Rondo vs the other elite PGs: by the numbers

[drza44]

  Ok, a few things: you said that point guard was the least important offensive position. When Rondo's on the court he's not our least important defensive player. And, yes, they do have those on/off numbers on 82games. They don't really take into account what lineups are in the game and whether you're going against starters or reserves. It's also a fairly noisy stat, by the way. There are plenty of advanced stats that have Rondo in the top 10-20 defenders in the league. He holds opposing players to among the lowest scoring efficiency every year and he forces a lot of turnovers in addition to being a good rebounder.

  I'd also like to point out that you're claiming that a point guard's defense is less important than his offense because of the impact of his defense on the team, while at the same time you're comparing point guards to each other based mainly on individual offensive production. What's a bigger impact, Westbrook having a higher PER and a higher offensive win share or Rondo directing the team with the best eFG% in the league?

1) I don't believe I said that PG was the least important offensive position, and if I did that was a mis-statement on my part because PGs can obviously have a HUGE impact on team offense.  I was saying that point guard was not as important a position DEFENSIVELY, not offensively.  And I stand by that, even on our team.  The big man slots are MUCH more important to our defense than PG.  Then, there's a fairly even split between the 3 perimeter slots.  The PGs in our defense are tasked more with ball-disruption and steals, but the wings generally have to play better 1-on-1 perimeter D against more elite scorers.  Without looking it up, I'd bet that Rondo's and Pierce's on/off defensive stats from the last few years would look similar, solidly behind Perk, who himself is solidly behind KG.

2) Again, I agree that Rondo is a good defensive point guard.  I've never argued differently.  What I've argued is that the impact he can have on a team's defense is smaller than what a big can have by definition, and thus his defensive impact can't be large enough to overcome a dominant offensive advantage like Paul or Nash have over him this year.  And I guarantee you that none of the stats you refer that have Rondo anywhere near the top-10 or 20 are looking at team impact.  As I mentioned in my last post, something like a defensive win shares or defensive rating would be good to Rondo because they're math models that rely heavily on steals and the other box score stats instead of looking at actual impact. 

Likewise, I could believe that Rondo holds his individual assignment to lower than usual efficiency and that he generates a lot of possession changes for a guard.  But again, what makes bigs dominant defenders is that they hold down not only their own man, but they do a lot of helping out in the paint as well.  Look at Perk, for example, there was an article on this site last week talking about how much he limits opposing post players 1-on-1...but he's also one of our key help defenders in the paint, on a level that Rondo isn't.  You could argue that Rondo is better among PG defenders than Perk is among big man defenders, but that's a distinction without value.  At the end of the day, Perk's going to be the more important defender simply because of his size and his role than Rondo. 

And again, that's no shame on Rondo or his D, it's an indication that at his size/position what he can accomplish on defense just isn't as much as what a) big men can accomplish on defense or b) perimeter players can accomplish on offense.

3) As for your last point, my whole reason behind the OP was the exact opposite of what you state.  My point was that something like PER, though widely used/accepted these days, focuses too much on individual offensive production (specifically volume scoring) to be an accurate stand-alone measure of value.  That is specifically why I look at the other 4 stats, 1 of which is purely +/- based, 2 of which specifically include a defensive component, and a 4th (wins produced) who's biggest criticism is that it UNDER-emphasizes individual scoring/offense.  If we just looked at PER and Win Shares alone then yes, Westbrook would have a clear advantage over Rondo.  But looking across many statistical angles, as I suggest here, we see that on the contrary Rondo is actually likely the slightly more valuable of the 2 though both are still excellent players with arguments for top-5 in the league.  To me, that's both a more accurate and a more intuitive conclusion than what you'd get looking at just PER or watching Sportscenter highlights.  Especially for an NBA fan that isn't a huge fan of either the Celtics or the Thunder, this would be a good way to supplement the handful of times you might get to see them play with numbers that detail different aspects of the strengths/weaknesses in their games.

[drza44]

Doc claims the PG pressure is extremely important to the defense so that a lot of time is taken off the shot clock before the other team gets into their offense.

In other words, the PG could mean 5 or 6 less seconds of playing defense per possession for teammates. This also means more bad shots by opposing teams and less defensive breakdowns.

