Rondo vs the other elite PGs: by the numbers

[drza44]

I didn't account for home and away games either, that's something that we could do. Typically HCA is worth an expected +3 efficiency differential if I remember that right. (not sure if its +3 raw points or +3 per 100 possessions)

For what it's worth in the 9 games Rondo was out the Cs played 5 road games (@tor, @Atl, @NJN, @Orl, @Ind) and 4 home games while in the 8 games that KG was out (that Rondo played in) the Cs played 2 road games (@Tor, @Chi) and 6 home games.

In a quick search I couldn't find whether the 3 points were supposed to be added to the total or on a per-100 possessions basis, but I did come across another idea that was interesting for small sample sizes like this.  Instead of using a team's overall record in the SoS, instead this guy suggested using their home record (for home games) or their away record (for away games) to give a more accurate idea about how strong the team actually was.

For example, the 76ers are 15 - 8 at home but 6 - 18 on the road.  Thus, playing them at one location means you're playing one of the better teams in the league and at the other you're playing one of the worst on that given day.  If I get a few minutes I may re-calculate the strengths of schedules using that modification.

[drza44]

So, with the adjustment I just mentioned for using home record and away record:

KG out/Rondo in (8 games): Opponents were 95 - 118 (SOS .446)
Celtics went 6 - 2, +6.7 differential (per Fafnir)

Rondo out/KG in (9 games): Opponents were 75 - 118 (SOS .389)
Celtics went 7 - 2, +10.4 differential (per Fafnir)

[Celtics18]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

[BballTim]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

  .475 should be either .436 or .44 depending on which website you go to.

[Celtics18]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

  .475 should be either .436 or .44 depending on which website you go to.

Thanks.  One site that I looked at recommended the .475 number.  Maybe, I'll go back and try the other ones to see how significantly the overall numbers are changed.

[Edgar]

where are Jammer and Miller?

[Fafnir]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

The actual estimates for possessions ia more complicated that if you use Dean Olivers formula. It involves adding together offensive and defensive rebounds to get an estimate, you do the calculation for both teams, and then you average the two to get an estimate for possessions.

I have the book at home on my kindle, I can post the entire formula that Basketballreferrence is using then.

Your formula is a solid one, but Oliver made a few tweaks to get it to be more accurate in his view.

[Celtics18]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

The actual estimates for possessions ia more complicated that if you use Dean Olivers formula. It involves adding together offensive and defensive rebounds to get an estimate, you do the calculation for both teams, and then you average the two to get an estimate for possessions.

I have the book at home on my kindle, I can post the entire formula that Basketballreferrence is using then.

Your formula is a solid one, but Oliver made a few tweaks to get it to be more accurate in his view.

Thanks.  I don't like the sound of "more complicated".  I just feel like it would be valuable for discussions like this if we where all using the same numbers. 

I also like having something that I can relatively easily plug into a boxscore and get an immediate result from.

[BballTim]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

The actual estimates for possessions ia more complicated that if you use Dean Olivers formula. It involves adding together offensive and defensive rebounds to get an estimate, you do the calculation for both teams, and then you average the two to get an estimate for possessions.

I have the book at home on my kindle, I can post the entire formula that Basketballreferrence is using then.

Your formula is a solid one, but Oliver made a few tweaks to get it to be more accurate in his view.

  I think if he changes his free throw multiplier he'll get close enough to the numbers they use, at least on basketballreference.

[drza44]

where are Jammer and Miller?

Yeah, Billups could be in there too.  Like I mentioned earlier, I started this as a best-of-the-best comp but I had 6 guys (Rondo/Paul/Rose/Williams/Westbrook/Nash).  Then, I thought about how Felton was getting All Star game buzz so I added him.  I threw in Wall because he got early buzz, and then added Curry and Parker to get it to 10 for a nice round number.  So my list isn't necessarily a true top-10, more of a comparison of 10 PGs that get a lot of buzz.  I could go back and add in a few more, but I don't really think they'd change the story that much.

[Fafnir]

Hey, guys, please help me out with this; how come the offensive and defensive efficiency numbers on NBA Reference are so different from Hollinger's efficiency numbers? Both claim to be representative of points per possession.

To come up with points per possession for the games being discussed without either Rondo or Garnett, I plugged this formula, FGAs-Orebs+TOs+(.475 x FTAs), into the individual boxscores for those games. This gave me the number of possessions for each game. Then I divided the total points scored by the number of possessions and moved the decimal point to make it per 100 possessions, and that was it.

Did I do this wrong?  Why are my numbers so different from Fafnir's?

The actual estimates for possessions ia more complicated that if you use Dean Olivers formula. It involves adding together offensive and defensive rebounds to get an estimate, you do the calculation for both teams, and then you average the two to get an estimate for possessions.

I have the book at home on my kindle, I can post the entire formula that Basketballreferrence is using then.

Your formula is a solid one, but Oliver made a few tweaks to get it to be more accurate in his view.

Thanks.  I don't like the sound of "more complicated".  I just feel like it would be valuable for discussions like this if we where all using the same numbers.  

I also like having something that I can relatively easily plug into a boxscore and get an immediate result from.

Here's the Dean Oliver formula:

Poesssions = FGA - OREB/(OREB + DREB) * (FG -FGM) * 1.07 + TOV + 0.4 * FTA

I believe this this the formula that Basketball Reference uses. Typically you calculate this for both teams and then average them to get a pretty good guess at the pace of the game.

[ElGee]

Yes, I track those events too. Rondo had 7 in last year's playoffs, 52nd in the playoffs (of 133 qualifiers) per possession at doing this.

ElGee, how'd you you even know someone from here had referenced your blog?  You got some kind of esp or something, lol?

Blog tells me.

[Eja117]

Those are some great stats.

I think it's hard to take him seriously when he can't hit a three or a free throw.

But on the other hand he has the "Yeah, but none of those guys can do that factor" like some of the passes he gets or the nights with 25 assists or the times he gets a ton of rebounds. Very few of those guards get as many rebs as him.

The fact that he plays with 3 all stars and they still gear defenses against him.

But I'd love to see what Nash, Paul, and Derron could do with teammates like this