Hi Antonio,
I'm a bit puzzled that by just randomizing the factors
you get now smaller errors compared to a fit!?
Cheers,
Heiko
On Fri, 15 Sep 2006, Antonio Petrella wrote:
> Hi Heiko,
> what I recomputed are the systematics on datasets with B/D reweighing
> and with different ip cuts (without fitting S/P but randomizing the
> factors...)
>
> ip > 0.20 ~12%
> ip > 0.30 ~14%
> ip > 0.40 ~9%
> ip > 0.50 ~32%
>
> Cheers,
> Anotonio
>
> Heiko Lacker wrote:
> > Does this affect also the other numbers?
> > In particular, the systematics from S/P when only randomizing
> > the numbers bin-by-bin?
> >
> > Could you please provide all plots showing the
> > PartialBRBR distribution from the randomization?
> >
> > Cheers,
> > Heiko
> >
> > On Fri, 15 Sep 2006, Antonio Petrella wrote:
> >
> >> Hi all,
> >>
> >> unfortunately there was a bug in the macro to compute the systematics
> >> due to S/P fit. I found it and re run the computation for the cases:
> >>
> >> 1- linear fit starting from the 2" bin;
> >> 2- gaussian fit.
> >>
> >> In both cases there is a worsening wrt the previous estimate:
> >>
> >> 1: 35.7% (wrong old estimate: 22.5%)
> >> 2: 41.4%. (wrong old estimate: 22%).
> >>
> >> Plots of the PBrBr distribution are attached with this mail.
> >>
> >> Cheers,
> >> Antonio
> >>
>
|