Re: [APBR_analysis] Re: Tendex rating
- On Fri, 2 Nov 2001, Dean Oliver wrote:
> Model analysis can be evaluated on 3 (or 4) philosophies:[...]
> 1. Simplicity
> 2. Reality
> 3. Conservatism
> (4. Consistency)
>#2. My favorite example is that people used to think that orbitalI think the word you're looking for is "epicycles". Unless you're looking
>physics was all built on circles. The sun went around the earth in a
>big circle. If it wasn't just a circle, it was circles within
>circles. There was a good name for all of this that fails me right
>now and I know someone out there can spell this out better. People
for the more general term of "adding more and more little adjustments to
the model to make it fit reality, so much so that the model starts falling
apart under its own complexity". Thomas Kuhn (referred to in Stuart
McKibbin's posting) might've had a word for this too, but I forget. When
the old model gets pushed aside by a new different one, that's a "paradigm
Hey didn't you go to Caltech and shouldn't you know this stuff? ;) But
you did nail the correct answer to the ellipse modeller: Kepler.
> What's the 3rd thing? Well, conservativeness is important when usingI'm not sure about this one. Because an overly "conservative" list of
> a model to set policy (something I do at work). Say you're
> interested in only the good defensive players in the league (for some
> reason) and you want your stat to get those guys. Well, you want to
> make sure you don't get the mediocre ones. Your statistic should
> come out and there should be no argument that the method you chose is
> only going to get good defensive players. Not sure it matters much
> for basketball, except if you're helping the league to rewrite rules.
good defensive players will STILL get arguments -- from people who
complain that the list left out players X, Y, and Z, who are great
It's analogous to statistics: you can be "conservative" and minimize the
probability of a Type I error by choosing a small significance level. But
in doing so, you are automatically raising the probability of a Type II
A list of "great defenders" which is conservative will avoid Type I
errors, but will be making more Type II errors.
Decision theory tells us we should look at the relative cost of Type I
errors and Type II errors and choose a signficance level which balances
out the likely errors so as to minimize the costs.
In other words, sometimes we want a "conservative" list of great
defenders, but other times a not so conservative one. Depending on the
purpose of the list.
> I frankly hate the 4th one. A lot of times, someone has doneWell there's another kind of consistency, one which is a good thing to
> something stupid before, but because of "consistency", we have to do
> the same stupid thing again.
have: logical self-consistency. E.g. rating systems should avoid
double-counting (unless there is a reason to put a heavy weight on that
--- In APBR_analysis@y..., "Dean Oliver" <deano@t...> wrote:
> Hmm, one of the more controversial subjects here, I think, is the
> Admiral. I heard more complaints about him being in the NBA Top 50
> than about anyone else. No killer instinct, they said.
David Robinson is probably a great person, as well as a colossal
figure in the history of the game.
He left a clue to his attitude in an interview I saw some years ago.
Invoking his Christian ideals, he said his faith in God elevates his
game; and that (quoting from memory) "when going up against an
Olajuwon or a Robert Horry, I feel like David against Goliath..."
In other words, he doesn't feel physically superior to Robert Horry?
> How has the Admiral performed in the playoffs relative to the
> season, then relative to other people? Has he fallen off much more
> than others in the playoffs? .......
Per-36-minute standardized equivalents for David Robinson, career
Games Min. Pct. Sco. Reb. Ast PF Stl TO Blk Total
Regular 845 35.8 .570 26.5 11.7 2.8 3.1 1.5 2.7 3.3 45.7
Playoff 96 37.5 .532 22.8 12.2 2.8 3.4 1.4 2.6 2.9 41.8
The ratio of this playoff rate to regular season rate is .915.
The "average" playoff rate is around .940 (for 515 alltime players).
> Mike -- this statement:
> > What makes David over-the-hill is that he cannot produce at that
> > rate for as many minutes as before.
> The implication is that people can be as good, as efficient as they
> get older, but just not as long. Was his "rate" (per minute) as
> before as it is now? Any sense for how this goes down with age
> relative to minutes played?
> Dean Oliver
> Journal of Basketball Studies
It depends on the player, the team situation, and the coach. Popovich
may be "babying" DRob by using him only 30 mpg, but it has been
several years now since he has had back problems and other ailments.
Jordan seems willing to go 40 minutes at age 38, but that is just
another chapter-in-the-making of a unique case.
Per-minute rates are down for any player past his prime. The more
involved issue is : how to get the most production. Over the course
of a whole season (playoffs particularly), and over the course of a
- --- In APBR_analysis@y..., "Michael K. Tamada" <tamada@o...> wrote:
>stuff? ;) But
> Hey didn't you go to Caltech and shouldn't you know this
> you did nail the correct answer to the ellipse modeller: Kepler.Hey, going to Caltech means I know the formulas, how to derive them
from first principles, what all the little math symbols mean, and how
to go 40 hours without sleeping. It doesn't mean I know the names of
the famous dead guys. Heck, by going to Caltech, I have a right to
> I'm not sure about this one. Because an overly "conservative" list
> good defensive players will STILL get arguments -- from people whominimize the
> complain that the list left out players X, Y, and Z, who are great
> It's analogous to statistics: you can be "conservative" and
> probability of a Type I error by choosing a small significancelevel. But
> in doing so, you are automatically raising the probability of aType II
> error.This is pretty much my point. In policy making, a policy maker
really wants to reduce one of those types of errors. Usually the
policy makers don't care about the cost-benefit of Type I vs. Type II
errors. Their job is minimize one type and fight with everyone else
who wants to minimize the other type.
Not sure how that relates to any hoops stuff we're doing right now,
but it might in the future, when we start using all those funky math
> Well there's another kind of consistency, one which is a good thingto
> have: logical self-consistency. E.g. rating systems should avoidthat
> double-counting (unless there is a reason to put a heavy weight on
> variable).That's true and a good . I can add that to the routine speech I give
at work. Then when my people look at me funny, I can blame it on you.
I would phrase what you're talking about a little differently here,
though. A method makes assumptions at the start and those
assumptions should remain true at the end. Thinking off the top of
my head -- if Tendex assumes all those things to be worth one point,
shouldn't they all be worth one point at the end, too? Does this
mean Tendex should add up to points scored?
I know all the arguments pro and con with Tendex. I always point out
that Tendex, when applied to teams has only about a 70-80%
correlation with winning percentage. Which means it's not terribly
reliable for predicting winning teams and probably no more reliable
for predicting winning players. I also don't like the fact that it
really just encourages a lot of shooting. I don't know how well
Tendex correlates with points scored. But unlike baseball, where
position players are pretty much responsible for offense and pitchers
for defense, basketball players are responsible for both. That's why
I try to keep offensive and defensive contributions separate for
individuals. Doug Steele has an offensive and defensive Tendex
rating based on some conversations we had in '94. It was a start.