 |
 |
Hello Tim, Thank you for you comments and arguments. Yes I will make presentation about this on Tuesday. Just to comment quickly on the size of the v0 collection. With this cuts the v0 size (and # of events) increase is about 10%. The dsts V0 sizes are "26Mv vs 23 Mb". On the full recon it is too small about 1%. This was the main reason to loosen cuts, where I have filling cuts might be not so safe, and therefore I think it will be worth to have 10% more events in the skim, but feel safer. Rafo On 12/1/18 10:49 PM, Nelson, Timothy Knight wrote: > Hi Rafo, > > Thanks for the email, but something like MOUSE cuts don’t get changed this way: I object to making such changes without reviewed presentations and discussion. These current cuts and their results were shown many, many times with a lengthy validation process by Miriam, and approved with comments solicited after each presentation. > > As far as chisq cuts (what happened to track chisq probability?) I do not understand cuts this loose. The MOUSE cuts were already very loose. If I recall correctly, some tracks let in by looser cuts are obviously garbage when looked at in an event display (e.g. changing the sign of tan(lambda) along their trajectory!). Can you show us that these tracks are useful? > > I think you are defeating the purpose behind the OtherElectron collection with your last change. The whole point of that collection was to have electrons that would not be used in analysis, but would be needed for other purposes kept separately from those we want to make V0 from. Why do you want FEE mixed in with the electrons we will make V0 from? This will simply blow up the V0 collections and data volume again. > > We can discuss this on Tuesday, but the place to start is to review all the work shown in validating the current cuts, and determine through a similar process whether there is really a strong case for changing things. > > Tim > >> On Dec 1, 2018, at 6:32 PM, Rafayel Paremuzyan <[log in to unmask]> wrote: >> >> Dear all, >> >> I modified a little bit MOUSE cuts, in particular loosened them. >> >> <maxVertexP> 2.8 </maxVertexP> >> was 1.2*Eb, thic cut is almost the same as the mosue >> >> >> <minVertexChisqProb> 1.7e-16 </minVertexChisqProb> >> Vertex fit chi2, this corresponds to chi2?NDF to be less than 20, before it was around 6, but this is not increasing the number of events too much, >> since most of events are at low chi2/NDF >> >> >> <maxVertexClusterDt> 2.5 </maxVertexClusterDt> >> time difference between clusters, was 2 >> >> >> <maxTrackChisq5hits> 60. </maxTrackChisq5hits> >> <maxTrackChisq6hits> 72. </maxTrackChisq6hits> >> Track fit Chi2 cuts. Was around, now this corresponds to chi2/NDF = 12 >> >> >> The most important modification is the electron momentum cut in the final state particles >> <maxElectronP> 2.6 </maxElectronP> >> was 0.75 Eb, When we apply all other selection cuts, we don't see any FEE peak in the electron momentum distribution, >> therefore I feel safe to release it effectively no cut. >> The consequence is that "OtherElectron" collection mostly will be empty. >> >> I applied MOUSE cuts to pass1 data except the electron momentum cut, and >> attached is the comparison of the one with MOUSE cuts, and the one from pass1. >> The red histogram represents 1 new cooked file with MOUSE cuts, and is scaled N_recon(pass1)/N_recon(new file), >> and as they agree with statistical uncertainty. >> >> Please let me know, if you have comments on this >> >> Rafo >> <P_em_validate.png>_______________________________________________ >> Hps-analysis mailing list >> [log in to unmask] >> https://mailman.jlab.org/mailman/listinfo/hps-analysis > > ######################################################################## > Use REPLY-ALL to reply to list > > To unsubscribe from the HPS-SOFTWARE list, click the following link: > https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.slac.stanford.edu_cgi-2Dbin_wa-3FSUBED1-3DHPS-2DSOFTWARE-26A-3D1&d=DwIGaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=0HDJrGO9TZQTE97J9Abt2A&m=g8C7unOiqEeF-lvOyfbvwcadx_8Emq-VZn5nGdjPqHQ&s=E-iC-970vIYIPafaEuCbCj6ekTwBrdq4f__iw13mcX4&e= ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the HPS-SOFTWARE list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1