[From Rick Marken (980123.1300)]
David Goldstein (1/23/98) --
Thanks Rick. For those of us who are too lazy to go to he data
source you mentioned, how does it compare?
Remarkably similar. Maybe I'll scan the graph and post it with
the demo (flagrantly ignoring copyright laws); but maybe not. The
model fielder seems to back up more than the real fielders (the
Science article has plots of the ground path of the fielders). I
may be able to minimize the backing up by adjusting the control
system parameters. I'll keep playing with the program.
Also, if these are the two perceptual variables really involved,
would it follow that if we could measure how well a person could
judge these velocities, then we should be able to predict how
well a person could catch a fly ball?
I don't think measuring the ability to judge velocity would tell
you much. What you want to know is what most scouting reports
probably tell you: how fast the guy can run (forward, back
and laterally). The fielder's ability to generate the outputs
that are used to control his perceptions of the ball is probably
your best predictor of his ability to catch fly balls.
I am surprised that a sense of a person's own body in space is not
involved somehow.
I'm sure there are kinesthetic perceptions involved, especially when
it comes to getting the ball into the glove (I haven't modeled that
yet -- probably won't; too hard)
Bruce Nevin (980123.0948 EST) --
The backing away seems undiminished.
Yes. I was kindda surprised.
What third perception might be controlled?
That might be what's going on. As I suggested to David, the backing
up might also be reduced by adjusting the parameters of the control
systems. But there is backing up going on, even on balls hit
laterally with repect to the fielder, in the ground path plots
in the McBeath, et al _Science_ article. The backing up
just doesn't seem to be as pronounced as it sometimes is in the
model but McBeath always had the fielder standing in the same
place -- my fielder moves toward or away from home on
each trial -- and he used only "short" fly balls so the fielder
always had to run foward (ultimately) to get it.
Me:
The plot on the left shows the projection of the path of the ball
on the fielder's retina (relative to the projection of home
plate -- the open square at the bottom of the plot).
Bruce:
Having trouble interpreting this. The ball keeps rising relative
to home plate?
The _projection_ of the ball keeps rising relative to the _projection_
of home plate _on the eye_. This is how I described the right hand
plot it in a private post to Tim Carey:
This is what you would see if you were standing right behind the
fielder's retina and the retina were made out of glass. The
image of the ball is being "painted" onto this glass retina over
time.
Bruce:
(A nit: you mean the plot on the right.)
Thanks. I fixed it. And it was a _shoulder_ mounted camera -- not
head mounted as I said originally.
Me:
Compare these retinal plots to those obtained by McBeath, Shaffer
and Kaiser (reported in Science, 1995, v. 268, pp. 569-573) using
a head mounted video camera.
Bruce:
I don't have access to these. How do they compare?
I really would love to put the McBeath graph up on the net; the
similarity of my model-based plots to their real fielder based
plots is quite impressive, I think. They called their plots
straight lines but most of their lines actually deviate slightly
from linearity, just like mine. The apparent fit of model to real
data is, as my daughter would say, major cool.
Best
Rick
···
--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken