Hi,
this is a preliminary study on the effect of a different resultion on
data in the fit reults. I tried two effects:
- an extra gaussian smearing with different sigmas
- a shift in the Mx reconstruction respect to the MC one
These are the results:
- smearing
the x axis is in GeV
http://www.slac.stanford.edu/~daniele/myplots/smear.eps
is the result of the fit (the true value is 0.0166, the line in
the plot) as a function of the sigma of the smearing
http://www.slac.stanford.edu/~daniele/myplots/relsmear.eps
is the difference from the generated value divided by the true value
(result-0.0166)/0.0166.
- shift
http://www.slac.stanford.edu/~daniele/myplots/shift.eps
http://www.slac.stanford.edu/~daniele/myplots/relshift.eps
both as above
We can notice that we can live with an extra smearing in the resolution
if it is below 60 MeV (the shift is below 3%).
The shift introduces a very large bias. We should take it really under
control since a shift of only 10 MeV introduces a bias close to 10%.
I will try to redo these plots with a different set of cut, improving for
instance, the ratio S/B.
Daniele
|