## Re: Predict regular season winning percentage here....

Expand Messages
• ... I think that s to be expected -- fundamentally, projections have to be pretty conservative to be accurate and place teams relatively closer to .500 than
Message 1 of 19 , Nov 3, 2004
• 0 Attachment
> ^^ Another prediction system that doesn't expect any team to win
> fewer than 22 games.

I think that's to be expected -- fundamentally, projections have to
be pretty conservative to be accurate and place teams relatively
closer to .500 than they actually are. Injuries are usually the
reason that doesn't turn out.

> Who needs that Zero? I know statisticians are fond of them. It
> looks like clutter to me; like any other insignificant digit.

I just put it there because it's in the NBA's actual standings and I
copied over them for the HTML.

> What's the Real reason you didn't include the Sonics ?

I think you explained that in the second post.
• ... From: thedawgsareout [mailto:kpelton08@hotmail.com] Sent: Wednesday, November 03, 2004 8:09 AM ... Yup, it is often a good idea to shrink or regress
Message 2 of 19 , Nov 3, 2004
• 0 Attachment
-----Original Message-----
From: thedawgsareout [mailto:kpelton08@...]
Sent: Wednesday, November 03, 2004 8:09 AM

>> ^^ Another prediction system that doesn't expect any team to win
>> fewer than 22 games.
>
>I think that's to be expected -- fundamentally, projections have to
>be pretty conservative to be accurate and place teams relatively
>closer to .500 than they actually are. Injuries are usually the
>reason that doesn't turn out.

Yup, it is often a good idea to "shrink" or "regress" forecasts
and estimates towards the mean. Even if this results in biased
estimates, they may nonetheless have lower expected error. A
canonical example is batting averages early in the season; after
a week or two, the leaderboard will have some guy hitting .456.
Standard simple statistics would say that that represents a sample
of his hitting and we should estimate his eventual batting average
by using his sample mean, i.e. .456. Common sense, as well as
more sophisticated "shrinkage to the mean" statistics, tell us that
by the end of the season, he'll be batting something well under
.456.

The best easy-to-read article that I've seen explaining this phenomenon
is by the statistician Bradley Efron, in a 1977 _Scientific
American_ article, "Stein's Paradox in Statistics".

>> Who needs that Zero? I know statisticians are fond of them. It
>> looks like clutter to me; like any other insignificant digit.
>
>I just put it there because it's in the NBA's actual standings and I
>copied over them for the HTML.

Depends on the circumstance. Leading 0s can be highly valuable
looking at a column of numbers such as

56
.81
4.2

it may not be instantly obvious that one of those numbers is much
smaller than the others. There are some circumstance where the
extra 0 is unnecessary or unhelpful, but there are other circumstances

--MKT
• ... What have these teams done to rate as highly (or better) than the seemingly-unbeatable Pistons of last year ? John Hollinger and Kevin Pelton seem to have
Message 3 of 19 , Nov 10, 2004
• 0 Attachment
--- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
> KP Tm. JH
> 49 Dal 56
> 55 SAS 59
> 49 Det 52
> 49 Mem 51
> 56 Ind 55
> 51 Uta 47
> 51 Sac 45

What have these teams done to rate as highly (or better) than the
seemingly-unbeatable Pistons of last year ?

John Hollinger and Kevin Pelton seem to have spent some time looking
over the rosters, grinding the data, and tweaking the results with
rookie forcasts and age-discrimination. And one or both guys likes
these other teams' chances as much or more than Detroit's.

The Pistons aren't an aging group of players (except Rasheed);
they've added McDyess (who has a nonzero chance of playing);
basically a bunch of guys with perfect chemistry and in their primes.

