APBR_analysis is a Restricted Group with 454 members.
 APBR_analysis

 Restricted Group,
 454 members
Re: nice methods
Expand Messages
 That last post certainly was hard to read. This is a bad place to
try to post in columns, as all spaces are compressed to a single
space.
Anyway, I did notice what was the big difference between my rankings
and John Craven's. My list had Paul Pierce pretty high up, where his
didn't include Pierce at all, nor Antoine Walker or Stackhouse.
These 3 guys played for very bad teams last season. This seems to
disqualify members of such teams from consideration as great players,
in these equivalentwins (or individualwins) methods.
What I wonder is, does anyone have a winsbased evaluation method
that dates back several years? If so, are players like Pierce
considered "bad" players as long as their team is bad, and then
suddenly become "good" players at the moment their team improves (or
they move to a better team)?
Mike Goodman
p.s. If you hit "Reply", a post will appear with columns restored.   In APBR_analysis@y..., "mikel_ind" <msg_53@h...> wrote:
> That last post certainly was hard to read. This is a bad place to
rankings
> try to post in columns, as all spaces are compressed to a single
> space.
>
> Anyway, I did notice what was the big difference between my
> and John Craven's. My list had Paul Pierce pretty high up, where
his
> didn't include Pierce at all, nor Antoine Walker or Stackhouse.
players,
>
> These 3 guys played for very bad teams last season. This seems to
> disqualify members of such teams from consideration as great
> in these equivalentwins (or individualwins) methods.
(or
>
> What I wonder is, does anyone have a winsbased evaluation method
> that dates back several years? If so, are players like Pierce
> considered "bad" players as long as their team is bad, and then
> suddenly become "good" players at the moment their team improves
> they move to a better team)?
I don't have it with me here at work, but I have been doing wins
based stuff for a long time and it shows this kind of change for
supporting cast kind of players  the Steve Kerr's of the world who
are valuable on good teams, but not so valuable on poor teams. By my
numbers last year, Pierce was one of the best players in the league,
with a winloss record of 12.24.5. Kobe, for comparison, was 11.6
3.4. Those are some very solid numbers.
Another example I think of as an interesting one was Mitch Richmond.
I consistently had him winning a lot of games for Sacramento, then he
tanked upon being traded to Washington. That was weird.
I had Andre Miller at 8.95.1 last year. McDyess was 8.64.3.
Stackhouse was 10.08.3. Jamison was 6.210.5. AWalker was 7.710.1.
I think Miller's record may be typical for winloss records of very
good players on bad teams. The reason is defense. Miller and Pierce
and Kobe are not the kind of players who can completely turn around a
defense, make it good. Only big men can really do that (and maybe
Jason Kidd).
Also, if you have a team that wins only 15 games, it doesn't make
sense to have one player on that team who wins 16 games like a Shaq
or a Jordan or Duncan do. Those guys make winning teams. Elton
Brand, though a good player, clearly doesn't add more than the 15
wins that Chicago had last year; it's impossible. My first cut was
an addition of 6 wins by Brand. I frankly am coming to believe that
it was more like 9 wins, but that's all a little theoretical right
now. And it always seems strange to me if someone contributes more
than half of his team's wins (especially since the Bulls are on pace
to beat 15 wins this year). It's hard to argue that even Jordan ever
won half his team's games when he was scoring 35 ppg and the Bulls
were winning only 40 games. Actually, Jordan should be a very good
example of what MikeG was asking for. I don't have all my info in
front of me (again), but
http://www.rawbw.com/~deano/articles/JordanvsOlaj.html
shows that Jordan was 19.40.7 in 1988. I don't think the Bulls were
that good that year, maybe 4042 (help?). So, sure, it is indeed
very possible for winbased ranking methodologies to show stars (or
superstars in this case) on mediocre teams.
Dean Oliver
Journal of Basketball STudies
>
> Mike Goodman
>
> p.s. If you hit "Reply", a post will appear with columns restored. > > What I wonder is, does anyone have a winsbased evaluation method
who
> > that dates back several years? If so, are players like Pierce
> > considered "bad" players as long as their team is bad, and then
> > suddenly become "good" players at the moment their team improves
> (or
> > they move to a better team)?
>
> I don't have it with me here at work, but I have been doing wins
> based stuff for a long time and it shows this kind of change for
> supporting cast kind of players  the Steve Kerr's of the world
> are valuable on good teams, but not so valuable on poor teams. By
my
> numbers last year, Pierce was one of the best players in the
league,
> with a winloss record of 12.24.5. Kobe, for comparison, was 11.6
Let me back up a little here. I do see some variation for even good
> 3.4. Those are some very solid numbers.
>
players who change teams IF the team defense of the two teams are
different. Most players do not have significant effects on team
defense, big men being the exception (and apparently Jason Kidd and
maybe MJ). Offensively, stars don't really change much from team to
team. Role players can, but don't necessarily. Defensively, it's a
mixed bag. And since I don't calculate wins and losses as an
explicit rating method (usual Bill Jameslike disclaimer: I don't
believe in one number for overall ratings of players), they are just
meant to reflect a player's contribution to his team. They are a
pseudomeasurement, not an overall rating. Because they are pseudo
measurements, they are not subjective and, hence, a bit more
predictable and meaningful than stupid awards.
Pierce, by the way, is a pretty unusual player. His winloss record
has been above 0.500 since entering the league, a starlike quality.
5.91.8 as a rookie. 8.54.5 as a 2nd year guy. 12.24.5 last
year. He's a good but not great defender. You put him on one of
these poor defensive teams (like Cleveland) and they may get a little
better defensively. They should get better offensively.
> Also, if you have a team that wins only 15 games, it doesn't make
that
> sense to have one player on that team who wins 16 games like a Shaq
> or a Jordan or Duncan do. Those guys make winning teams. Elton
> Brand, though a good player, clearly doesn't add more than the 15
> wins that Chicago had last year; it's impossible. My first cut was
> an addition of 6 wins by Brand. I frankly am coming to believe
> it was more like 9 wins, but that's all a little theoretical right
pace
> now. And it always seems strange to me if someone contributes more
> than half of his team's wins (especially since the Bulls are on
> to beat 15 wins this year).
This is an interesting case. The Clips are better with Brand. The
Bulls are better without Brand. So how good is Brand? Context
sensitive.
> It's hard to argue that even Jordan ever
were
> won half his team's games when he was scoring 35 ppg and the Bulls
> were winning only 40 games. Actually, Jordan should be a very good
> example of what MikeG was asking for. I don't have all my info in
> front of me (again), but
>
> http://www.rawbw.com/~deano/articles/JordanvsOlaj.html
>
> shows that Jordan was 19.40.7 in 1988. I don't think the Bulls
> that good that year, maybe 4042 (help?). So, sure, it is indeed
OK. I'm back home and the Bulls were 5032 in 1988. In 1987, the
> very possible for winbased ranking methodologies to show stars (or
> superstars in this case) on mediocre teams.
