The
LHC Olympics --- Calibration Samples
Updated 11/9/05 to correct and clarify number of events in the samples
and the effect of the trigger.
There are two calibration samples
provided:
- 190,000 events with pair
production of top quarks and antiquarks.
- 40,000 events with diboson pair
production: W W pairs, W Z pairs, and
Z Z pairs, in the proportions roughly expected in the standard model.
Note that these are by no means perfect simulations of
standard
model
backgrounds, which have large next-to-leading-order corrections and
competing processes that are not included here. However, they
should serve participants quite well for calibrating
their
understanding of the PGS detector, for testing any software
that
they might write or wish to use, or for trying out any new analysis techniques they might
develop.
- Here
are some
inclusive signature plots
for the top pair
and diboson
data. General information on
inclusive signatures is HERE.
- The
raw data files can be accessed HERE;
you will be asked for your email address, simply to ensure that we can
inform you by email if there are amy important announcements or changes
to the data sets.
- The
ttbar data set is broken up into fourteen files, named ttbar_1005.dat.#.gz where # runs
from 0 to 13. These can be concatenated by the user into a single
file. General
information on how to read
and use the data files is HERE.
Note that 500,000 events were
simulated, but less than 200,000 were accepted into the sample because
of the effect of the trigger.
- The
diboson data set is broken up into two files, named diboson_1005.dat.#.gz where # runs
from 0 to 1. These can be concatenated by the user into a single
file. General
information on how to read
and use the data files is HERE.
Note that 200,000 events were
simulated, but only about 40,000 were accepted into the sample because
of the effect of the trigger.
- The
samples were created using a standard event generator, Pythia 6.324 for
showering/hadronization, and PGS as a detector simulation. See the general
blackbox page for additional information.
There's also quite a lot to be learned about how far data differs from
naive theory at hadronic colliders! For instance, tree-level
theory would predict the cross-section, the number of electrons, muons
and taus, the number of b quarks, and the number of jets. But
as you can see in the inclusive signatures information pages, the truth
in the data is far from this prediction --- for reasons of triggering,
detector details, b-tagging, and QCD radiation.
Moreover, we can't
consistently include important loop corrections, or account for
uncertainties in parton distribution functions, etc., without more
sophisticated methods, all of which are more subtle than we are ready
for (and not all of which currently exist) so in addition
to being complicated by all the detector issues, this data is simply
not in agreement with the standard model to better than a factor of 2,
or worse. Apologies; but this is simply the nature of hadronic
physics when simulated at tree-level.