|
People
Salvatore Rappoccio
Daniel Sherman Joao Guimaraes da CostaAndy Foland Melissa Franklin Harvard University Email the authors Links
|
|
Question from Taka
Material effects, Are you still using the same alpha value to take into
account the material effects and the mistag positive/negative
asymmetries?
We have re-measured the light flavor asymmetry in CdfNote 7585.
We obtain 1.27 +- 0.13. We have updated the measurement to
simultaneously fit for the scaling of the negative tails of the
heavy flavor distributions, and the positive excess rates of heavy
and light flavor. We obtain a result of 1.8 to scale the negative heavy
flavor component, whereas a value of 1.6 was used in Gen4. If we use
the same Gen4 value of 1.6, our light flavor asymmetry is 1.08, as was
naively estimated. This results in an increase of 2.5 background events in the signal
region.
Question from Un-Ki
The Non-W background 5-jet double tag estimate is larger than the
single tags. As this is not physical, can you estimate the Non-W
double tags better?
As we use both the "tag method" and the "B-rate" methods to calculate
the double-tag estimate, we are sensitive to statistical fluctuations
due to the low statistics in the "tag method". Because of this,
we have changed to use the "B-rate method" only to calculate the
double tags. We still use the weighted average of the "tag method"
and the "B-rate method" for the single tags. See the
table here for the comparison.
Question from Weiming
You should take the systematic error on Fnonw into account
when you take the weighted average of the "tag" and "B-rate"
methods for the Non-W single tag estimate.
We have updated our results to use the systematic in the
weighted average. The results were consistent with the
values we obtained without using the systematic in the
weighted average. See table here.
Question from Tom W.
Have you considered the correlations between the MET and ISO after
the MT cut?
We have examined the Met Vs iso plane before (see here)
and after (see here) the Mt > 20 cut. There is a
sculpting of the distribution at low Met. We have also examined the
Mt distribution of events with Met between 10 and 20 GeV (see here).
Because the Mt seems to be consisting of two components, one above 20 and
one below, we have applied the Mt cut to Regions A and B to pick up the
right normalization of the Non-W events. The effect of this change is shown
here. There is an increase of 1.3 +- 1.1 background events in the signal
region because of this change.
Questions from Weiming/ Un-Ki
The Mt distribution for isolated and non-isolated leptons is not exactly
the same. How can you justify using the non-isolated, high Met sample
as a template for the Non-W backgrounds for the optimization of the Mt cut?
The template in question is shown here.
We have examined the Mt distribution of events with Met between 10 and 20 GeV
(see here). While there will be a slight
difference between isolated and non-isolated leptons, the distributions are
qualitativly similar. We feel the optimization at 20 GeV is still the
correct choice because there is a clear seperation in all the distributions
around 20 GeV. We have also shown the Mt distribution for Region D pretags
(see here). The distribution below 20 GeV is still
persistent. Given all these plots, we feel that the Mt > 20 cut is correctly
removing Non-W events as expected.
Met Vs Iso, No Mt Cuts 
Met Vs Iso, Mt > 20 
Mt For 10 < MET < 10, 1+2 Jet Bin 
Mt Template For Optimization, From Region C pretags 
Mt Distribution for Met > 20, 0.05 < iso < 0.2 
Table Comparing Different Non-W Configurations. 
Table Comparing Tag and Pretag Methods for Non-W Double Tags. 
Posted on April 11 by JG |