While this does not in itself mean that Rondo is necessarily a defensive lynch pin, it does point to the potential high impact of the PG.

Add to that the amount of time that the opposing PG has the ball and needs to be defended and pressured to disrupt the offense and the prevalence of pick and roll plays w/ PGs and I question why PGs should be any less significant than other players on defense.

1) Her's a link to an article from a guy that tracked every possession of every game in the 2010 postseason while looking for what he calls "defensive usage".  http://elgee35.wordpress.com/2011/01/22/defensive-usage/ The key take-home from the article is that about 14% of possessions ended in turnovers, about 30% ended in fast break attempts or intentional fouls (end of games) or things like that, and about 56% ended in contested shot attempts/FTs.  Of those contested shots, point guards were defending on about half as many shots (13%) as PFs (24%) or centers (31%).  The big guys, because they help at the rim, are just in on trying to prevent a lot more shots than the perimeter guys are.  Now, the article isn't perfect and I'm not endorsing this as another definite stat, I'm merely using his findings to point out that the intution is true: big guys face a lot more shots, and also a lot more high-percentage shots than little guys do, which is why in general big guys are a lot more important to a team's defense than little guys.

2) Again, it seems as though I need to clarify the difference between "Rondo's not a good defender" (which I'm not saying) and "Rondo's good defense isn't as important to his impact as his or other great point guardss offensive impact" (which I am saying).

Let's look at an example.  The Spurs have been a great defense for the last decade.  A great defense has many important contributors, all of which have a part to play.  Defensive scheme is huge.  Having great perimeter defenders is huge.  But having an elite defensive big is the most important piece.  So you take a guy like Bruce Bowen, he was an EXCELLENT defender.  One of the best perimeter wings every year, perennial all defense, challenged for DPoY a couple times.  Bowen not only did a great job on his own man, but he also made everyone else's life easier on defense.  He was a key player.  But.

If you took that exact Spurs defense and removed Bowen, they would still be a very good defensive team.  Maybe they drop from #1 to top-5, but they're still relatively the same defense.  But if you removed Duncan, on the other hand, the drop would be more dramatic.  Bowen was an excellent defender, but (using the estimate from that article) Duncan was about twice as important.  Bowen could impact about 15 - 20% of the defense, but replacing him with a replacement level player could only impact that 15 - 20%.

Meanwhile, a guy like Nash has his hands in almost everything that the Suns offense does.  If we were putting numbers to it, Nash is more like 50% of the Suns' offense than the 20% Bowen might mean to the Spurs' defense.  As the point guard, Nash could make a HUGE difference vs a replacement level guy replacing him.

Now, bring it back to Rondo.  Suppose for the sake of argument you say he's a Bowen-level defender.  At the end of the day, it's his offense that gives him the lion-share of his value to this team.  This year the Celtics are +11 offensively with Rondo on the court, and -3 defensively.  Forget even the sign of the number, and just look at the magnitude.  Rondo's making his big impact on this team OFFENSIVELY, much more than defensively.

And we saw it when Rondo got hurt.  The defense and their production look very similar during the games when Rondo was out.  Our offense, on the other hand, stagnated without our point guard.  We survived because we have talented individual offensive players besides him, but the assists went way down, the pace stagnated, and the offense as a whole just looked a lot worse without Rondo.  THAT'S where he's making his biggest impact.  Yes, he's a good defender, but even if he were Bruce Bowen the impact that he's making on defense just isn't as big as the impact that he's making on offense.

Does his defense make a difference?  Sure, to both his team and to himself in individual rankings.  But it's a smaller component than his offensive foundation, so when you compare him to someone like Nash or Paul that have big offensive advantages, Rondo's defense just isn't enough to bridge the gap pretty much no matter how you analyze it.

[BballTim]

  Ok, a few things: you said that point guard was the least important offensive position. When Rondo's on the court he's not our least important defensive player. And, yes, they do have those on/off numbers on 82games. They don't really take into account what lineups are in the game and whether you're going against starters or reserves. It's also a fairly noisy stat, by the way. There are plenty of advanced stats that have Rondo in the top 10-20 defenders in the league. He holds opposing players to among the lowest scoring efficiency every year and he forces a lot of turnovers in addition to being a good rebounder.

