Re: Fixed vs. floating offsets

From: Lifan Wang (lifan@panisse.lbl.gov)
Date: Wed Feb 19 2003 - 15:43:12 PST

  • Next message: Robert A. Knop Jr.: "Re: Fixed vs. floating offsets"

    Rob,

       I totally agree that it is impossible to make this perfect. But
    you did not answer my question of if the fixed slope fit of the 92ag has
    converged. So I am attaching my arguement below that it did not converge.

    > > For 92ag, I am surprised and bothered by the fact that your best fit
    > > needs a baseline offset of only 0.03. It seems to me that your other
    > > fit (with fixed baseline) actully did not converge to the real minimum
    > > chi^2. One way to see this is to add 0.03 to your best fit (with floating
    > > zero baseline) which then gives a fit with the baseline fixed to zero.
    > > Visually, I would expect this to be a much better fit than the fit given
    > > in the figure with fixed zero.
    >
    > I really think what's going on is that those low points are driving the
    > curve far more than they ought to. There's a whole bunch of them, with
    > really tiny error bars, down there.
    >
    > In the end, I don't think that any of this haggling is going to affect
    > the results much. Look at the figure that compares the P99 confidence
    > intervals to my refits of the same data. My refits include the new
    > lightcurve fits as well as new K-corrections. Any sort of tuneups we do
    > on this sort of thing is going to make *less* difference than that, but
    > will take a lot of time.
    >
    > We have to choose: do we want to try to converge toward perfection, or
    > do we want to publish the paper? Because those are two inconsistent
    > goals.
    >

    From your figure the chi2 for 92ag are

            | fixed base | floating base
    chi2 | 1.35e+03 | 101.
    dof | 29. | 31.

    Now, shift the best fit of the floating case by 0.034 to restore the
    case with the base fixed to zero. What is the chi2 of such a fit ?

    Let Xi be the ith model data points, Mi the ith model points of
    the one with floating base

          chi2 = sum((Xi-Mi)^2/error_i^2)

    Now remove the 0.03 offset, the chi^2 is

          chi^2 = sum((Xi-Mi-0.03)^2/error_i^2)
                = sum((Xi-Mi)^2/error_i^2) + sum(0.03^2/error_i^2) +
                    2.*sum((Xi-Mi)*0.03/error_i^2)

    The last term should be zero for a correct fit ==>

         chi^2 = sum((Xi-Mi)^2/error_i^2) + sum(0.03^2/error_i^2)
               ~ 101. + sum(0.03^2/error_i^2)
                
    This chi^2 is apparently much less than 1.35e+03. So I do not think
    the fit with fixed baseline has converged.



    This archive was generated by hypermail 2.1.4 : Wed Feb 19 2003 - 15:43:26 PST