Bulls were 4042 and Jordan was 17.33.7. That's about as close to
half a team's wins I can quickly find. Jordan scored 37 ppg with
Oakley and John Paxson as principal surrounding cast. With Pippen
and Grant around the following year, the team D got a little better
and Jordan's O got a little more efficient (his FG% went from 48% to
54%). Was Jordan a better player in 1988 than in 1987 because his
winloss record was better, because his offensive and defensive
numbers improved? Hell, I don't know. I don't really care. It was
obvious that he would improve both an offense and a defense. If the
Lakers offered me Magic Johnson at the time for Jordan, would I have
taken it? I guess we'll never know....
Dean Oliver
Journal of Basketball Studies On Tue, 5 Feb 2002, HoopStudies wrote:
> Also, if you have a team that wins only 15 games, it doesn't make
^^^^^^^^^^^
> sense to have one player on that team who wins 16 games like a Shaq
> or a Jordan or Duncan do. Those guys make winning teams. Elton
> Brand, though a good player, clearly doesn't add more than the 15
> wins that Chicago had last year; it's impossible. My first cut was
> an addition of 6 wins by Brand. I frankly am coming to believe that
> it was more like 9 wins, but that's all a little theoretical right
> now. And it always seems strange to me if someone contributes more
> than half of his team's wins (especially since the Bulls are on pace
I think it depends on how were define "contributes"; see below.
> to beat 15 wins this year). It's hard to argue that even Jordan ever
> won half his team's games when he was scoring 35 ppg and the Bulls
The 19.40.7 wonloss record for Jordan I am okay with. From your
> were winning only 40 games. Actually, Jordan should be a very good
> example of what MikeG was asking for. I don't have all my info in
> front of me (again), but
>
> http://www.rawbw.com/~deano/articles/JordanvsOlaj.html
>
> shows that Jordan was 19.40.7 in 1988. I don't think the Bulls were
> that good that year, maybe 4042 (help?). So, sure, it is indeed
> very possible for winbased ranking methodologies to show stars (or
> superstars in this case) on mediocre teams.
article, it appears to be based on a solid notion of looking at a player's
offensive and defensive ratings, comparing those to what teams of similar
off. and def. ratings would achieve in wonloss terms, and calling that
the player's wonloss record. I have no problem with that.
But I don't think Jordan's 19.40.7 wonloss record can be regarded as
being on the same measurement scale as the Bulls' 4042 record (actually
they were 5032 in 1988, but that doesn't matter). Nor can Brand's 6 or 9
individual victories be compared to the Bulls' total of 15.
We can calculate the Bulls' individual wonloss records in 1988, but we
cannot say that the sum of those wonloss records should equal 4042 (or
5032), nor can we say that 19.4 is almost half of 40. Those are apples
and oranges.
If we do want to claim that those 19.4 individual wins can be directly
compared to the team's 40 or 50 wins, we are imposing an overly simplistic
model upon how a team's record is determined by its players production.
Implicitly, it requires that the model be: Bulls Wins == sum of Jordan's
wins + Pippen's wins + Grant's wins + etc. etc. And that equation is
almost certainly an incorrect one for determining how individual players,
when put on a team, determine the team's wonloss record. It is extremely
unlikely that the correct model is a simple linear sum.
And if it is not a simple linear sum, then we can't directly compare
Jordan's 19.4 wins to the Bulls' 40 or 50 wins.
An analogy: if someone gets a 1400 SAT score, and such students usually
get 3.6 GPAs in college, we cannot say that the student's college GPA is
3.6/1400 = .0026 of their SAT score. Well we can say it, but it's not a
useful calculation. Nor is it useful to say that Brand or Jordan
contributed to half of their teams totals, based on their individual
wonloss stats.
It's much the same problem that you've pointed out with linear weights
systems: much of the world is not linear. If 10% of Microsoft's costs
are spent on systems analysts, can we claim that systems analysts
contribute to 10% of Microsofts production? It's not a useful ratio (the
second one I mean; the 10% of costs figure is useful for analyzing costs);
if Microsoft cut its systems analysts roster in half, would its production
fall by half? If it doubled its roster, would its production double? No
and no. Nor can we say that Jordan's 19.4 wins are about half of the
Bulls 40 wins.
Yet another way of looking at it: divvying up the 40 wins and saying that
soandso is responsible for x of them is an exercise doomed to failure.
How many of the 40 wins were due to Jordan, Pippen, etc.? After we finish
divvying them up, we then better ask: wait, how many wins would the Bulls
have had if they didn't have a coach? And for that matter an equipment
manager, ticket takers, stadium maintenance, etc.
Just as we can't look at Microsoft's sales of x million pieces of software
and say "Bill Gates produced y million pieces of software, Steve Ballmer
produced z million of them, the new programmer they hired produced w of
them, etc." That linear divvying up of production is not how the
production function works. Nor can a team's wins be linearly divvied up
among its players.
What we CAN do with players is try to estimate their MARGINAL value: how
many wins did they contribute compared to how many player X would have
contributed (where X could be a player that Jordan was traded for, or a
chosen comparison player, or a replacement level player if we could agree
on what the replacement level is, or whatever). And the 19.4 wins and 0.7
losses might be good estimates of that marginal value.
BUT: there is no requirement that the sum of players' marginal wins
equate to the team's total wins. Only with (in economics terms) "constant
returns to scale"  e.g. a simple linear sum  would that happen.
The marginal values can often be wellapproximated by linear methods.
But at extreme values even the marginal wins can't be interpreted
literally, or in a linear fashion. Does adding Jordan add 19 wins to a
team's total? Could be, in the case of the 2002 Wizards compared to 2001.
But if we're looking at the 1972 6913 Lakers, it is mathematically
impossible that adding Jordan to their roster would add 19 wins to their
total.
Similarly, if Jordan were to play for a really bad team that won only 15
games, would we say that subtracting Jordan from that team would cause
them to decrease their win total by 19?
Yet 19.40.7 might still be quite a good measure of Jordan's prowess. But
we can't interpret that 19.4 figure as one that can be directly compared
to the Bulls' 40 wins, or 15 wins, or 69 wins, or whatever their total is.
We can say that Jordan "contributed" 19.4 wins at the margin, but that
does not literally mean that he would add 19 victories to a team's total.
Wizards, maybe yes. 1972 Lakers no.
Nonlinearity. Individual wonloss records can be a fine way of
measuring players' production, but the jump from those individual wonloss
records to the team's actual wonloss record is not a simple one. Team
stats are a complex function of the stats of the individual players. Not
a simple linear sum.
And therefore Brand's, or Jordan's, individual victories cannot be
directly compared to their team's victories.
MKT   In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
>
wonloss
> Nonlinearity. Individual wonloss records can be a fine way of
> measuring players' production, but the jump from those individual
> records to the team's actual wonloss record is not a simple one.
Team
> stats are a complex function of the stats of the individual
players. Not
> a simple linear sum.
Whether or not the theoretical analysis is ok, it is tempting to do
>
> And therefore Brand's, or Jordan's, individual victories cannot be
> directly compared to their team's victories.
exactly this because the sum of individual winloss records does
almost always come very close to the team winloss total. I agree
that the model is simplistic. The fact that the sum is the team
total is why I call it a pseudomeasurement. There is a reality
check on it to some degree. And the model for getting there is
simple. Eh. Whatevah.