  I'd also like to point out that you're claiming that a point guard's defense is less important than his offense because of the impact of his defense on the team, while at the same time you're comparing point guards to each other based mainly on individual offensive production. What's a bigger impact, Westbrook having a higher PER and a higher offensive win share or Rondo directing the team with the best eFG% in the league?

1) I don't believe I said that PG was the least important offensive position, and if I did that was a mis-statement on my part because PGs can obviously have a HUGE impact on team offense.  I was saying that point guard was not as important a position DEFENSIVELY, not offensively.  And I stand by that, even on our team.  The big man slots are MUCH more important to our defense than PG.  Then, there's a fairly even split between the 3 perimeter slots.  The PGs in our defense are tasked more with ball-disruption and steals, but the wings generally have to play better 1-on-1 perimeter D against more elite scorers.  Without looking it up, I'd bet that Rondo's and Pierce's on/off defensive stats from the last few years would look similar, solidly behind Perk, who himself is solidly behind KG.

  Ok, it was late. I meant that you said that point guard was the least important position defensively, which I don't think is the case. BTW, I checked Rondo and Perk's on/off stats for the prior 3 seasons, and the difference was about .5 points per 100 possessions. Again, those numbers don't take things into account like lineups and opponents.

2) Again, I agree that Rondo is a good defensive point guard.  I've never argued differently.  What I've argued is that the impact he can have on a team's defense is smaller than what a big can have by definition, and thus his defensive impact can't be large enough to overcome a dominant offensive advantage like Paul or Nash have over him this year.  And I guarantee you that none of the stats you refer that have Rondo anywhere near the top-10 or 20 are looking at team impact.  As I mentioned in my last post, something like a defensive win shares or defensive rating would be good to Rondo because they're math models that rely heavily on steals and the other box score stats instead of looking at actual impact. 

  I'd just like to note that, when gathering your stats, you claimed that you don't like adjusted plus/minus because it's too noisy. You're basing this argument on *unadjusted* numbers.

Likewise, I could believe that Rondo holds his individual assignment to lower than usual efficiency and that he generates a lot of possession changes for a guard.  But again, what makes bigs dominant defenders is that they hold down not only their own man, but they do a lot of helping out in the paint as well.  Look at Perk, for example, there was an article on this site last week talking about how much he limits opposing post players 1-on-1...but he's also one of our key help defenders in the paint, on a level that Rondo isn't.  You could argue that Rondo is better among PG defenders than Perk is among big man defenders, but that's a distinction without value.  At the end of the day, Perk's going to be the more important defender simply because of his size and his role than Rondo. 

And again, that's no shame on Rondo or his D, it's an indication that at his size/position what he can accomplish on defense just isn't as much as what a) big men can accomplish on defense or b) perimeter players can accomplish on offense.

  Somewhat true, but that's only in the halfcourt setting. The Celts are always a top team in transition defense and forcing turnovers. Those are important to defensive efficiency, and I'd argue that Rondo's more responsible for those areas than the bigs are.

3) As for your last point, my whole reason behind the OP was the exact opposite of what you state.  My point was that something like PER, though widely used/accepted these days, focuses too much on individual offensive production (specifically volume scoring) to be an accurate stand-alone measure of value.  That is specifically why I look at the other 4 stats, 1 of which is purely +/- based, 2 of which specifically include a defensive component, and a 4th (wins produced) who's biggest criticism is that it UNDER-emphasizes individual scoring/offense.  If we just looked at PER and Win Shares alone then yes, Westbrook would have a clear advantage over Rondo.  But looking across many statistical angles, as I suggest here, we see that on the contrary Rondo is actually likely the slightly more valuable of the 2 though both are still excellent players with arguments for top-5 in the league.  To me, that's both a more accurate and a more intuitive conclusion than what you'd get looking at just PER or watching Sportscenter highlights.  Especially for an NBA fan that isn't a huge fan of either the Celtics or the Thunder, this would be a good way to supplement the handful of times you might get to see them play with numbers that detail different aspects of the strengths/weaknesses in their games.

