From: Tony Spadafora (ALSpadafora@lbl.gov)
Date: Thu Jul 10 2003 - 12:32:31 PDT
Rob,
This binning strikes me as kind of strange and will need to be clearly
explained in the caption. It's an OK way of combining points (weighted
average over fixed binwidth but variable location) but I'm not sure I'd
call it "binning". Do regular (fixed location) bins produce an ugly plot?
-Tony
"Robert A. Knop Jr." wrote:
>
> On Thu, Jul 10, 2003 at 11:15:31AM -0700, Greg Aldering wrote:
> > Also, I think there is merit in showing the full redshift range, but
> > then the cosmologies are not well separated on the scale of the plot.
> > Should we consider plotting binned residuals instead? Riess plots
> > the residuals wrt an OM=0, OL=0 universe, for example.
>
> I'll slam together a residual version and add it to the page.
>
> One option for the full redshift range is to play the same game that I
> played with Figure 5-- show the full redshift range, and then an inset
> with the 0.05 binning.
>
> (Note that the binning was done slightly strangely-- I didn't just
> bin at regular 0.05 intervals, but rather found a supernova, binned
> whose z was no more than 0.05 above that supernova's z (with
> variance-weighted combination), and then found the next highest z
> supernova. Repeat. That's a slightly odd way to do it, but does reduce
> the incidence of "orphans" in the case of smaller bin steps. In any
> event, the plotted points are still meaningful-- although as Don would
> remind us, they're only good for plotting purposes, and not what you'd
> want to do an actual fit to.)
>
> -Rob
>
> --
> --Prof. Robert Knop
> Department of Physics & Astronomy, Vanderbilt University
> robert.a.knop@vanderbilt.edu
-- Tony Spadafora ALSpadafora@lbl.gov Physics Division Tel: (510) 495-2316 Lawrence Berkeley National Lab FAX: (510) 486-6738 1 Cyclotron Road BLDG 50R5032 Berkeley, CA 94720-8160
This archive was generated by hypermail 2.1.4 : Thu Jul 10 2003 - 12:32:37 PDT