defination of entries per event

[Karabo-Mosala]

Question

Hi,
I was reading lhco file in MadAnalysis5. Ma5 report gives me 2 numbers "events kept" (given in table of Cut section) and "Integral" (given in table of Histogram section). "Event kept" refers to remaining events after cuts which are generated by MadGraph. These two numbers are related by "entries per event". How conversion between these quantities is done, can you please help?

Thanks

[jackaraz]

Hi @Karabo-Mosala, I'm not sure if I understand your question. Ma5 counts MC events that are passing the cuts, but MC events have no physical meaning, but each has a weight, $w_i$, attached to it. So after cuts, we calculate the probability of events passing the cuts $p = \sum w^{pass}_i / \sum w_i$. So physical events that are passing each cut are $N = p\sigma\mathcal{L}$ where $\sigma$ is the cross-section, and $\mathcal{L}$ is the luminosity. The integral of the histogram is simply the integral of the curve that you are getting.

[Karabo-Mosala]

Hi,

My question is, looking at the picture attached below - for the histogram, we have entries per event on distribution. How is that value calculated.

[jackaraz]

Hi @Karabo-Mosala entries per event means "sum of event-weights over entries" divided by "sum of event-weights over events".

[Karabo-Mosala]

Hi Jack,

Thanks for the help.

In some processes this ratio "entries per events" is 1 and in some it is some number.
I have some confusion like;

1. In my example case, entries per event = 234527/171330 = 1.37
   Here sum of events-weights over enteries = 234527 is physical number, right?
2. We get this physical number by formula as you explained in first reply. MA5 doesn't show at what luminosity we get this number.
   Thanks for your patience.

[jackaraz]

Hi @Karabo-Mosala

1. No, this number is just for you to estimate how sufficient your MC simulation is. In case your particular event needs more events or not. This is not a physical number. The only physical number is the ratio between the initial and final sum of event weights which gives you the efficiency of the cut or histogram bin.
2. It's a ratio, luminosity cancels out. Columns with events kept and rejected are computed with luminosity.