Using last year's stats, we don't see the team that rolled over
everyone after Rasheed came. Using the last 1/3 of the season and
playoffs we see That team.
• ... Well, I don t know what they see or don t see in Detroit, but one thing that I see is that their bench seems a lot weaker than last year. Their starters
Message 4 of 19 , Nov 10, 2004
• 0 Attachment
--- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
>
> --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
> > KP Tm. JH
> > 49 Dal 56
> > 55 SAS 59
> > 49 Det 52
> > 49 Mem 51
> > 56 Ind 55
> > 51 Uta 47
> > 51 Sac 45
>
> What have these teams done to rate as highly (or better) than the
> seemingly-unbeatable Pistons of last year ?
>
> John Hollinger and Kevin Pelton seem to have spent some time looking
> over the rosters, grinding the data, and tweaking the results with
> rookie forcasts and age-discrimination. And one or both guys likes
> these other teams' chances as much or more than Detroit's.
>
> The Pistons aren't an aging group of players (except Rasheed);
> they've added McDyess (who has a nonzero chance of playing);
> basically a bunch of guys with perfect chemistry and in their primes.
>
> Using last year's stats, we don't see the team that rolled over
> everyone after Rasheed came. Using the last 1/3 of the season and
> playoffs we see That team.

Well, I don't know what they see or don't see in Detroit, but one
thing that I see is that their bench seems a lot weaker than last
year. Their starters are playing 35+ minutes a game and their best
reserve seems to be McDyess (which by the way seems a downgrade
compared to Okur).
• Good points. If Detroit was projected based on how they played once they got Rasheed, I assume they would fare much better (they went 20-6 in the regular
Message 5 of 19 , Nov 10, 2004
• 0 Attachment
Good points. If Detroit was projected based on how they played once
they got Rasheed, I assume they would fare much better (they went 20-6
in the regular season once they got him).

And so far, Detroit's bench does look weaker. However, it's definitely
too soon to say that for sure as the new players will need time to
and if the bench remains inferior to last year's Detroit's overall win
total might not be as high. But in the playoffs, the bench won't
matter much because they will get fewer minutes. Last year, the
Pistons won it all despite very poor play from their bench in the
playoffs (Williamson and Okur basically disappeared).

--- In APBR_analysis@yahoogroups.com, "Carlos" <carlosmanuel@b...>
wrote:
>
> --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...> wrote:
> >
> > --- In APBR_analysis@yahoogroups.com, "Mike G" <msg_53@h...>
wrote:
> > > KP Tm. JH
> > > 49 Dal 56
> > > 55 SAS 59
> > > 49 Det 52
> > > 49 Mem 51
> > > 56 Ind 55
> > > 51 Uta 47
> > > 51 Sac 45
> >
> > What have these teams done to rate as highly (or better) than the
> > seemingly-unbeatable Pistons of last year ?
> >
> > John Hollinger and Kevin Pelton seem to have spent some time
looking
> > over the rosters, grinding the data, and tweaking the results with
> > rookie forcasts and age-discrimination. And one or both guys
likes
> > these other teams' chances as much or more than Detroit's.
> >
> > The Pistons aren't an aging group of players (except Rasheed);
> > they've added McDyess (who has a nonzero chance of playing);
> > basically a bunch of guys with perfect chemistry and in their
primes.
> >
> > Using last year's stats, we don't see the team that rolled over
> > everyone after Rasheed came. Using the last 1/3 of the season and
> > playoffs we see That team.
>
>
> Well, I don't know what they see or don't see in Detroit, but one
> thing that I see is that their bench seems a lot weaker than last
> year. Their starters are playing 35+ minutes a game and their best
> reserve seems to be McDyess (which by the way seems a downgrade
> compared to Okur).
• ... I don t think that s the issue. I rate Detroit s end-of-season roster at 59 wins last year because they added so much value by trading spare parts for
Message 6 of 19 , Nov 11, 2004
• 0 Attachment
> Using last year's stats, we don't see the team that rolled over
> everyone after Rasheed came. Using the last 1/3 of the season and
> playoffs we see That team.

I don't think that's the issue. I rate Detroit's end-of-season
roster at 59 wins last year because they added so much value by
trading spare parts for Rasheed and Mike James.

By comparison, I have Minnesota at 55, Indiana at 51, the Lakers at
48, Sacramento at 56 and San Antonio at 54. So I don't think I'm
underrating the 03-04 Pistons.