In terms of predictions, individual winloss records don't work,
despite my hopes nearly 15 years ago. I have actually come close to
proving that it is theoretically impossible to have a simple number
that allows you to predict a team's winloss record with that player
in place of another. Even the net points stuff I have, which comes
closer. It is practically impossible to remove context. Your 6913
Laker team is a good example, I think, reflecting how context is
important in making predictions. Or consider a team that wins by 10
ppg. Replace a player who contributes net 1 ppg with one who
contributes net 6 ppg is very unlikely to make that team win by 15
ppg because the team doesn't need the extra 5 ppg to win. Very
context sensitive.
Anyway, we can find reasons to discard EVERY single number we
calculate here. Individual winloss records are simple scans of
contribution that do sum to the team total, giving them a reality
check that linear weights do not have. I like that conceptually. I
don't claim it's predictive (why I pointed out the Brand conundrum),
but no one is putting forth any way to make those predictions.
Unfortunately.
Dean Oliver
Journal of Basketball Studies  On Wed, 6 Feb 2002, HoopStudies wrote:
>  In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
[...]
> >
> > Nonlinearity. Individual wonloss records can be a fine way of
> > measuring players' production, but the jump from those individual
> wonloss
> > records to the team's actual wonloss record is not a simple one.
> despite my hopes nearly 15 years ago. I have actually come close to
If "simple" includes "linear" yes, models that are that simple will not
> proving that it is theoretically impossible to have a simple number
work.
> that allows you to predict a team's winloss record with that player
Yes, diminishing marginal returns, and other nonlinearities abound.
> in place of another. Even the net points stuff I have, which comes
> closer. It is practically impossible to remove context. Your 6913
> Laker team is a good example, I think, reflecting how context is
> important in making predictions. Or consider a team that wins by 10
> ppg. Replace a player who contributes net 1 ppg with one who
> contributes net 6 ppg is very unlikely to make that team win by 15
> ppg because the team doesn't need the extra 5 ppg to win. Very
> context sensitive.
> Anyway, we can find reasons to discard EVERY single number we
I hope I made it clear that I was not criticizing the individual wonloss
> calculate here. Individual winloss records are simple scans of
records as measures of player quality, nor as measures of marginal
contributions to wins.
> contribution that do sum to the team total, giving them a reality
This is the part that is troublesome. It's nice that they sum to the team
> check that linear weights do not have. I like that conceptually. I
total, but on the whole I think that doesn't really tell us much about the
validity of the model. With suitable normalization, most or at any
rate many rating schemes could be made to have sums which come close to
adding up to the team's win total.
> don't claim it's predictive (why I pointed out the Brand conundrum),
I suspect it's part of the datafitting problem in statistics. When we
have a set of data, it's pretty easy to come up with a model that fits
that data set really well. But such models usually perform poorly when
used to make actual predictions on outofsample data (i.e. real world
predictions).
Good predictive models are very hard to create. Just ask any economist to
try to predict when the next recession will come. Or any geologist when
the next big earthquake will hit Los Angeles.
> but no one is putting forth any way to make those predictions.
Yes, part of the Holy Grail again: how do individual players' qualities
> Unfortunately.
(and statistics measuring those qualities) combine into determining the
team's outcome? A problem that is difficult enough in baseball and harder
still in basketball.
MKT   In APBR_analysis@y..., "HoopStudies" <deano@r...> wrote:
>
"The Bulls are better without Brand" is only half a comment.
>...... The Clips are better with Brand. The
> Bulls are better without Brand. So how good is Brand? Context
> sensitive.
"...than they would be if they still had him"?
or "...than they were when they had him"?
One might imagine that 1036 is a better mark than 1567, but the
Bulls' average score is 85.794.4, compared to 87.596.6 last year.
No significant change in the scoring.
Ron Artest is suddenly a star this year. Brad Miller and Marcus
Fizer are suddenly serious players. Mercer and Hoiberg have dropped
off, but Anthony has come along, with Oakley, while nobody
significant has been dumped.
With the coaching change, I would agree "the Bulls are better"; but
with Brand they might actually be contending.   In APBR_analysis@y..., "mikel_ind" <msg_53@h...> wrote:
> >...... The Clips are better with Brand. The
year.
> > Bulls are better without Brand. So how good is Brand? Context
> > sensitive.
>
> "The Bulls are better without Brand" is only half a comment.
>
> "...than they would be if they still had him"?
>
> or "...than they were when they had him"?
>
> One might imagine that 1036 is a better mark than 1567, but the
> Bulls' average score is 85.794.4, compared to 87.596.6 last
> No significant change in the scoring.
dropped
>
> Ron Artest is suddenly a star this year. Brad Miller and Marcus
> Fizer are suddenly serious players. Mercer and Hoiberg have
> off, but Anthony has come along, with Oakley, while nobody
Speculation. Would Fizer and Artest and Miller have "suddenly"
> significant has been dumped.
>
> With the coaching change, I would agree "the Bulls are better"; but
> with Brand they might actually be contending.
improved with Brand there? (I don't think Artest is a star, but
haven't fully looked at his numbers.) Maybe Brand was a negative
influence, keeping down the hopes of these guys. It's a plausible
story, if just because Brand was getting all the touches.
The only thing that is clear is that it wasn't that hard to make up
for Brand's loss; they didn't drop to a 4 win franchise and it's hard
to name any 15 win team that got worse by losing its "best player".
DeanO   In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
> > contribution that do sum to the team total, giving them a reality
conceptually. I
> > check that linear weights do not have. I like that
>
the team
> This is the part that is troublesome. It's nice that they sum to
> total, but on the whole I think that doesn't really tell us much
about the
> validity of the model. With suitable normalization, most or at any
close to
> rate many rating schemes could be made to have sums which come
> adding up to the team's win total.
I agree that normalization can make anything work and that is
cheating. The normalization I do is only on the "games" that
individuals play. I make that sum to 82 because there is no real
concept of what consists of an "individual game". The indiv win%
comes directly from the way James does it in baseball.
Dave Berri obviously had another method where the win sum was about
the team total. His results were different than mine. There are a
lot of rating schemes that can sum to the team's total. The
justification for using them is the process in which they were
developed. If you like Berri's theory, you can use his. At least he
makes an attempt to relate individual performance to team success. I
just personally don't like some aspects of how he does it. I don't
particularly like the determination of "individual games" in my
method, but I like it overall better than Berri's. And that's all
there is (maybe  I think Craven has something, but I don't know the
process).
>
conundrum),
> > don't claim it's predictive (why I pointed out the Brand
>
When we
> I suspect it's part of the datafitting problem in statistics.
> have a set of data, it's pretty easy to come up with a model that
fits
> that data set really well. But such models usually perform poorly
when
> used to make actual predictions on outofsample data (i.e. real
world
> predictions).
economist to
>
> Good predictive models are very hard to create. Just ask any
> try to predict when the next recession will come. Or any geologist
when
> the next big earthquake will hit Los Angeles.
Or ask a hydrogeologist like me when predicting how contaminants
>
migrate through groundwater. The best thing you can do is provide
ranges of realistic estimates, based on a rigorous process using as
much data as possible and using as much physics about the
interactions as possible. You can do better.
DeanO   In APBR_analysis@y..., "HoopStudies" <deano@r...> wrote:
>
For that matter, would Mercer and Hoiberg have slipped, with Brand
> Speculation. Would Fizer and Artest and Miller have "suddenly"
> improved with Brand there?
still there?
>(I don't think Artest is a star, but
Artest's standardized rates are (thru 23 games):
> haven't fully looked at his numbers.)
18.9 pts, 6.5 reb, 3.6 ast, 3.1 steals, 2.9 TO, .9 blocks
.516 combined shooting
The 3.1 steals rate leads the league by a large margin (2nd is
Iverson at 2.4)
> Maybe Brand was a negative
I was never discouraged by having a great player on my team, so I
> influence, keeping down the hopes of these guys.
> It's a plausible
> story, if just because Brand was getting all the touches.
don't fathom this thinking.
> The only thing that is clear is that it wasn't that hard to make up
If 15 wins is good enough, I guess.
> for Brand's loss;
>they didn't drop to a 4 win franchise and it's hard
player".
> to name any 15 win team that got worse by losing its "best
>
The "regression to the mean" principle would suggest that any 15win
> DeanO
team is almost certain to improve, no matter how you mix it up. On Wed, 6 Feb 2002, HoopStudies wrote:
[...]
> The only thing that is clear is that it wasn't that hard to make up
> for Brand's loss; they didn't drop to a 4 win franchise and it's hard
> to name any 15 win team that got worse by losing its "best player".
This is one of the extreme cases where the nonlinearities become
important. If we measure players by their individual wins and losses and
furthermore require that those sum to 15, then no Bull can appear to be
as highly productive as Jordan was with his 19 individual wins.
But in situations such as these where we're looking at wonloss
percentages, it's probably a better idea to look at odds instead of
probabilities, or even the logarithm of odds (known as a logit
transformation). Things may become linear with respect to odds or to
logits, which are nonlinear with respect to probabilities.
Two examples: odds ratios are what's behind the log5 method for
predicting win probabilties that some mathematician friend introduced Bill
James to.
And here's an example of how logits could be applied: For the 1567 Bulls
is ln(15/67) ~ 1.50. (Their odds of winning were 15/67 = .22, and their
probability of winning of course was 15/82 = .18.)
If we were lucky and life were relatively simple, Elton Brand's
contributions to the Bulls might be linear with respect to a logit, e.g.
subtracting him from the Bulls and replacing him with a pretty much
useless (for this year) high school player might hurt the Bulls to the
tune of 0.4 logits. For a team that had been 1567, the new logit would
be 1.9, the new odds would be exp(1.9) = .13, the new probability would
be .13/(.13+1) = .13, and the number of victories would be 10.7. So
losing Brand would cost the Bulls about 4 victories (which could of course
be counteracted by increased production from Artest, etc.  another
nonlinearity that we'd have to deal with).
Adding Brand to the Clippers, assuming that the other players' didn't
change (probably not a good assumption, nonlinearities again), would help
them by +.4 logits. So their 3151 2001 team, which had had a logit of
.50, now has a logit of .10, and therefore odds of .90, probability of
.475, and 39 wins. So Brand adds 8 wins to the Clips, in contrast to the
loss of 4 wins by the Bulls. (Obviously some of the Clips' wins would
therefore have to come at the expense of some team other than the Bulls,
nonlinearity again.)
That's a nice simple yet nonlinear model: Brand's quality measure stays
constant at .4 logits, but that translates into 4 marginal victories for
the Bulls and 8 marginal victories for the Clippers.
Unfortunately, this all assumes that (a) the logit function is the correct
functional form and (b) that the other players' production stays constant
(and of course there will be other roster changes which add even further
complications).
Life is undoubtedly not so simple, so I'm not claiming that that model
will actually work in terms of predictive value.
One thing which I've been meaning to try for years but never gotten around
to however is to use this kind of model to look just at rebounding. It's
a smaller, simpler task than trying to model offenses, defenses, or team
wins. It's clearly going to be a nonlinear process: if Tim Duncan gets
added to the Spurs and replaces ... who'd he replace, Carl Herrera?
Anyway, if Herrera was getting 4.5 rebounds per game and Duncan gets 12
per game, it is clearly not correct to predict that the Spurs will gain an
additional 7.5 rebounds per game. Some of Duncan's rebounds will come at,
so to speak, the expense of teammates. Yet he clearly should cause some
improvement to the Spurs' rebounding. I wonder if odds or logit measures
could be used so that players' rebounding quality stays constant even
though their teammates' and team's rebounds may change.
Such a measure wouldn't meet the "David Wesley" test that alleyoop2
suggested: we know that players' rebound stats will change when they
change positions (centers get more than power forwards, thanks to their
inside position). But it might pass the Greg Anthony test: a good
rebounder going to different teams or having different teammates (maybe
Dennis Rodman, maybe Vin Baker) might end up with a constant rebounding
score using these models, even though his reboundsper48minutes would
change.
MKT  In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
> > The only thing that is clear is that it wasn't that hard to make
up
> > for Brand's loss; they didn't drop to a 4 win franchise and it's
hard
> > to name any 15 win team that got worse by losing its "best
player".
>
losses and
> This is one of the extreme cases where the nonlinearities become
> important. If we measure players by their individual wins and
> furthermore require that those sum to 15, then no Bull can appear
to be
> as highly productive as Jordan was with his 19 individual wins.
Some of my final thoughts (before I go absolutely insane from working
>
on too many things at the same time).
1. I like playing devil's advocate. It was clear that Brand was the
best Bull last year and that he is at least a good player. His loss
was easily replaced because shaking up a bad team just helps improve
things. You add noise to your team when you make big changes to
raise it out of the consistent stinkhole that it resides in. No
offense to Chicagoans intended. It is hard to see how the Bulls
would be much better than their current record if Brand were
2. It is possible, though unlikely, for your hypothetical situation
to occur, where a Jordan with 19 wins plays on a team with 15 wins.
This is because I do not normalize to constrain the sum to 15, nor
did I develop the individual games to make the wins sum to the right
number.
3. Thanks for the info on logits. Saves me from reading about them
in some bori..., err, fun economics book! They do illustrate the
effect that definitely occurs in basketball, where one player is more
beneficial for one team than for another. I'm not convinced that
logits are the right functional form or how I'd use them yet. But
it's good to understand them well enough to consider it as I continue
research. Haven't seen the log5 method. Where is it?
4. If basketball were simple, none of us would be discussing this.
5. Where is Bob Chaikin? He has his simulation program that can be
used to determine the effect of replacing Brand with other guys. I'd
be curious to hear what it says.
Gotta go have cake.
DeanO
> But in situations such as these where we're looking at wonloss
to
> percentages, it's probably a better idea to look at odds instead of
> probabilities, or even the logarithm of odds (known as a logit
> transformation). Things may become linear with respect to odds or
> logits, which are nonlinear with respect to probabilities.
introduced Bill
>
> Two examples: odds ratios are what's behind the log5 method for
> predicting win probabilties that some mathematician friend
> James to.
67 Bulls
>
> And here's an example of how logits could be applied: For the 15
> is ln(15/67) ~ 1.50. (Their odds of winning were 15/67 = .22, and
their
> probability of winning of course was 15/82 = .18.)
e.g.
>
> If we were lucky and life were relatively simple, Elton Brand's
> contributions to the Bulls might be linear with respect to a logit,
> subtracting him from the Bulls and replacing him with a pretty much
the
> useless (for this year) high school player might hurt the Bulls to
> tune of 0.4 logits. For a team that had been 1567, the new logit
would
> be 1.9, the new odds would be exp(1.9) = .13, the new probability
would
> be .13/(.13+1) = .13, and the number of victories would be 10.7. So
course
> losing Brand would cost the Bulls about 4 victories (which could of
> be counteracted by increased production from Artest, etc.  another
didn't
> nonlinearity that we'd have to deal with).
>
> Adding Brand to the Clippers, assuming that the other players'
> change (probably not a good assumption, nonlinearities again),
would help
> them by +.4 logits. So their 3151 2001 team, which had had a
logit of
> .50, now has a logit of .10, and therefore odds of .90,
probability of
> .475, and 39 wins. So Brand adds 8 wins to the Clips, in contrast
to the
> loss of 4 wins by the Bulls. (Obviously some of the Clips' wins
would
> therefore have to come at the expense of some team other than the
Bulls,
> nonlinearity again.)
stays
>
>
> That's a nice simple yet nonlinear model: Brand's quality measure
> constant at .4 logits, but that translates into 4 marginal
victories for
> the Bulls and 8 marginal victories for the Clippers.
correct
>
> Unfortunately, this all assumes that (a) the logit function is the
> functional form and (b) that the other players' production stays
constant
> (and of course there will be other roster changes which add even
further
> complications).
model
>
> Life is undoubtedly not so simple, so I'm not claiming that that
> will actually work in terms of predictive value.
around
>
>
> One thing which I've been meaning to try for years but never gotten
> to however is to use this kind of model to look just at
rebounding. It's
> a smaller, simpler task than trying to model offenses, defenses, or
team
> wins. It's clearly going to be a nonlinear process: if Tim
Duncan gets
> added to the Spurs and replaces ... who'd he replace, Carl Herrera?
gets 12
> Anyway, if Herrera was getting 4.5 rebounds per game and Duncan
> per game, it is clearly not correct to predict that the Spurs will
gain an
> additional 7.5 rebounds per game. Some of Duncan's rebounds will
come at,
> so to speak, the expense of teammates. Yet he clearly should cause
some
> improvement to the Spurs' rebounding. I wonder if odds or logit
measures
> could be used so that players' rebounding quality stays constant
even
> though their teammates' and team's rebounds may change.
they
>
> Such a measure wouldn't meet the "David Wesley" test that alleyoop2
> suggested: we know that players' rebound stats will change when
> change positions (centers get more than power forwards, thanks to
their
> inside position). But it might pass the Greg Anthony test: a good
(maybe
> rebounder going to different teams or having different teammates
> Dennis Rodman, maybe Vin Baker) might end up with a constant
rebounding
> score using these models, even though his reboundsper48minutes
would
> change.
>
>
> MKT   HoopStudies <deano@...> wrote:
>  In APBR_analysis@y..., "mikel_ind" <msg_53@h...>
The Bulls appeared to be headed toward 1012 wins tops
> wrote:
> > >...... The Clips are better with Brand. The
> > > Bulls are better without Brand. So how good is
> Brand? Context
> > > sensitive.
> >
> > "The Bulls are better without Brand" is only half
> a comment.
> >
> > "...than they would be if they still had him"?
> >
> > or "...than they were when they had him"?
> >
> > One might imagine that 1036 is a better mark than
> 1567, but the
> > Bulls' average score is 85.794.4, compared to
> 87.596.6 last
> year.
> > No significant change in the scoring.
> >
> > Ron Artest is suddenly a star this year. Brad
> Miller and Marcus
> > Fizer are suddenly serious players. Mercer and
> Hoiberg have
> dropped
> > off, but Anthony has come along, with Oakley,
> while nobody
> > significant has been dumped.
> >
> > With the coaching change, I would agree "the Bulls
> are better"; but
> > with Brand they might actually be contending.
>
> Speculation. Would Fizer and Artest and Miller have
> "suddenly"
> improved with Brand there? (I don't think Artest is
> a star, but
> haven't fully looked at his numbers.) Maybe Brand
> was a negative
> influence, keeping down the hopes of these guys.
> It's a plausible
> story, if just because Brand was getting all the
> touches.
>
> The only thing that is clear is that it wasn't that
> hard to make up
> for Brand's loss; they didn't drop to a 4 win
> franchise and it's hard
> to name any 15 win team that got worse by losing its
> "best player".
early in the season. Then two things improved the team
dramatically. Bill Cartwright took over as coach and
Ron Artest returned.
Under Bill they're 713 with a 89.893.2 point diff.
Hardly a threat to the Lakers (even though they swept
that season series. Figure those odds), but definitely
out of the stinkhole. Artest has played like a star. I
wouldn't go so far as to call him a star just yet. He
plays in such a frenzied style that he can be
extremely erratic. I suspect he has a serious slump in
him before the season ends.
As for Brand, there's little doubt in my mind that
this would be a team on the fringes of the EC playoffs
had Brand stuck around. That said, I have to say that
I do like the deal the Bulls made. A solid, usable PF
is something that's pretty easy to find. Not one of
Brand's calibre, but one that a team can get by with.
Tyson Chandler has a chance to be a special player. It
may take another year or two, but the talent is
obvious. I've don't think I've ever seen a player this
tall who was as athletic.
Ed Weiland
__________________________________________________
Do You Yahoo!?
Send FREE Valentine eCards with Yahoo! Greetings!
http://greetings.yahoo.com   In APBR_analysis@y..., Ed Weiland <weiland1029@y...> wrote:
> Tyson Chandler has a chance to be a special player. It
You've said this before. Given how rarely the Bulls are on TV, it's
> may take another year or two, but the talent is
> obvious. I've don't think I've ever seen a player this
> tall who was as athletic.
no surprise I haven't seen him, but I am really curious, though. Who
do you think he is most similar to? Most people say "tall"
and "athletic" and they are referring to Kevin Garnett. Is that
realistic?
Dean Oliver
Journal of Basketball Studies  Imagine a taller, faster David Robinson with no offensive game.
Original Message
From: deano@...
Sent: Thursday, February 07, 2002 9:27 AM
To: APBR_analysis@yahoogroups.com; deano@...
Subject: [APBR_analysis] Re: nice methods
 In APBR_analysis@y..., Ed Weiland <weiland1029@y...> wrote:
> Tyson Chandler has a chance to be a special player. It
> may take another year or two, but the talent is
> obvious. I've don't think I've ever seen a player this
> tall who was as athletic.
You've said this before. Given how rarely the Bulls are on TV, it's
no surprise I haven't seen him, but I am really curious, though. Who
do you think he is most similar to? Most people say "tall"
and "athletic" and they are referring to Kevin Garnett. Is that
realistic?
