Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
From: [email protected] (Jonathan King)
Subject: teams as organisms; stats or "stats" (was Re: Jack Morris)
Organization: University of California, San Diego
Lines: 124
Distribution: na
NNTP-Posting-Host: cogsci.ucsd.edu
Summary: let's talks stats and not "stats"
Note: I'm not posting this as part of an argument with Roger Meynard,
but as an independent sort of thread. I do actually quote some things
that Roger Meynard wrote, but it might be better to think of this as
"sampling" his post (in the hip-hop sense) because it fits in with what
I want to say.
[email protected] (Roger Maynard) writes:
>But the point is that the only decision making pro-
>cess used to determine the "best" is the score of the game and it re-
>lates to the *teams*. Not the individual players. There is no method
>inherent in baseball of comparing individual performances. And that
>is how it should be, because, after all, baseball is a team game.
There's an interesting parallel between this way of viewing a baseball
team and some people's conception of a biological organism. In the
biology context, we would very likely read "fitness" for "the score of
the game" and "organisms" for "teams". How we interpret "players" is
trickier, but either "organs", or "genes" might seem reasonable
choices depending on what point we were trying to make. A "genes"
interpretation actually might be really interesting in this case,
but that would be a different and probably longer post.
If, however, we take the "organ" view, then our knowledge of biology
should make us pause before we start saying things like "species X is
more fit than species Y because of a better organ Z". Given what we
know about the interdependence of organs, we would often be suspicious
of such claims. (But note that this type of argument is quite often
made when you map "species X" onto 'humans', and "organ Z" onto
'brain'). On the other hand, some statements of this kind do seem
more reasonable than others, as far as we can test them (e.g. 'brain'
above might be more reasonable than 'pancreas' assuming no gross
pathology, particularly if species Y is a primate).
Even when you make such statements, you should be concerned with the
functioning of the whole organism, and the possibility that one organ
might be more crucial for one species and a second organ in another.
(Not to mention the possibility that no organ is particularly crucial
in some third species.) However, if we are non-vitalists with any
kind of reductionsit streak, we will want to say that an organism is
not some completely magical unanalyzable "whole" but an intriguing
process made up of various subprocesses that interact in ways that are
potentially observable. Some of these processes might be localized to
particular organs, while others may be distributed across multiple
organs. In a way, this is just like a baseball team, except that I
think it is pretty clear that the processes and interactions involved
in baseball are *much* simpler and less numerous than in most organisms.
>To say that one player is better than another is to be able to say ab-
>solutely that player A's team would have played better with player B
>in their lineup. Sheer speculation. Impossible to ascertain.
One thing that is quite difficult about baseball is that perfectly
controlled experiments are sometimes very tough to do. But, of
course, this has never stopped researchers from doing the best they
can, and sometimes deriving very powerful conclusions even in the
absence of certainty. Most of this goes far beyond sheer speculation,
but even sheer speculation can motivate further interesting research.
>If you want to select a group of statistics and claim that Clemens
>has done better [than another pitcher] with those statistics as a
>criteria, then fine.
In this cases, we're seeing the word "statistics" means "summary of
observed events", where the events themselves can be viewed as the
output of some process, and possibly inputs for other processes.
Thus, if we have any valid notion of how the processes are put together
into the functioning organism, data in the form of statistics might
give us a basis to test particular hypotheses.
>But you have to be able to prove that those statistics measure the
>individual's contribution to winning the WS - because that is the only
>measure of "best" that has any meaning in the context of baseball.
This statement brings us back to the concept of fitness again.
Fitness is defined in terms of both an organism and its environment;
you might be fit in one situation and not another. Moving to
baseball, it is clear that each team spends the entire season in an
environment including all the other teams in the league. In at least
a nominal sense, the division winners are the fittest teams in the
league, in that they (on average) had better fitness scores than any
of their competing opponennts. But in a real sense, there is a fairly
large random component in the performance of each team that is
difficult if not impossible to account for in terms of factors
intrinsic to (or interesting for) baseball. The same is true in
biology. But here is also no direct biological equivalent of the
World Series in basebal. In the world series, the random component
may be greatly magnified by the small number of games that are played,
and both teams suddenly experience huge changes from the environement
where they were originally successful. It might be fun to watch, but
it's unclear what it all really means.
***
Now just one more un-related point:
>I have yet to see that any of you can predict a
>WS winner with any greater accuracy than Jeanne Dixon.
On the other hand, you have seen some of us who can predict the
outcome of the divisional races better than a random assignment of
teams to finishes, and maybe some of us (e.g. me) who can do this
better than the other participants in this forum on a regular basis.
But this is probably only due to the fact that a 162-game schedule
gives you a little hope that bad hops aren't the only difference
between the winners and the losers.
Moreover, you've had the opportunity to see some analysis of the World
Series situation that makes the strong claim that *nobody* can predict
the WS winner with reliably greater accuracy than a coin biased only
to reflect the well-known home vs. road effect on winning percentage.
>The stats are a nice hobby and that's about it. There is no new
>knowledge being produced.
Since stats are summaries of events, it's true that if you know the
events you can derive the stats. But if somebody is trying to
understand the process behind the stats, then the stats produce new
knowledge, and some of this might even be reliable, repeatable, and
useful. Speaking of which, I should get back to producing knowledge
in a different field. That is, of course, if I can produce knowledge
even though I'm relying on stats to do it.
jking
