Hi all,

 as shown in my previous postings, we had two problems in PID:

 1) the muon PID killing at Beta level seemed to be wrong. I observed
large disagreements in p* distribution between data and MC. Once I
re-applied the PID killing at recoilNtp level the situation was
really improved.

 2) the data showed large differences with respect to the old production
in p* spectrum (both electrons and muons). More events at low p* were
present. The same effect was present in Mx distributions (excess in data
at high Mx). MC was pretty unchanged.


 Then I found an explanation to both the problems:


 1) The problem in muons for MC is only in the vtight selector. Look at

 http://www.slac.stanford.edu/~daniele/testPID/m1000.eps

 this is the comparison between muon tight and verytight selectors for MC.
As you see the vtight (histogram) has a lot of events at low momenta and
this does not make sense. The vtight selector was the only selector that
showed very strange behaviours.
Probably this is caused by a wrong killing at Beta level and it should be
investigated. BTW we used to apply tight selection for muons (no reason
to use the vtight one). Then I think we can switch to tight one.


 2) The differences in data-MC comparison are due to a wrong
mes sideband subtraction. In the new data the purity is lower then before.
The histogram based comparison (overall sideband subtraction) is not
adequate. This is shown in the following table of comparison data-MC

 In

 http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/datagene0930_all/datagene0930_all-comp.html

 you can find the comparison based the overall comparison. Disagreement in
pcms and mxhadfit plot is evident


 In

 http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/datagene0930_all_tree/datagene0930_all_tree-comp.html

 you see the same comparison on the same samples with the correct
sideband subtraction bin by bin. All the effects are gone (except the
mm2... sigh). Look in particular at the mxhadfit plot (before and after
the lepton cut). The agreement is pretty good.


  Daniele

 p.s. The last QA plots (70fb-1 with muon tight selector) are in the web
page usual page

 http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/QAcomparisons.html