  Again, though, you're finding different ways to measure individual output on offense, but you're by and large ignoring the impact of the player on the TEAM. On defense, you're arguing the opposite. You're skewing your arguments to lower Rondo's value, even if it's not your intention. For instance, in the 11 games when Rondo was out Ray shot about 35% on threes, but he's hitting about 50% of them when Rondo plays. Which of your stats accounts for that?

 

[drza44]

  Ok, it was late. I meant that you said that point guard was the least important position defensively, which I don't think is the case. BTW, I checked Rondo and Perk's on/off stats for the prior 3 seasons, and the difference was about .5 points per 100 possessions. Again, those numbers don't take things into account like lineups and opponents.

I'd just like to note that, when gathering your stats, you claimed that you don't like adjusted plus/minus because it's too noisy. You're basing this argument on *unadjusted* numbers.

Somewhat true, but that's only in the halfcourt setting. The Celts are always a top team in transition defense and forcing turnovers. Those are important to defensive efficiency, and I'd argue that Rondo's more responsible for those areas than the bigs are.

I'm grouping these together to respond to because I fear that in our growing discussion we could lose sight of the main point where we initially disagreed, namely whether Rondo's defense is enough to bridge the gap for a player like Nash's offensive advantage over him.  As you point out, our discussion was devolving a bit more into generalities and unadjusted numbers to make more encompassing points when we have specific numbers we could look at.

Re: +/- stats.  Yes, adjusted +/- for a given year is very noisy.  But over time things get a lot less noisy, and the story comes out.  I found a 5-year APM calculation on the APBRmetrics board (2005 - 2010), and there's another 6 year APM calculation publicly available (2003 - 2009, https://spreadsheets.google.com/ccc?key=0AnGzTFTtSPx_dE9xcTVITjhRSDdfRkJ5MHJ2cU1nb0E&hl=en#gid=0 ) that each help get the point across for these specific players.

For both APMs, Nash measures out with by-far the biggest offensive adjusted +/- at any position in the NBA (+8.05 in the 5-year study, +8.84 in the 6-year study).  On the flip-side, Rondo's defensive APM in both studies was about +2, and in both cases he was right around the best measured PG in the league and ranked in the 50 - 70 range in the NBA overall on defense. 

I didn't include these types of multi-year comparisons in the OP because we're looking specifically at this year, but from the 2 +/- based stats I did include in the OP it's pretty clear that the same trend would show up this year as well.  Nash's offensive impact is much larger than anything Rondo could do defensively, and in fact even if you look at the sum impact that includes both offense and defense Nash measures out higher.  Which has been my point all along, though sometimes that can get a bit hidden in the analogies and generalizations as I tried to think of better ways to emphasize that.

  Again, though, you're finding different ways to measure individual output on offense, but you're by and large ignoring the impact of the player on the TEAM. On defense, you're arguing the opposite. You're skewing your arguments to lower Rondo's value, even if it's not your int ention. For instance, in the 11 games when Rondo was out Ray shot about 35% on threes, but he's hitting about 50% of them when Rondo plays. Which of your stats accounts for that? 

Both the adjusted +/- stat I included and the Roland rating that is based in part of net +/- should account for that.  Both of those stats consider the impact of the p layer on the team offense...in fact, they're the same stats that also help account for the player's impact on team defense.  In fact, I'm a bit confused now as to where you're coming from with this particular rebuttal point...how could I be using the same +/- based stats to argue in opposite ways for team defense and offense?  The conclusions might change based on what the stat tells us, but the methodology for the analysis is the same.

[jimmywolfrey]

maybe there should be a weighted ranking system like the BCS? Hollinger could be one of the polls used.

[jimmywolfrey]

This probably affect the top 5 or 6 positions but what about not limiting the players in the pool and instead use a 10 to 1 point system with any player outside the top 10 getting a zero? 

I'm sure in some of the rating systems, John Wall and others wouldn't rank in the top 10.

[winsomme]

Doc claims the PG pressure is extremely important to the defense so that a lot of time is taken off the shot clock before the other team gets into their offense.

