Hi Sheila,
I had a look at your plots and have a few (maybe a lot) comments. Some
were already mentioned yesterday in the meeting.
Plotting:
---------
- when you plot number of leptons (or e, m, ..) you should fix the maximum
of your plot. Either to 1 or test for the maximum of both histograms.
- I like histograms better (with error bars). Maybe you can try something
like
h1->Draw("histo")
h1->Draw("e1same")
h2->Draw("histosame")
h2->Draw("e1same")
Depends a bit if you can still see something, but I really appreciate
the histo option for the spectra
- the binning for the spectra could be finer. Maybe 50 MeV?
- you should say more clearly what you have plotted in the captions
to your plots. Number of particle is Number of leptons per BReco event.
You say it in the text but maybe you can also add it to the plot caption.
- Your normalisation makes only sense, if you look at a quantity, which
you count for all events. Clearly, the spectra are only for events with
at least one lepton (or e, m). Here you should normalise to the number
of entries.
- for the spectra, do you take only the most energetic particle in an event
or all? The most energetic particle should be better simulated than all,
so take only that.
- how do you get the gif plots? Do you know that you can use
canvas.Print("plot.gif")?
Number of Breco vs run number:
------------------------------
- I was thinking of something like Toyoko showed in one of the
Wednesday meetings on page 4:
http://www.slac.stanford.edu/BFROOT/www/Physics/WednesdayMeetings/W072005/orimoto.pdf
There, the mes yield/fb-1 is plotted, but for us the number of
BReco events is still fine. The binning is 200 runs, maybe 100 is
also fine. For data you clearly need the luminosity, which you can
get from the lumiscript with adjusted tcl files.
I am not sure how this is done for MC. I guess there is always the
same number of events in a run and the runnumbers are evenly
distributed. But then you would expect a more or less straight line
for MC. Maybe this needs some further investigation, but I am not sure
if we learn something from this.
R2 with R3 data comparision:
----------------------------
- the muon momentum for R2 is missing.
- the number of leptons for R3 is missing.
- the number of leptons for R2 looks stange, you don't have entries
without a lepton (as for electrons and muons seperately).
- apart from the normalisation the electron and muon spectrum look good
and could be compatible. I am not sure for the lepton spectrum.
- the number of leptons (e, m, l) look not as expected. I would expect a
lower number of muons in R3 compared to R2 because of the decreasing
efficiency. I don't expect such a difference for electrons. This can be
cross checked with the PID control histograms from the PID group.
- the number of Kaons looks good.
R2 data/MC comparison:
----------------------
- the muon spectrum for MC is missing.
- the number of lepton for data is missing.
- for data you only should use OnPeak. You can check if the OffPeak
data is important to the analysis. Interesting numbers are number
of Breco events per fb-1, number of leptons per fb-1 and number of
events with BReco, lepton and lepton momentum > 1 GeV. This number
should be very low. Already the number of BRecos should be quite
low in the OffPeak data. This is one reason why we use the BReco
sample.
- The same argument should hold for ccbar and uds events. If you
still have a discrepancy between data and MC you should also look
at these samples.
- the number of ... plots don't look too promissing. For example the
number of electrons differ by absolute ~5% in the second bin. Maybe
this is coming from the OffPeak. Let's wait for the updated plots.
Numbers:
--------
A few numbers would be helpful:
- number of events with at least one lepton/electron/muon per fb-1
(N/intLum), which is the average of the ... vs run number plots.
Same for BReco.
Cheers,
Wolfgang
|
