Higgsino WH request update


Abstract. Estimate of the updated stats for the production of the Higgsino \(\to{}WH\) samples. The decay \(H\to\tau\tau\) will be disabled.


For now I am just looking at the change in efficiency when skipping the \(H\to\tau\tau\) events with the current samples. We are mostly interested in having a decent number of raw events, so that we don’t run out of statistics at final selection. There are two effects:

  1. change in yield before any selection

  2. change in yield at final selection

The change in yield before selection is “how much we save” when we generate events without the \(H\tau\tau\) decay. The change in yield at final selection is “how much we loose” because some of the \(H\to\tau\tau\) events are currently in our current final selection.

The final selection is the same one that I defined back in March. It was defined trying to optimize \(Z_n\). However, because of the limited stats, there were significant fluctuations on the Z_n values from one signal sample to the next. So the choice of the selection criteria was also relying on the efficiency separation between signal and background. That study was made with n0115 ntuples normalized to 13/fb. This update is done witn n0139 ntuples normalized to 21/fb. I have not updated the \(Z_n\) plots because I need to perform some cross-checks with Josephine and Matt on the background yields.

Here is the link to the previous estimate https://test-gerbaudo.web.cern.ch/test-gerbaudo/tmp/2013-03-20/yields/yields_mc_req_2013-03-19.html and tot he savannah request http://savannah.cern.ch/task/?40831

Signal regions

2l, \(\geq\) 2j regions sign flavor
SR6 Opp. Same
SR7 Opp. Diff
SR8 Same Same
SR9 Same Diff

Selection criteria

For SR6 and for SR7 the selection requirements are, addition to sign/flavor:

\(\Delta R (l_1, l_2)\) < 2.0
\(p_{T,ll}\) > 50
\(MET_{rel}\) > 50
\(m_{ll}\) < 80
\(abs(m_{\tau\tau} - m_{Z})\) > 40
\(m_{T}(ll, MET)\) > 50
\(m_{T, min}(l, MET)\) > 50

for SR8 and for SR9 the only requirement in addition to sign/flavor is \(MET_{rel}\) > 50. The plots below show thesignal yield, normalized to 13/fb, and the raw counts.



The signal efficiency for the \(H\to\tau\tau\) veto before selection, figure \ref{fig:fig1}, is \(50\%\) for all signal samples.

The signal inefficiency for the \(H\to\tau\tau\) veto varies from \(0\%\) to \(10\%\) in SR6 and SR7 (figures \ref{fig:eff-sr6}, \ref{fig:eff-sr7}), and \(0\%\) to \(20\%\) in SR8 and SR9 (figures \ref{fig:eff-sr8}, \ref{fig:eff-sr9}).