## Re: [APBR_analysis] Pace

Expand Messages
• ... From: John Maxwell To: Sent: Wednesday, January 30, 2002 9:12 AM Subject: Re:
Message 1 of 7 , Jan 30, 2002
• 0 Attachment
----- Original Message -----
From: "John Maxwell" <John.Maxwell@...>
To: <APBR_analysis@yahoogroups.com>
Sent: Wednesday, January 30, 2002 9:12 AM
Subject: Re: [APBR_analysis] Pace

> Just curious as to how you calculated "pace."

Team FGA + Opponent FGA + Team TO + Opponent TO, basically, then divided by
the league average and multiplied by 1000.

>
> For those who haven't done so, may I suggest that you check out DeanO's
site
> at http://www.rawbw.com/~deano/ to see how he figures this stuff out. For
> my money it's the best methodology out there, and he doesn't toot his own
> horn nearly enough.

Yeah, he's done some interesting stuff; I just wish he'd update more often
(on his STATS page, the "latest" year is still 95-96, and he doesn't have
any articles newer than 1998). Maybe the best thing he's got is the
groundwork for an NBA-style Project Scoresheet (see: his Possession Scoring
System) which would allow us to track defense and passing quite a bit
better.

Another thing that would be kind of cool to get started is a Baseball
Prospectus-style web page... haven't seen anything like that for basketball
just yet (RealGM comes closest in terms of analysis, but they don't really
do anything with stats as of yet).

John Craven
• John -- The pace that I typically calculate is team possessions or FGA - OR + TO + 0.4*FTA Calculate it using both the offensive and defensive team stats, then
Message 2 of 7 , Jan 30, 2002
• 0 Attachment
John --

The pace that I typically calculate is team possessions or

FGA - OR + TO + 0.4*FTA

Calculate it using both the offensive and defensive team stats, then
average. The two estimates are usually within 1-2% of each other and
I've checked the estimate against games I've scored. In theory, a
team and its opponents have the same number of possessions in each
game (+ or -2 and this + or - usually balances out over the season).

I use that for creating "Offensive Ratings" and "Defensive Ratings",
which are just Points scored and allowed per 100 possessions,
respectively. Fortunately, my relative rankings are almost identical
to yours so we're doing something similar. By my methodology,
Milwaukee, LA, Dallas, then Utah were top offenses last year. My top
5 D's were almost the same, with Sacramento in there, but in
different order (though all of them were _very_ similar). Bob
Chaikin has another way of estimating possessions.

You can also do things like compensate for how teams slack off
because they're involved in blowouts. That starts to get at why the
Lakers' D improved so much in the playoffs, but, as you say, they
turned it on then. (See Toying With 'Em at
http://www.rawbw.com/~deano/articles/aa052097.htm or They Say Defense
Wins Championships at
http://www.rawbw.com/~deano/articles/aa082197.htm)

The method I use for pace (and other things) is explained here:

http://www.rawbw.com/~deano/estabmthf.html

Dean Oliver

--- In APBR_analysis@y..., "John Craven" <john1974@u...> wrote:
> One thing that's always bothered me is that when people talk about
the best
> and worst offensive and defensive teams in the league, they nearly
always
> rank them by raw points scored for and against. Sometimes this is a
valid
> method, but not always - teams like the Mike Fratello Cavaliers are
bound to
> be overrated by this method, while teams like the George Karl
Seattle
> Supersonics (see my bias? ;) ) will be perennially underrated. Last
year, I
> observed that the Sacramento Kings really weren't half-bad on D,
even though
> they give up a lot of points per game, whereas Chicago wasn't even
close to
> average defensively.
>
> If you've ever watched the Kings play, you know that they like to
push the
> ball upcourt. Chicago last year did not. The net result is,
> ton more possessions (and opponent possessions) than did Chicago.
Think of
> this like park factors in baseball: just as some teams (the
Rockies, for
> instance) will have skewed hitting and pitching numbers because of
the park
> they play in, so will some teams in basketball have skewed offense
and
> defense numbers because of how fast they like to play.
>
> Without further ado, here are the top 5 and bottom 5 teams in pace:
>
> Top 5 (1000 = average)
> 1. Golden State 1049
> 2. Detroit 1046
> 3. Sacramento 1045
> 4. Orlando 1034
> 5. Dallas 1017
>
> The two most average-paced teams last year were Boston (999) and
Cleveland
> (1001).
>
> Bottom 5
> 1. New York 943
> 2. Miami 950
> 3. Portland 973
> 4. Charlotte 974
> 5t. Utah 977
> 5t. San Antonio 977
>
> You'll notice that the Knicks and Heat slowed things down more than
any
> single team sped things up. Overall, the bottom 5 pretty uniformly
have good
> defensive reputations; #7 (Chicago with a 978) is the first that
does not.
> Conversely, all five teams on the top of the list had reputations
for
> playing matador d, which isn't entirely true if you adjust for pace:
>
> Top 10 teams, Pace-Adjusted Opp. PPG
> 2. San Antonio 90.5
> 3. Phoenix 91.1
> 4. Miami 91.2
> 5. New York 91.3
> 6. Sacramento 91.8
> 7. Charlotte 92.2
> 8. Detroit 93.0
> 9. Orlando 93.3
> 10. Indiana 93.5
>
> Anyone who saw a lot of Philly last year knows that that #1 ranking
is
> accurate. Interestingly enough, a number of slow-paced teams (San
Antonio,
> New York, Miami) are still prominently featured. The biggest
surprise is
> Phoenix at #3. Some may be surprised that the Lakers aren't on the
list; all
> I have to say is, the team that played in the playoffs was not the
same one
> who played most of the season, defensively speaking. Let's take a
look at
> the worst 10:
>
> Bottom 10
> 29. Washington 99.2
> 28. Chicago 98.8
> 27. New Jersey 97.8
> 26. Denver 97.8
> 25. Vancouver 97.2
> 24. Golden State 96.8
> 23. Boston 96.7
> 22. Houston 96.6
> 21. Cleveland 96.5
> 20. Seattle 96.5
>
> Not surprisingly, these are some of the worst franchises in the
league. The
> only playoff team to make the list is Boston, though they only made
it in
> because of the weak East; most years, 36 wins isn't going to be
enough to do
> it. The only two teams with winning records on the list are Houston
and
> Seattle, and both are pretty far down it. The top offenses are a
bit more
> surprising:
>
> 1. Milwaukee 99.9
> 2. Utah 99.4
> 3. LA Lakers 99.3
> 4. Houston 99.0
> 5. Dallas 98.8
> 6. San Antonio 98.4
> 7. Portland 98.1
> 8. Toronto 97.5
> 9. Sacramento 97.4
> 10. Minnesota 96.8
> (Seattle was #11 with 96.6)
>
> 2 of the top 3 teams should surprise no one. In fact, the Lakers
being
> ranked so high should answer the question of how they won so many
games
> despite an average defense. But Utah? Yeah. Think about it.
Stockton and
> Malone are getting older, but they're still like clockwork, and
last year
> they had a third guy, Donyell Marshall, in the mix. They like to
slow things
> way down, but when they don't turn the ball over much and make a
lot of
> their shots. All the top 10 teams had plus-.500 records, and only
one
> (Houston) missed the playoffs. And they were the only top-10
offense team
> that was also on the bottom-10 defense list (Seattle just missed
out on that
> "honor").
>
> The worst 10:
> 29. Golden State 88.2
> 28. Chicago 89.5
> 27. Atlanta 90.4
> 26. Detroit 91.3
> 25. Vancouver 91.4
> 24. Cleveland 92.3
> 23. Washington 92.6
> 22. New Jersey 92.7
> 21. LA Clippers 93.0
> 20. Indiana 93.3
>
> Again, the worst offensive teams in the league were also, by and
large, the
> worst teams in the league. None of these guys reached .500. Indiana
came the
> closest to making the playoffs; not surprisingly, they were only
the 10th
> worst offense, and they had the 10th best defense to boot. With the
possible
> exception of the Clippers, none of these names should be
surprising. Chicago
> had the 2nd worst offense and 5th worst defense; that's the recipe
for
> losing 65 games, folks.
>
>
> John Craven
• ... by ... Apologies for the double-post, but here s my formula in full: All statistics are the team totals plus opponents totals.
Message 3 of 7 , Jan 30, 2002
• 0 Attachment
> > Just curious as to how you calculated "pace."
>
> Team FGA + Opponent FGA + Team TO + Opponent TO, basically, then divided
by
> the league average and multiplied by 1000.

Apologies for the double-post, but here's my formula in full:

All statistics are the team totals plus opponents' totals.

((FGM+OR+DR+TO+0.64055*PF)/211.61565)*1000

211.61565 was the league average number of "possessions" per game.

John Craven
• ... From: HoopStudies [mailto:deano@rawbw.com] Sent: Wednesday, January 30, 2002 12:53 PM To: APBR_analysis@yahoogroups.com Subject: [APBR_analysis] Re: Pace
Message 4 of 7 , Jan 30, 2002
• 0 Attachment
-----Original Message-----
From: HoopStudies [mailto:deano@...]
Sent: Wednesday, January 30, 2002 12:53 PM
To: APBR_analysis@yahoogroups.com
Subject: [APBR_analysis] Re: Pace

John --

The pace that I typically calculate is team
possessions or

FGA - OR + TO + 0.4*FTA

[Dean LaVergne] Dean, could you expound upon the
derivation of the 0.4 multiplier?

Thanks,
DeanL

__________________________________________________
Do You Yahoo!?
Great stuff seeking new owners in Yahoo! Auctions!
http://auctions.yahoo.com
• ... Almost purely empirical. I scored a bunch of games, counting possessions. I recognized that possessions occur when a team takes a field goal and does not
Message 5 of 7 , Jan 30, 2002
• 0 Attachment
--- In APBR_analysis@y..., Arthur LaVergne <deanlav@y...> wrote:
> The pace that I typically calculate is team
> possessions or
>
> FGA - OR + TO + 0.4*FTA
>
> [Dean LaVergne] Dean, could you expound upon the
> derivation of the 0.4 multiplier?

Almost purely empirical. I scored a bunch of games, counting
possessions. I recognized that possessions occur when a team takes a
field goal and does not get the offensive rebound (FGA - OR), turns
the ball over (TO), or on some free throws. (There is some junk on
the offensive rebounds since some team rebounds are awarded and some
OR's are off of FTA's.) Looking for the simplest fit and left with
the FTA's as the last factor, I just looked for a multiplier that fit
my data. 0.4 worked pretty well and didn't give the false impression
of being an exact number (like 0.387). Martin Manley also used the
0.4 for estimating the number of "plays" (he called them possessions)
as FGA + TO + 0.4*FTA.

So it's purely empirical. It would change if you had a 3-to-make-2
rule. (Though I tend to support Bob Chaikin's proposal for the 3-to-
make-2 rule, I've dreaded having to refigure out the empirical
number.) I have actually checked this multiplier on a smaller scale
in college and found 0.4 to work ok there even though they have the 1-
and-1. I would need to do a thorough study to check it there,
though.

DeanO
Your message has been successfully submitted and would be delivered to recipients shortly.