In other words, the PG could mean 5 or 6 less seconds of playing defense per possession for teammates. This also means more bad shots by opposing teams and less defensive breakdowns.

While this does not in itself mean that Rondo is necessarily a defensive lynch pin, it does point to the potential high impact of the PG.

Add to that the amount of time that the opposing PG has the ball and needs to be defended and pressured to disrupt the offense and the prevalence of pick and roll plays w/ PGs and I question why PGs should be any less significant than other players on defense.

they're definitely not less significant. I just think these advanced stats can take us further from the actual game rather than helping us understand what is actually happening on the court.

I think you have to pass the eye test first and for me, watching the games, it's pretty clear that Rondo has a big impact on the defensive end.

So stats that say otherwise are flawed and why they are flawed is less important to me.

If you want to criticize Rondo's game why go to these advanced stats when all we need to look at is a good old standby FT%...

I can't for the life on me understand why he struggles so much at the line. Other than that, Rondo is aces. The guy is electric and has off the charts court vision.

[Vermont Green]

Thank you Drza44 for your excellent and neutral analysis of PG's  Unfortunately you are swimming upstream against a current of opinion that I believe overrates Rondo as a PG.

I think your point about the Spurs being a top defense even though they have Tony Parker as their PG is very compelling. It should make people think.  Parker is a far better offensive player and with his shooting range, opens up shots for others, even if he doesn't record an assist.  Plus apparently his defense is not hurting the team all that much.

That is the metric that is missing.  What is the impact on the Celtics team offense when Rondo's man is playing 10 feet off him and providing an instant double whenever he wants.

The Celtics are a good team and Rondo is a fine PG but there are 5 or 6 PGs that would make the Celtics even better, just as your statistical analysis indicates.

[winsomme]

Thank you Drza44 for your excellent and neutral analysis of PG's  Unfortunately you are swimming upstream against a current of opinion that I believe overrates Rondo as a PG.

I think your point about the Spurs being a top defense even though they have Tony Parker as their PG is very compelling. It should make people think.  Parker is a far better offensive player and with his shooting range, opens up shots for others, even if he doesn't record an assist.  Plus apparently his defense is not hurting the team all that much.

That is the metric that is missing.  What is the impact on the Celtics team offense when Rondo's man is playing 10 feet off him and providing an instant double whenever he wants.

The Celtics are a good team and Rondo is a fine PG but there are 5 or 6 PGs that would make the Celtics even better, just as your statistical analysis indicates.

unfortunately statistical analysis can't run an offense or see open PP cutting to the basket...

[Surferdad]

This is the never-ending argument about who is the "best" PG.  Frankly, I think Rose is better than Rondo, but Rondo is better for this team.

[BballTim]

Doc claims the PG pressure is extremely important to the defense so that a lot of time is taken off the shot clock before the other team gets into their offense.

In other words, the PG could mean 5 or 6 less seconds of playing defense per possession for teammates. This also means more bad shots by opposing teams and less defensive breakdowns.

While this does not in itself mean that Rondo is necessarily a defensive lynch pin, it does point to the potential high impact of the PG.

Add to that the amount of time that the opposing PG has the ball and needs to be defended and pressured to disrupt the offense and the prevalence of pick and roll plays w/ PGs and I question why PGs should be any less significant than other players on defense.

1) Her's a link to an article from a guy that tracked every possession of every game in the 2010 postseason while looking for what he calls "defensive usage".  http://elgee35.wordpress.com/2011/01/22/defensive-usage/ The key take-home from the article is that about 14% of possessions ended in turnovers, about 30% ended in fast break attempts or intentional fouls (end of games) or things like that, and about 56% ended in contested shot attempts/FTs.  Of those contested shots, point guards were defending on about half as many shots (13%) as PFs (24%) or centers (31%).  The big guys, because they help at the rim, are just in on trying to prevent a lot more shots than the perimeter guys are.  Now, the article isn't perfect and I'm not endorsing this as another definite stat, I'm merely using his findings to point out that the intution is true: big guys face a lot more shots, and also a lot more high-percentage shots than little guys do, which is why in general big guys are a lot more important to a team's defense than little guys.