> The Pistons aren't an aging group of players (except Rasheed);
> they've added McDyess (who has a nonzero chance of playing);
> basically a bunch of guys with perfect chemistry and in their
> primes.

It's the bench, mostly. I have James and Okur rated as worth a
combined nine wins last year (James had a heck of a year, let's not
forget), Williamson another win and a half.

*Combined*, I have this year's reserves worth 2.3 wins. Now might
they be better than that? Certainly. I think Hunter may be
underrated by his individual traditional stats, while Delfino's
rating is a SWAG and that Antonio McDyess was coming off a knee
injury last year isn't factored in.

Also, only Rasheed amongst Detroit's starters missed more than four
games last year. I conservatively project at most 75 games played,
so that lessens their value a little bit.

> What have these teams done to rate as highly (or better) than the
> seemingly-unbeatable Pistons of last year ?

If I might answer that question in just two words, they would be
Brent. And Barry.
• For less analytical takes on basketball, please chck out my blog: http://highfivehoopschool.blogspot.com ... Do you Yahoo!? Check out the new Yahoo! Front
Message 7 of 19 , Nov 13, 2004
• 0 Attachment

http://highfivehoopschool.blogspot.com

Do you Yahoo!?
Check out the new Yahoo! Front Page. www.yahoo.com
• Just to agree with everyone else here -- it s the bench. You can see it already now that they ve had an injury or two -- they ve got NOBODY to fill in, whereas
Message 8 of 19 , Nov 15, 2004
• 0 Attachment
Just to agree with everyone else here -- it's the bench. You can see
it already now that they've had an injury or two -- they've got
NOBODY to fill in, whereas last year they had a ton of guys.

McDyess, IMHO, will be exposed as a bust, although I must admit he's
played better than I expected so far. If you look at his stats at the
end of last year in Phoenix, they were nothing, but because people
get all goofy about per game stats the Pistons looked at his last 10
games and talked about him as a double-double guy. Ain't happening.
And he definitely won't replace the production of Okur.

They gave away Williamson and James and replaced them with
replacement-level Ronald Dupree and Delfino, who based on my Euro
conversions looks way over his head offensively.

That leaves:
-- 38-year old Elden Campbell. Useful at times, but a poor man's
McDyess.
-- Kneeless Derrick Coleman. No value.
-- Lindsey Hunter, who is 35, can't play point but forced to be the
backup point guard, and due for a fall after coming off his best year
since leaving Milwaukee
-- Darvin Ham, a locker room guy with the worst jumper in the league.

So instead of James-Williamson-Okur as the three bench guys who see
the most action, the Pistons have Hunter-Delfino-McDyess. That is a
MAJOR downgrade, enough to make Indiana and San Antonio better
choices as the favorite this year.

--- In APBR_analysis@yahoogroups.com, "thedawgsareout"
<kpelton08@h...> wrote:
>
> > Using last year's stats, we don't see the team that rolled over
> > everyone after Rasheed came. Using the last 1/3 of the season
and
> > playoffs we see That team.
>
> I don't think that's the issue. I rate Detroit's end-of-season
> roster at 59 wins last year because they added so much value by
> trading spare parts for Rasheed and Mike James.
>
> By comparison, I have Minnesota at 55, Indiana at 51, the Lakers at
> 48, Sacramento at 56 and San Antonio at 54. So I don't think I'm
> underrating the 03-04 Pistons.
>
> > The Pistons aren't an aging group of players (except Rasheed);
> > they've added McDyess (who has a nonzero chance of playing);
> > basically a bunch of guys with perfect chemistry and in their
> > primes.
>
> It's the bench, mostly. I have James and Okur rated as worth a
> combined nine wins last year (James had a heck of a year, let's not
> forget), Williamson another win and a half.
>
> *Combined*, I have this year's reserves worth 2.3 wins. Now might
> they be better than that? Certainly. I think Hunter may be
> underrated by his individual traditional stats, while Delfino's
> rating is a SWAG and that Antonio McDyess was coming off a knee
> injury last year isn't factored in.
>
> Also, only Rasheed amongst Detroit's starters missed more than four
> games last year. I conservatively project at most 75 games played,
> so that lessens their value a little bit.
>
> > What have these teams done to rate as highly (or better) than the
> > seemingly-unbeatable Pistons of last year ?
>
> If I might answer that question in just two words, they would be
> Brent. And Barry.
• I was wondering about your comments on Delfino. How do you make these Euro conversions? How accurate have they been in the past? From watching a few Pistons
Message 9 of 19 , Nov 15, 2004
• 0 Attachment
Euro conversions? How accurate have they been in the past? From
watching a few Pistons games so far, he has impressed me on offense.
Obviously his game is still a work in progress, and his jumper has
been poor, but he's been very active and is fearless going to the
hoop. He's had some spectacular drive and dunks. Defense has been his
main problem from what I've seen.

