Fielder Model

CSGnet Archive CSG2000
1

[From Rick Marken (2000.09.23.0950)]

Bruce Abbott (2000.09.22.1620 EST)

(For example, Rick and I differ in our opinions of how
successful his fielder model is in capturing the details of
real fielding.)

I didn't know that. I'd really like to hear what you have to
say about the success (and lack thereof) of the model. I think
I am aware of its limitations but I would really like to hear
your comments. It might help me get back into working on it
again.

By the way, did you know that my paper describing the version
of the model that is on the net was accepted for publication
in _American Journal of Psychology_? It's scheduled to appear
in Fall of 2001.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: marken@mindreadings.com
mindreadings.com

2

[From Bruce Abbott (2000.09.23.2240 EST)]

Rick Marken (2000.09.23.0950) --

Bruce Abbott (2000.09.22.1620 EST)

(For example, Rick and I differ in our opinions of how
successful his fielder model is in capturing the details of
real fielding.)

I didn't know that. I'd really like to hear what you have to
say about the success (and lack thereof) of the model. I think
I am aware of its limitations but I would really like to hear
your comments. It might help me get back into working on it
again.

It's been quite a while since I played with the model, and my recollections
are now more than a little fuzzy. Shortly after you briefly described your
fielder model on CSGnet I wrote my own version of the fielder simulation,
using your description as a starting point and working out the rest
logically. With certain parameter values and ball trajectories the fielder
would initially back up but then reverse direction and run like crazy for
the ball, in some cases failing to reach the ball before it hit ground.
This behavior resembles what real fielders sometimes do, and as I recall,
this resemblance is partly why you felt at the time that the model must be
on the right track.

Now here's where my memory gets fuzzy, but my I recall becoming dissatisfied
with my fielder's performance after observing it under a variety of
conditions (different ball trajectories, limits to running speed, etc.) --
it just didn't look realistic. If I'm recalling correctly (and I'm not at
all sure of this), it would always back up while the ball was rising and
then reverse direction as the trajectory reached its peak. Real fielders
may do this on occasion but in my experience it is not a stable feature of
their fly-fielding behavior. So there must be something else that fielders
control for when fielding a pop fly, either in addition to or instead of the
variable employed in your fielder demo.

By the way, did you know that my paper describing the version
of the model that is on the net was accepted for publication
in _American Journal of Psychology_? It's scheduled to appear
in Fall of 2001.

Congratulations! How did they come to accept a paper with no empirical data
against which to test the model? Or did you provide such data? Did you say
Fall of 200_1_? I've heard of publication delays, but that's incredible.
How about summarizing the paper for us?

Bruce A.

3

[From Rick Marken (2000.09.21.0900)]

Bruce Abbott (2000.09.23.2240 EST)--

It's been quite a while since I played with the model

Try the latest version at:

http://home.earthlink.net/~rmarken/ControlDemo/CatchXY.html

The site also presents some of the relevant data.

I recall becoming dissatisfied with my fielder's performance
after observing it under a variety of conditions (different
ball trajectories, limits to running speed, etc.) -- it just
didn't look realistic.

Again, you can compare the plots of model behavior to plots
of actual fielder behavior at the site listed above.

it would always back up while the ball was rising and
then reverse direction as the trajectory reached its peak.

Yes. The amount of backup can be adjusted by changing parameter
values. Including a vertical component eliminates a lot of
backup; most of the backup occurs when the ball is his directly
toward the fielder, and this shows up in the data as well.

How did they come to accept a paper with no empirical data
against which to test the model?

There was empirical data; see the site above.

Or did you provide such data?

Yes. I used other people's data (mainly that of McBeath,
Schaeffer and Kaiser).

Did you say Fall of 200_1_? I've heard of publication
delays, but that's incredible.

I think they are having staffing problems at AJP. The editor
said it would be published in early 2000 if I got the revised
manuscript in before Jan 1, 2000. Obviously, he was unable to
keep his promise for some reason.

How about summarizing the paper for us?

I used the fielding model as an example of one approach
to testing for controlled variables. Three possible controlled variables
were considered: rate of change in optical angle, acceleration of
optical angle and linear optical trajectory
(LOT, the McBeath, et al hypothesis). Modeling eliminates
optical acceleration as a possible controlled variable. I
suggested experiments that would allow a decision between
optical rate and LOT.

I had a nice visit with Schaeffer when he was here in LA LA
Land a couple years ago and he said they (the _Science_ baseball
team) were going to do some experiments that would provide
the kind of data I needed. I haven't heard from him in several
months -- he was a graduate student in transition when I saw
him -- but I will try to get in touch with him again soon.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: marken@mindreadings.com
mindreadings.com