The acceptancies of SM W, H = 90, 110, 130 GeV samples through our selection cuts have been calculated. (Acceptancies). However, these samples do not give an accurate measure of what we expect to see in data, for several reasons.
In MC production we assume 100% branching ratio for tt -> W+,b,H-,b. In reality, there are 4 possible decays. (W+,b,W-,b / H+,b,H-,b / W+,b,H-,b / H+,b,W-,b) Only 2 of these are our signal. So the number of signal events will be reduced by a factor 2 x Br(t -> H+b) x Br(t -> Wb)
In MC we assume 100% branching ratio for H -> cs. The number of signal events will be reduced by a factor Br(H+ -> cs)
In MC we assume 100% branching ratio for W -> l,nu. The number of signal events will be reduced by a factor Br(W-> l,nu)
So, the expected number of signal events can be calculated using:
N = Luminosity x Cross section(ttbar) x Acceptancy x Filter
Where Acceptancy is the acceptancy of a sample though the event selection cuts.
Filter = 2 x Br(t -> H+b) x Br(t -> Wb) x Br(H+ -> cs) x Br(W-> l,nu)
Using tan beta = 1, the branching ratios can be extracted from charged higgs root files. Branching ratios.
A table of expected numbers of events using Luminosity = 10fb-1 and cross section = 830pb-1 (Table).
A similar thing occurs with the standard model ttbar samples.
In MC we force both sides to decay as t -> Wb. So the expected number of events is decreased by a factor Br(t -> Wb) x Br(t -> Wb)
In MC we have a mixture of di-lepton and semi-leptonic events. i.e. one top is forced to decay as t -> l,nu while the other can decay in any way. In real life the lepton can be on either side, so the number of events will be reduced by a factor 2 x Br(W -> l,nu). This assumes that in MC, the same side was always fixed to decay leptonically. Need to check that this is the case, as it will make a factor of 2 difference to the expected number of events.
So the expected number of events is reduced by a filter: Filter = 2 x Br(t ->Wb) x Br(W -> l,nu)
This highlights an important issue. In all cases, the expected numbers from the muon channel are considerably higher. This is because of differences in acceptancy through our event selection.
Large variation in the number of events passing event selection cuts in the electron and muon channels:
Number of events in electron channel: 197
Number of events in muon channel: 332
(Of 3750 events)
Event selection requires:
Exactly 1 tight lepton with Pt > 20 GeV and eta < 1 (No constraints on the number of loose electrons)
MET > 20GeV
At least 4 jets with Pt > 20 GeV, eta < 2.5
2 jets btagged
Add an additional selection requiring the tight lepton from event selection to be matched to a true lepton from W decay:
Number of matched events in electron channel: 181 (92% of above)
Number of matched events in muon channel: 268 (81% of above)
Suggests that our detector is more efficient at finding muons than electrons, since this requirement removes background muons, eg from K, pi decays. But this matching selection still removes more muons than electrons. (Twice as many.)
Discrepacy may be due to different isolation critera - looser isolation cut on tight muons would result in more tight muons than electrons. Look at the delta R distribution between lepton and the closest jet. Find more muons at low deltaR, electrons peak at a higher delta R. Plots But this does not require matching of lepton to true lepton from W decay
If we require only events where the tight lepton can be matched to the true lepton from W decay, there are fewer muons at low delta R. i.e. requiring lepton from W decay removes many of the non-isolated muons, although a significant fraction still remain. Plots We expect that requiring matching should remove e.g. muons from semi-leptonic b decay
But this is counter-intuitive. Information on the topview wiki page suggests that the same isolation requirements have been applied to both, but still some muons from W decay seem not to be well isolated.