  The article says:
__________________________________________________________________________________

  This means that the range of defensive impacts will be much wider for interior players than perimeter plays. The difference between a good and bad defensive point guard won’t be nearly as pronounced as the difference between a good and bad defensive center. Using a simple mathematical example, imagine the following:

    * A “Bad” Defender allows 50% eFG shooting
    * A “Good” Defender allows 40%e FG shooting

Assume free throw accuracy is a constant (the league average). Based on these shooting percentages, at the center position the difference between our bad defender and good defender is 2.9 pts/100 (a difference in efficiency this year between an average team and the 8th-best team). But at PG, the difference between our good and bad defenders shrinks to 1.2 pts/100.
_______________________________________________________________________________

  I agree with all of this. But consider Rondo in particular. He gets about 2 steals per 100 possessions more than the average guard, 3 more than a bad defender. If you include turnovers forced the gulf is probably wider. That would push his impact on the game at least into the same range as the impact difference between a good and bad defensive center.
  

[BballTim]

Thank you Drza44 for your excellent and neutral analysis of PG's  Unfortunately you are swimming upstream against a current of opinion that I believe overrates Rondo as a PG.

I think your point about the Spurs being a top defense even though they have Tony Parker as their PG is very compelling. It should make people think.  Parker is a far better offensive player and with his shooting range, opens up shots for others, even if he doesn't record an assist.  Plus apparently his defense is not hurting the team all that much.

  Tony Parker isn't really a good outside shooter and he's a somewhat average defender.

That is the metric that is missing.  What is the impact on the Celtics team offense when Rondo's man is playing 10 feet off him and providing an instant double whenever he wants.

The Celtics are a good team and Rondo is a fine PG but there are 5 or 6 PGs that would make the Celtics even better, just as your statistical analysis indicates.

  The metric is there. Rondo's leading the league in assists by a good rate, the Celts hit a much higher percentage of their shots when he's in the game and they have (I'm pretty sure) the highest percentage of assisted baskets of any team in the league when he plays. I don't agree with your assessment that someone who can shoot better than Rondo but can't pass or defend or control the pace of the game as well will make the Celts better.

[greenpride32]

The fact that we need 20 different threads to disucss why Rondo is an elite PG only proves one thing; that he's not.

When Michael played nobody debated whether he was the best or not, it was universally known and accepted that he was.  Guys like Magic and Stockton in their prime were always regarded as the top PG's.  Same with CP3 for many years.

People are digging under rocks to find reasons why Rondo is #1.  And then when his PER doesn't quite stack up it means it's a flawed system.

Look at these numbers if you want ot see true eliteness:

http://www.basketball-reference.com/players/s/stockjo01.html    

Now I love Rondo's play and I'm thrilled he's on our team.  But that doesn't make him the best.

We won the title in 2008 without him being a key piece.  Obviously if we win this year he will be a key, but one of several.

[BballTim]

The fact that we need 20 different threads to disucss why Rondo is an elite PG only proves one thing; that he's not.

When Michael played nobody debated whether he was the best or not, it was universally known and accepted that he was.  Guys like Magic and Stockton in their prime were always regarded as the top PG's.  Same with CP3 for many years.

People are digging under rocks to find reasons why Rondo is #1.  And then when his PER doesn't quite stack up it means it's a flawed system.

Look at these numbers if you want ot see true eliteness:

http://www.basketball-reference.com/players/s/stockjo01.html    

Now I love Rondo's play and I'm thrilled he's on our team.  But that doesn't make him the best.

We won the title in 2008 without him being a key piece.  Obviously if we win this year he will be a key, but one of several.

  Some people will look at that page and notice that while Rondo's generally considered a top pg in the league (it's true, whether you believe it or not), Stockton wasn't even a starter at Rondo's age. When we won the title Rondo *was* a key piece. He was arguably our best player in two of our wins against the Lakers.

  If your argument is that Rondo's not on a MJ/Magic level, he's joined in that shortcoming by every other pg in the league.

[winsomme]

the only way to truly show the impact of Rondo would be by losing him.

personally, I'd rather win with him and let people still trumpet the idea that we'd be better with another PG than be without Rondo, lose, and be left saying I told you so...

as long as we have Rondo, people are going to continue to complain about him.