Dean Oliver
Journal of Basketball Studies
To unsubscribe from this group, send an email to:
APBR_analysisunsubscribe@yahoogroups.com
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/  On Thu, 7 Feb 2002, HoopStudies wrote:
>  In APBR_analysis@y..., Ed Weiland <weiland1029@y...> wrote:
I'm curious too. I've only seen him once, in the LA Pro Summer League
> > Tyson Chandler has a chance to be a special player. It
> > may take another year or two, but the talent is
> > obvious. I've don't think I've ever seen a player this
> > tall who was as athletic.
>
> You've said this before. Given how rarely the Bulls are on TV, it's
> no surprise I haven't seen him, but I am really curious, though. Who
> do you think he is most similar to? Most people say "tall"
> and "athletic" and they are referring to Kevin Garnett. Is that
> realistic?
(actually I think it was called the Dada Summer League this past summer).
At that point he was extremely raw and the two crossroads for him that I
saw led to (a) Kevin Garnett and (b) Brad Sellers. It was absolutely too
early to tell which way he would end up.
People who have seen him more often, throughout the season, are in a much
better position to evaluate. But unless he starts making some obvious
Kobetype teenage strides, it could easily be 34 years before we know
where he'll end up. Or maybe it'll only take 12 years as Ed Weiland
says.
MKT  On Wed, 6 Feb 2002, Michael K. Tamada wrote:
> Two examples: odds ratios are what's behind the log5 method for
I deleted the email, but I think someone asked about Bill James' "log5
> predicting win probabilties that some mathematician friend introduced Bill
> James to.
method". That's simply his name for his formula (not as well known but
much cleverer than his Pythagorean formula) for calculating the expected
win probability when, say, a 75% winprobability team plays a 25%
winprobability team. I don't know where the log or the 5 comes from, but
the formula can be derived from standard probability formulas, I think
with a small assumption about functional form thrown in.
The really fantastic more general version of the formula is for situations
which are not inherently 5050 balanced, such as batters' probability of
getting a hit against a pitcher. Someone told me that version can also be
derived from probability theory, but I haven't been able to do it.
Despite the name, the formulas use odds ratios, not logarithms. Actually
come to think of it I don't think Bill James put the formulas in terms of
odds ratios, he used probabilities. But the formulas are much simpler
and cleaner when cast in odds terms.
MKT
P.S. For those who are interested, the formulas.
First, odds are calculated from probabilities by this definition:
odds = p/(1p)
e.g. a 75% probability is equivalent to 75/25 = 3:1 odds.
The log5 formula says to find the probability of Team A beating Team B,
look at Team A's odds of winning (based on their overall wonloss record,
or whatever source of probability estimates you want to use), call those
OddsA. Call Team B's odds OddsB.
Then Team A's odds of winning against Team B are simply OddsA/OddsB.
Example: if A wins 75% of the time, and B wins 25% of the time, then
intuitively Team A should have an extremely high probability of beating
Team B, we're talking Sacramento Kings vs Chicago Bulls. Their respective
odds are 3 and 1/3, so Team A's odds of beating Team B are 3/(1/3) = 9.
9:1 odds convert into a 90% probability, using the inverse of the odds
definition: p = (odds/(1+odds).
Next, the more general version of the formula, let's use a baseball
example: a .333 hitter faces a pitcher who gives up hits at a .333 rate.
What is the expected probability that the batter will get a hit?
Intuitively, it's going to be a lot larger than .333, because we're
clearly talking about an ace hitter and a bad pitcher here.
Because the overall odds (overall meaning throughout baseball) of getting
a hit are not 1:1 (the probability is not 5050, unlike teams' wonloss
records), we need a third parameter: the overall odds of getting a hit.
Let's assume that on average, batters hit .280. So their typical odds are
.280/.720 = .389, lets call this OddsO for overall odds.
The .333 hitter's odds, let's call them OddsH, are .333/.667 = .500. For
the pitcher, we need to look at his odds of success (not his odds of
giving up a hit); the pitcher gets batters out .667 of the time so his
odds of success are .667/.333 = 2, let's call the pitcher's odds OddsP.
The log5 formula for the batter's odds of getting a hit are
(OddsB/OddsP)/OddsO
i.e. the same formula, but divided by the overall odds, OddsO.
In our example this is (.500/2)/.389 = 9/14 = .643.
Converting that into probabilities, the batter has a .643/(1+.643) = .391
probability of getting a hit.
Obviously these are general, overall calculations. There may be
individual quirks in the matchup that cause one player or team to do
unusually well or poorly against certain opponents (Bulls vs Lakers this
season, although that's probably a random fluke).
Bill James said in one of his Baseball Abstracts that he got these
formulas from a mathematician friend, but I have not seen a complete
citation or derivation of them.
MKT   "Michael K. Tamada" <tamada@...> wrote:
>
Garnett is probably a stretch. One spin on draft day
>
> On Thu, 7 Feb 2002, HoopStudies wrote:
>
> >  In APBR_analysis@y..., Ed Weiland
> <weiland1029@y...> wrote:
> > > Tyson Chandler has a chance to be a special
> player. It
> > > may take another year or two, but the talent is
> > > obvious. I've don't think I've ever seen a
> player this
> > > tall who was as athletic.
> >
> > You've said this before. Given how rarely the
> Bulls are on TV, it's
> > no surprise I haven't seen him, but I am really
> curious, though. Who
> > do you think he is most similar to? Most people
> say "tall"
> > and "athletic" and they are referring to Kevin
> Garnett. Is that
> > realistic?
>
> I'm curious too. I've only seen him once, in the LA
> Pro Summer League
> (actually I think it was called the Dada Summer
> League this past summer).
> At that point he was extremely raw and the two
> crossroads for him that I
> saw led to (a) Kevin Garnett and (b) Brad Sellers.
> It was absolutely too
> early to tell which way he would end up.
tried to sell Chandler and Curry as a future
ShaqGarnett tandem, which is pretty ridiculous, IMO.
I doubt Chandler will ever develop Garnett's
allaround game. He just doesn't seem to have the same
personality. He seems more like a Webber/Barkley type
personalitywise. That being an OK guy who sometimes
rubs folks the wrong way. Not exactly an MJ when it
comes to leadership. That's just my initial take on
him though. I could easily be way off the mark here. I
was concerned about Chandler being another Brad
Sellers at first too. I doubt that will happen.
Chandler is more athletic than Sellers and he doesn't
play soft. Reckless yes, but not soft.
Comparing him to other guys who skipped college in
their rookie years, right now his offense comes up
short compared to Garnett, Kobe and TMac. It seems
that most of Chandler's points come from put backs,
alleyoops and other high percentage shots. He's said
to have good range, but I've yet to see it during a
game. He also commits a ton of turnovers. On defense
he looks pretty good. He uses his height and quickness
very well. He does get burned on occasion, but that's
to be expected from any rookie, let alone one straight
from the preps. Also, he doesn't look completely
overmatched, as Jonathan Bender did his first couple
of seasons. Most of Chandler's problems seem to come
from being tentative on the court. That's a common
problem with rooks and it usually corrects itself in
time.
As far as comparing him to one guy, I would say
Stuart's assessment "A taller, quicker David Robinson
with no offense" is about as accurate as any. I
suppose the more pessimistic types would call him a
Brad Lohaus who can jump.