--- In APBR_analysis@yahoogroups.com, "John Hollinger" <alleyoop2@y...
> wrote:
>
> Just to agree with everyone else here -- it's the bench. You can see
> it already now that they've had an injury or two -- they've got
> NOBODY to fill in, whereas last year they had a ton of guys.
>
> McDyess, IMHO, will be exposed as a bust, although I must admit he's
> played better than I expected so far. If you look at his stats at
the
> end of last year in Phoenix, they were nothing, but because people
> get all goofy about per game stats the Pistons looked at his last 10
> games and talked about him as a double-double guy. Ain't happening.
> And he definitely won't replace the production of Okur.
>
> They gave away Williamson and James and replaced them with
> replacement-level Ronald Dupree and Delfino, who based on my Euro
> conversions looks way over his head offensively.
>
> That leaves:
> -- 38-year old Elden Campbell. Useful at times, but a poor man's
> McDyess.
> -- Kneeless Derrick Coleman. No value.
> -- Lindsey Hunter, who is 35, can't play point but forced to be the
> backup point guard, and due for a fall after coming off his best
year
> since leaving Milwaukee
> -- Darvin Ham, a locker room guy with the worst jumper in the
league.
>
> So instead of James-Williamson-Okur as the three bench guys who see
> the most action, the Pistons have Hunter-Delfino-McDyess. That is a
> MAJOR downgrade, enough to make Indiana and San Antonio better
> choices as the favorite this year.
>
>
>
> --- In APBR_analysis@yahoogroups.com, "thedawgsareout"
> <kpelton08@h...> wrote:
> >
> > > Using last year's stats, we don't see the team that rolled over
> > > everyone after Rasheed came. Using the last 1/3 of the season
> and
> > > playoffs we see That team.
> >
> > I don't think that's the issue. I rate Detroit's end-of-season
> > roster at 59 wins last year because they added so much value by
> > trading spare parts for Rasheed and Mike James.
> >
> > By comparison, I have Minnesota at 55, Indiana at 51, the Lakers
at
> > 48, Sacramento at 56 and San Antonio at 54. So I don't think I'm
> > underrating the 03-04 Pistons.
> >
> > > The Pistons aren't an aging group of players (except Rasheed);
> > > they've added McDyess (who has a nonzero chance of playing);
> > > basically a bunch of guys with perfect chemistry and in their
> > > primes.
> >
> > It's the bench, mostly. I have James and Okur rated as worth a
> > combined nine wins last year (James had a heck of a year, let's
not
> > forget), Williamson another win and a half.
> >
> > *Combined*, I have this year's reserves worth 2.3 wins. Now might
> > they be better than that? Certainly. I think Hunter may be
> > underrated by his individual traditional stats, while Delfino's
> > rating is a SWAG and that Antonio McDyess was coming off a knee
> > injury last year isn't factored in.
> >
> > Also, only Rasheed amongst Detroit's starters missed more than
four
> > games last year. I conservatively project at most 75 games played,
> > so that lessens their value a little bit.
> >
> > > What have these teams done to rate as highly (or better) than
the
> > > seemingly-unbeatable Pistons of last year ?
> >
> > If I might answer that question in just two words, they would be
> > Brent. And Barry.
Your message has been successfully submitted and would be delivered to recipients shortly.