayne Wade suck. Which we clearly don't, he's beat up on the C's before.

I get that you disagree with me on Daniels, but your rejection of estimating posessions used strikes me as strange.

[Roy H.]

I lied.  I'm going to take another stab at this.

Here was the original post I objected to:

Points per minute is a strange measure to use Roy. In fact I don't think anyone really uses it. Points per possession and/or TS% are much better ones. They show a bigger difference than using per minute numbers do.

Marquis in: 1.161 PPP with TS% .679
Everyone else: .91 PPP with TS% .551

First, I agree that "per possession" stats are more accurate than "per minute" stats.  However, those possessions -- like offensive rating, etc. -- are *team* possessions.  Possessions, rather than minutes, account for things like pace.  (As an aside, I think "per minute" stats are probably pretty valid when you're comparing one player's production in the same games against the same team, but in general I agree with the point.)

The point of looking at points per minute -- or possession -- is to determine how productive Lebron is against Marquis, or against players other than Marquis, during a finite interval.  Obviously, the lower the production output, the better for the defender.

By using your formula, and framing things solely in terms of efficiency, you're losing a large part of the picture.  For instance, take the following two examples:

A)  Lebron shoots 8-for-10 and has 0 turnovers in 17 minutes, scoring 16 points;

B)  Lebron shoots 1-for-1 and has 0 turnovers in 17 minutes, scoring 2 points.

Using your methodology, example "B" looks better.

A)  Total points / (FGA + FTA * .44 + Turnovers) = 16 / (10 + 0 + 0) = 1.6 PPP

B)  Total points / (FGA + FTA * .44 + Turnovers) = 3 / (1 + 0 + 0) = 2.0 PPP

Now, can anybody actually argue that in example B, Lebron was better than in example A?  Of course not.

However, if you look at points per minute, you get:

A) Total points / minutes = 16 / 17 = .941 points per minute

B) Total points / minutes = 2 / 17 = .118 points per minute

Obviously, this second example is a much, much more accurate way to assess how Lebron's defender was doing.  Now, looking at points allowed per *team* possession to account for fluctuations in pace would be the most accurate way, of course, but that's a lot different than framing the argument in terms of *individual* possessions.

I know one of your pet peeves is "moving the goal posts", but that's exactly what you did.  You started out using this "PPP" -- points per possession -- methodology, and now you're arguing that you're using basketball-reference's calculation regarding usage.  You haven't just moved the goal posts, you've totally constructed new ones in a stadium across town.

[Fafnir]

I lied.  I'm going to take another stab at this.

Here was the original post I objected to:

Points per minute is a strange measure to use Roy. In fact I don't think anyone really uses it. Points per possession and/or TS% are much better ones. They show a bigger difference than using per minute numbers do.

Marquis in: 1.161 PPP with TS% .679
Everyone else: .91 PPP with TS% .551

First, I agree that "per possession" stats are more accurate than "per minute" stats.  However, those possessions -- like offensive rating, etc. -- are *team* possessions.  Possessions, rather than minutes, account for things like pace.  (As an aside, I think "per minute" stats are probably pretty valid when you're comparing one player's production in the same games against the same team, but in general I agree with the point.)

The point of looking at points per minute -- or possession -- is to determine how productive Lebron is against Marquis, or against players other than Marquis, during a finite interval.  Obviously, the lower the production output, the better for the defender.

By using your formula, and framing things solely in terms of efficiency, you're losing a large part of the picture.  For instance, take the following two examples:

A)  Lebron shoots 8-for-10 and has 0 turnovers in 17 minutes, scoring 16 points;

B)  Lebron shoots 1-for-1 and has 0 turnovers in 17 minutes, scoring 2 points.

Using your methodology, example "B" looks better.

A)  Total points / (FGA + FTA * .44 + Turnovers) = 16 / (10 + 0 + 0) = 1.6 PPP

B)  Total points / (FGA + FTA * .44 + Turnovers) = 3 / (1 + 0 + 0) = 2.0 PPP

Now, can anybody actually argue that in example B, Lebron was better than in example A?  Of course not.

However, if you look at points per minute, you get:

A) Total points / minutes = 16 / 17 = .941 points per minute

B) Total points / minutes = 2 / 17 = .118 points per minute

Obviously, this second example is a much, much more accurate way to assess how Lebron's defender was doing.  Now, looking at points allowed per *team* possession to account for fluctuations in pace would be the most accurate way, of course, but that's a lot different than framing the argument in terms of *individual* possessions.

This argument is really about how important is usage, I agree that your role and usage can be factors roy. But in our two minute samples LeBron is taking shots and using possessions at a very similar rate.

He used up 15.52 poss.in 21 minutes, .73 uses per minute

He used up 53.44 poss. in 65 minutes, .82 uses per minute

This isn't a case where the numbers are skewed like you suppose.

I mean what you say is true you can't use efficiency and say Perkins is a better offensive player than KG just because he has a higher eFG% and TS% last year. But we're not comparing two players in different roles. LeBron is LeBron.

I'm using basketball reference usage not to move the goal posts Roy but to show you that my estimate for possessions is a reasonable one. Most sites stick with TS% for efficiency, I like to use my own PPP because I think turnovers should be counted that's all. I mean a player can have a very "efficient" game by TS% but turn it over 5 times and still have been a net negative to his team when it comes to efficiency.

Edit: The easy way to deal with the issue you raise is look at both how many points you score and the efficiency that its done with. You point out that he scores at roughly the same rate (5% different) on Marquis as he does Pierce. But he also is scoring those same points with greater efficiency.

[Rondo2287]

At this point we arent talking about basketball anymore we are talking about fractions.

[Lucky17]

At this point we arent talking about basketball anymore we are talking about fractions.

Fractions?! We're talking about fractions?! Not a game, not a game, but fractions?!