From: clidman (clidman@eso.org)
Date: Sat Jul 12 2003 - 16:21:37 PDT
Dear All,
If we are going to include an additional figure, I'd suggest:
- linear z scaling. As mentioned by Greg, a log z scaling emphasises the low-z data.
- Figure 6 (In Saul's original e-mail) was the most striking. The binning appears to
average about a dozen high z SNe, and fortuitously, some of the most deviant SNe
are included in this dozen We should be clear in the caption what we have done. The
residual plot (the lowest plot is Saul's figures) is probably biased, because we
have subjectively chosen a binning scheme that made the Hubble plot look very good.
- Stay in B&W. We can put the colour plot on the SCP web page.
- As Greg mentions, include only the averaged points (ie Saul's clean version).
Cheers, Chris.
Greg Aldering wrote:
> Personally I like the plot without the unaverage points. If someone wants
> the unaveraged points they can refer to Figure 5 (since we seem to have
> decided that this would be shown in addition to Figure 5). Also, IMHO, the
> point of averaging is to clarify the trend in otherwise noisy data -
> putting both the averaged and unaveraged points on the plot completely
> defeats this goal.
>
> I may also have preferred it when the error bars were black (You'd have
> to run ploterr and then oplot so the red circles are in front of the
> error bars though.) In that version the larger error bars were not as
> distracting to my eye. Also, I like the linear z because it shows *our*
> data rather then using 3/4 of the area to show the low-z data (or else
> having to resort to insets as in Figure 5).
>
> Cheers,
>
> Greg
>
> On Sat, 12 Jul 2003, Saul Perlmutter wrote:
>
> > ...If we like the more busy version, we may want to consider also
> > putting the open circle (unaveraged) points on the bottom panel.
> >
This archive was generated by hypermail 2.1.4 : Sat Jul 12 2003 - 15:20:30 PDT