> People who have seen him more often, throughout the
Keeping in mind of course that Ed Weiland is just a
> season, are in a much
> better position to evaluate. But unless he starts
> making some obvious
> Kobetype teenage strides, it could easily be 34
> years before we know
> where he'll end up. Or maybe it'll only take 12
> years as Ed Weiland
> says.
fan and possibly an overly optimistic one at that, one
might be better off trusting the experts here. Since
Chandler has moved into the starting lineup, I'm
guessing he'll get about 1000 more minutes this
season. At season's end we should have a better idea
of what he's going to become.
Ed Weiland
__________________________________________________
Do You Yahoo!?
Send FREE Valentine eCards with Yahoo! Greetings!
http://greetings.yahoo.com  I hate to mention this, but can we change the Subject of the posts
when the topic changes?
I personally don't have too much of an interest in Tyson Chandler,
but I would love to read more about the mathmatical methods used
here. Changing the subject prevents everyone from reading through
irrelevant material.
Thanks,
LKM
 In APBR_analysis@y..., Ed Weiland <weiland1029@y...> wrote:
>
>  "Michael K. Tamada" <tamada@o...> wrote:
> >
> >
> > On Thu, 7 Feb 2002, HoopStudies wrote:
> >
> > >  In APBR_analysis@y..., Ed Weiland
> > <weiland1029@y...> wrote:
> > > > Tyson Chandler has a chance to be a special
> > player. It
> > > > may take another year or two, but the talent is
> > > > obvious. I've don't think I've ever seen a
> > player this
> > > > tall who was as athletic.
> > >
> > > You've said this before. Given how rarely the
> > Bulls are on TV, it's
> > > no surprise I haven't seen him, but I am really
> > curious, though. Who
> > > do you think he is most similar to? Most people
> > say "tall"
> > > and "athletic" and they are referring to Kevin
> > Garnett. Is that
> > > realistic?
> >
> > I'm curious too. I've only seen him once, in the LA
> > Pro Summer League
> > (actually I think it was called the Dada Summer
> > League this past summer).
> > At that point he was extremely raw and the two
> > crossroads for him that I
> > saw led to (a) Kevin Garnett and (b) Brad Sellers.
> > It was absolutely too
> > early to tell which way he would end up.
>
> Garnett is probably a stretch. One spin on draft day
> tried to sell Chandler and Curry as a future
> ShaqGarnett tandem, which is pretty ridiculous, IMO.
> I doubt Chandler will ever develop Garnett's
> allaround game. He just doesn't seem to have the same
> personality. He seems more like a Webber/Barkley type
> personalitywise. That being an OK guy who sometimes
> rubs folks the wrong way. Not exactly an MJ when it
> comes to leadership. That's just my initial take on
> him though. I could easily be way off the mark here. I
> was concerned about Chandler being another Brad
> Sellers at first too. I doubt that will happen.
> Chandler is more athletic than Sellers and he doesn't
> play soft. Reckless yes, but not soft.
>
> Comparing him to other guys who skipped college in
> their rookie years, right now his offense comes up
> short compared to Garnett, Kobe and TMac. It seems
> that most of Chandler's points come from put backs,
> alleyoops and other high percentage shots. He's said
> to have good range, but I've yet to see it during a
> game. He also commits a ton of turnovers. On defense
> he looks pretty good. He uses his height and quickness
> very well. He does get burned on occasion, but that's
> to be expected from any rookie, let alone one straight
> from the preps. Also, he doesn't look completely
> overmatched, as Jonathan Bender did his first couple
> of seasons. Most of Chandler's problems seem to come
> from being tentative on the court. That's a common
> problem with rooks and it usually corrects itself in
> time.
>
> As far as comparing him to one guy, I would say
> Stuart's assessment "A taller, quicker David Robinson
> with no offense" is about as accurate as any. I
> suppose the more pessimistic types would call him a
> Brad Lohaus who can jump.
>
> > People who have seen him more often, throughout the
> > season, are in a much
> > better position to evaluate. But unless he starts
> > making some obvious
> > Kobetype teenage strides, it could easily be 34
> > years before we know
> > where he'll end up. Or maybe it'll only take 12
> > years as Ed Weiland
> > says.
>
> Keeping in mind of course that Ed Weiland is just a
> fan and possibly an overly optimistic one at that, one
>
> might be better off trusting the experts here. Since
> Chandler has moved into the starting lineup, I'm
> guessing he'll get about 1000 more minutes this
> season. At season's end we should have a better idea
> of what he's going to become.
>
> Ed Weiland
>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Send FREE Valentine eCards with Yahoo! Greetings!
> http://greetings.yahoo.com  Fine idea on the topic change.
Regarding Chandler, I see him as a 7'1" Ralph Sampson. Like Sampson,
he has intriguing physical skills, including something of a jumper,
and makes a lot of athletic plays, but has no post game and a strange
allergy to defensive rebounds.
I saw Chandler play 3 games in Oregon as a high schooler. Some of you
brought it up before, but what struck me is that he didn't seem to
have that disposition to dominate. One team put a 6'3" football
player on him and the guy totally got under Chandler's skin; his team
was actually lucky to even win the game.
That made me think he didn't have the maturity to leap to the NBA,
but perhaps I was wrong on that. He's certainly been better than the
other high schoolers, although I believe his first year numbers will
end up short of what Garnett and TMac did.
 In APBR_analysis@y..., "lk_maxwell" <lk_maxwell@h...> wrote:
> I hate to mention this, but can we change the Subject of the posts
> when the topic changes?
>
> I personally don't have too much of an interest in Tyson Chandler,
> but I would love to read more about the mathmatical methods used
> here. Changing the subject prevents everyone from reading through
> irrelevant material.
>
> Thanks,
>
> LKM
>
>  In APBR_analysis@y..., Ed Weiland <weiland1029@y...> wrote:
> >
> >  "Michael K. Tamada" <tamada@o...> wrote:
> > >
> > >
> > > On Thu, 7 Feb 2002, HoopStudies wrote:
> > >
> > > >  In APBR_analysis@y..., Ed Weiland
> > > <weiland1029@y...> wrote:
> > > > > Tyson Chandler has a chance to be a special
> > > player. It
> > > > > may take another year or two, but the talent is
> > > > > obvious. I've don't think I've ever seen a
> > > player this
> > > > > tall who was as athletic.
> > > >
> > > > You've said this before. Given how rarely the
> > > Bulls are on TV, it's
> > > > no surprise I haven't seen him, but I am really
> > > curious, though. Who
> > > > do you think he is most similar to? Most people
> > > say "tall"
> > > > and "athletic" and they are referring to Kevin
> > > Garnett. Is that
> > > > realistic?
> > >
> > > I'm curious too. I've only seen him once, in the LA
> > > Pro Summer League
> > > (actually I think it was called the Dada Summer
> > > League this past summer).
> > > At that point he was extremely raw and the two
> > > crossroads for him that I
> > > saw led to (a) Kevin Garnett and (b) Brad Sellers.
> > > It was absolutely too
> > > early to tell which way he would end up.
> >
> > Garnett is probably a stretch. One spin on draft day
> > tried to sell Chandler and Curry as a future
> > ShaqGarnett tandem, which is pretty ridiculous, IMO.
> > I doubt Chandler will ever develop Garnett's
> > allaround game. He just doesn't seem to have the same
> > personality. He seems more like a Webber/Barkley type
> > personalitywise. That being an OK guy who sometimes
> > rubs folks the wrong way. Not exactly an MJ when it
> > comes to leadership. That's just my initial take on
> > him though. I could easily be way off the mark here. I
> > was concerned about Chandler being another Brad
> > Sellers at first too. I doubt that will happen.
> > Chandler is more athletic than Sellers and he doesn't
> > play soft. Reckless yes, but not soft.
> >
> > Comparing him to other guys who skipped college in
> > their rookie years, right now his offense comes up
> > short compared to Garnett, Kobe and TMac. It seems
> > that most of Chandler's points come from put backs,
> > alleyoops and other high percentage shots. He's said
> > to have good range, but I've yet to see it during a
> > game. He also commits a ton of turnovers. On defense
> > he looks pretty good. He uses his height and quickness
> > very well. He does get burned on occasion, but that's
> > to be expected from any rookie, let alone one straight
> > from the preps. Also, he doesn't look completely
> > overmatched, as Jonathan Bender did his first couple
> > of seasons. Most of Chandler's problems seem to come
> > from being tentative on the court. That's a common
> > problem with rooks and it usually corrects itself in
> > time.
> >
> > As far as comparing him to one guy, I would say
> > Stuart's assessment "A taller, quicker David Robinson
> > with no offense" is about as accurate as any. I
> > suppose the more pessimistic types would call him a
> > Brad Lohaus who can jump.
> >
> > > People who have seen him more often, throughout the
> > > season, are in a much
> > > better position to evaluate. But unless he starts
> > > making some obvious
> > > Kobetype teenage strides, it could easily be 34
> > > years before we know
> > > where he'll end up. Or maybe it'll only take 12
> > > years as Ed Weiland
> > > says.
> >
> > Keeping in mind of course that Ed Weiland is just a
> > fan and possibly an overly optimistic one at that, one
> >
> > might be better off trusting the experts here. Since
> > Chandler has moved into the starting lineup, I'm
> > guessing he'll get about 1000 more minutes this
> > season. At season's end we should have a better idea
> > of what he's going to become.
> >
> > Ed Weiland
> >
> >
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Send FREE Valentine eCards with Yahoo! Greetings!
> > http://greetings.yahoo.com   In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
>
introduced Bill
>
> On Wed, 6 Feb 2002, Michael K. Tamada wrote:
>
> > Two examples: odds ratios are what's behind the log5 method for
> > predicting win probabilties that some mathematician friend
> > James to.
James' "log5
>
> I deleted the email, but I think someone asked about Bill
> method". That's simply his name for his formula (not as well known
but
> much cleverer than his Pythagorean formula) for calculating the
expected
> win probability when, say, a 75% winprobability team plays a 25%
from, but
> winprobability team. I don't know where the log or the 5 comes
> the formula can be derived from standard probability formulas, I
think
> with a small assumption about functional form thrown in.
situations
>
> The really fantastic more general version of the formula is for
> which are not inherently 5050 balanced, such as batters'
probability of
> getting a hit against a pitcher. Someone told me that version can
also be
> derived from probability theory, but I haven't been able to do it.
Actually
>
> Despite the name, the formulas use odds ratios, not logarithms.
> come to think of it I don't think Bill James put the formulas in
terms of
> odds ratios, he used probabilities. But the formulas are much
simpler
> and cleaner when cast in odds terms.
Things are coming together for me. I didn't know the method was
>
called log5. I called them matchup probabilities and use them a lot
myself. I can't say that I could quite derive the formula either (it
always seemed that the league average had to be some sort of prior
probability, if you framed it in a Bayes perspective). James said he
got the formula from Dallas Adams. I asked him once about a citation
and he didn't give me a specific one, so I spent some time looking
for it in math/stat journals and couldn't find it there.
>
Or I've got them documented at
> MKT
>
>
> P.S. For those who are interested, the formulas.
>
http://www.rawbw.com/~deano/methdesc.html#matchup
Dean Oliver
Journal of Basketball Studies  On Fri, 8 Feb 2002, HoopStudies wrote:
>  In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
[...]
> > Despite the name, the formulas use odds ratios, not logarithms.
I found a sort of citation at
> Actually
> > come to think of it I don't think Bill James put the formulas in
> terms of
> > odds ratios, he used probabilities. But the formulas are much
> simpler
> > and cleaner when cast in odds terms.
> >
>
> Things are coming together for me. I didn't know the method was
> called log5. I called them matchup probabilities and use them a lot
> myself. I can't say that I could quite derive the formula either (it
> always seemed that the league average had to be some sort of prior
> probability, if you framed it in a Bayes perspective). James said he
> got the formula from Dallas Adams. I asked him once about a citation
> and he didn't give me a specific one, so I spent some time looking
> for it in math/stat journals and couldn't find it there.
http://www.baseballprospectus.com/news/19980728kushner.html
Apparently Bill James first published the formula in his 1981 Baseball
Abstract (I didn't buy my first one until 1983 and stupidly gave it away
as a present). It sounds as though he came up with the formula himself,
and Dallas Adams provided empirical rather than theoretical evidence in
favor of the formula.
In essence, James' log5 measures of team quality are simply half the
team's winning odds. This permits him to use an additive formula for
calculating expected win probabilities  if Team A has log5 value "Alog5"
and Team B has "Blog5", then in James' formula, Team A's probability of
winning is Alog5/(Alog5+Blog5).
The author of the article in the URL, James Kushner, points out the log5
is simply equal to W/2L, where W is the team's wins and losses.
What these guys are all missing is that "W/2L" is in fact onehalf the
team's odds of winning, and that things become simpler still when
you use odds instead of probabilities. I mean you can't beat a formula
like OddsA/OddsB; that's even simpler than the formula above. Or to put
it another way, the "2" in "W/2L" is redundant.
I've written to Bill James a couple of times about using odds ratios (he
once asked readers for a formula for measuring home field advantage, and I
sent him essentially the formula that DeanO has on his website, and I
pointed out how this all flows simply and naturally from looking at odds
ratios), but never got a reply.
There was apparently discussion of "Dallas Adams formula" in 1997 on the
SABR listserv  see
http://www.pacificnet.net/~sroney/SABRL/index97.html
but I couldn't access the listserv archives, presumably because I'm not a
member of SABR.
MKT On Fri, 8 Feb 2002, Ed Weiland wrote:
[...]
> might be better off trusting the experts here. Since
> Chandler has moved into the starting lineup, I'm
> guessing he'll get about 1000 more minutes this
> season. At season's end we should have a better idea
> of what he's going to become.
Yeah, the starting, and the accompanying minutes should
have two benefits: he'll presumably learn more and faster,
and we observers will get a much better sense of how he's
developing.
And I think in cases like this, firsthand observations
such as the ones you've made are necessary. Those of us from
afar can only look at his stats, which will probably be lousy.
But for a 19year old rookie, lousy stats have to be expected,
it what's you see him doing or not doing on the court that is
probably a better forecast of his future.
MKT
Your message has been successfully submitted and would be delivered to recipients shortly.