<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    Hi,<br>
    <br>
    one thing that can be done is to drag along events gated on RF side
    peaks. Then you can see whether significant amounts of accidentals
    are left after full analysis.<br>
    <br>
    Su.<br>
    <br>
    <div class="moz-cite-prefix">On 20.02.23 11:33, Richard Jones via
      Halld-tagger wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CABfxa3QpAYofmie_UmXv23dg6Gny08Spr9ebpdKxdbe5OJrq9Q@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <div class="gmail_quote">
          <div dir="ltr">Hello Peter,
            <div><br>
            </div>
            <div>Thanks for asking, I noticed this as well, but I
              thought it was only being used for PWA where the primary
              focus is on the angular distributions. I agree that it is
              a concern for differential cross sections. You are right
              that this is not properly taking into account the
              accidentals that are present. Essentially it amounts to a
              hybrid between a fully tagged and a fully untagged
              experiment. Here are those two extremes:</div>
            <div>
              <ol>
                <li>A(untagged) -- the photon energy is inferred from
                  the reconstructed final state, and used to compute all
                  of the beam properties associated with the event: the
                  flux, the polarization, etc.</li>
                <li>B(fully tagged) -- the photon energy is inferred
                  from the unambiguously associated hit in the tagger,
                  which is used as input to the kinematic fit and to
                  lookup beam properties for the event</li>
              </ol>
              <div>At the rates of GlueX phase 2, we do not have the
                luxury of option 2 on an event-by-event basis, but we
                can achieve it by accidentals subtraction. Short of full
                accidentals subtraction there are several short-cuts you
                can use. All of these have uncontrolled systematics.</div>
            </div>
            <div>
              <ol>
                <li>best chi square - put them all in a ring and take
                  the last man standing as the winner with weight 1</li>
                <li>weighted average - count them all above some
                  chi-square acceptance cut and weight each event by 1/n
                  where n is the number of surviving tags</li>
              </ol>
              <div>Both of these methods reduce to tagging strategy
                A(tagged) at low rate, while they reduce to strategy
                B(untagged) at high rate. At GlueX Phase II intensities
                we are some intermediate hybrid of the two with these
                shortcuts, certainly not approximating B(fully tagged).</div>
              <div><br>
              </div>
              <div>To see what these short-cuts entail, consider the
                high-rate limit in the tagger. At high rate, the
                extracted cross section goes to infinity for a realistic
                tagger and an ideal GlueX detector. In reality, the
                asymptote would be something greater than one, channel
                and final state dependent, and probably run period
                dependent as well. The reason for this is that the
                tagger detection efficiency per beam photon goes down at
                high rate, while the accidentals continue to grow and
                generate a valid result for any reconstructed final
                state, tagged or untagged. So the flux that you need to
                put into the denominator under the yield for extracting
                a cross section will be different depending on the final
                state. Using the same flux regardless of final state
                could be a leading cause for why we are seeing different
                cross sections for the charged and neutral decays of
                eta.</div>
              <div><br>
              </div>
              <div>Beyond that, the shape of the flux spectrum (and the
                polarization spectrum for polar observables) is
                different from the shape of the energy dependence of the
                reconstructed yield. Take for example the
                energy-dependent cross section around the coherent edge.
                The tagged flux has a sharp edge, whereas the
                reconstructed yield washes out the edge with a
                resolution that depends on everything in sight: the
                kinematic fit cut, conditions in the detector, the
                particular final state, etc. One way to reduce our
                dependence on the different beam photon energy
                resolutions in the flux and yields is to average over a
                wide range in beam energy. As long as we are not
                interested in the s-dependence of the cross section,
                this might be justified and would reduce the systematics
                from these short-cut approaches.</div>
            </div>
            <div><br>
            </div>
            <div>In PWA, i understand that this avoids the pain of
              negative weights and so improves the statistical error
              from the fits (or at least it gives that feeling). In
              fact, it introduces a set of new systematic errors of its
              own that will probably drive us back to the more rigorous
              approach before we are done. For the moment I am not
              speaking up about this because we just need to get our
              first results out. But eventually this needs to be given a
              critical review. I hope to be part of that at some level,
              as soon as my work on photon beam systematics reaches a
              level where it can be used for publications.</div>
            <div><br>
            </div>
            <div>-Richard Jones</div>
          </div>
          <br>
          <div class="gmail_quote">
            <div dir="ltr" class="gmail_attr">On Sun, Feb 19, 2023 at
              11:13 AM Peter Hurck <<a
                href="mailto:Peter.Hurck@glasgow.ac.uk" target="_blank"
                moz-do-not-send="true" class="moz-txt-link-freetext">Peter.Hurck@glasgow.ac.uk</a>>
              wrote:<br>
            </div>
            <blockquote class="gmail_quote" style="margin:0px 0px 0px
              0.8ex;border-left:1px solid
              rgb(204,204,204);padding-left:1ex">
              <div>
                <div
style="background-color:rgb(255,235,156);width:100%;border-style:none;border-color:rgb(250,235,204);border-width:1pt;padding:10pt;font-size:11pt;line-height:12pt;font-family:Calibri;color:rgb(0,0,0);text-align:left"><span
                    style="color:rgb(156,101,0)"></span>*Message sent
                  from a system outside of UConn.*</div>
                <br>
                <div>
                  <div>Hi Richard,</div>
                  <div><br>
                  </div>
                  <div>During the collaboration meeting a few people
                    presented analyses which used a chi^2 ranking with a
                    +-2ns cut around the RF peak instead of tagger
                    accidental subtraction.</div>
                  <div><br>
                  </div>
                  <div>My initial thought was that this is wrong and not
                    recommended. Did the guidance by the beam line group
                    change regarding this issue? Given that the current
                    a2 cross-section analysis is using this method and
                    there is a big push to publish it asap I am
                    concerned that this might not be resolved properly
                    and might set a bad precedent going forward.</div>
                  <div><br>
                  </div>
                  <div>Given that you are the expert on this topic, what
                    are your thoughts on this issue? Is that a
                    legitimate way to perform analyses?</div>
                  <div><br>
                  </div>
                  <div>Cheers,</div>
                  <div>Peter</div>
                  <div><br>
                  </div>
                  <br>
                  <div>
                    <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                      <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                        <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                          <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                            <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                              <div dir="auto"
style="color:rgb(0,0,0);letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                                <div dir="auto"
style="letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
                                  <div><font color="#7b7b7b">----------------------------------------------------<br>
                                      Dr Peter Hurck (né Pauli)<br>
                                      <br>
                                      My new email address is</font></div>
                                  <div><font color="#7b7b7b"><a
                                        href="mailto:Peter.Hurck@glasgow.ac.uk"
                                        target="_blank"
                                        moz-do-not-send="true"
                                        class="moz-txt-link-freetext">Peter.Hurck@glasgow.ac.uk</a></font></div>
                                  <div><font color="#7b7b7b"><br>
                                      Research Associate<br>
                                      Nuclear and Hadron Physics
                                      Research<br>
                                      School of Physics and Astronomy<br>
                                      University of Glasgow</font></div>
                                </div>
                              </div>
                            </div>
                          </div>
                        </div>
                      </div>
                    </div>
                  </div>
                  <br>
                </div>
              </div>
            </blockquote>
          </div>
        </div>
      </div>
      <br>
      <fieldset class="moz-mime-attachment-header"></fieldset>
      <pre class="moz-quote-pre" wrap="">_______________________________________________
Halld-tagger mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Halld-tagger@jlab.org">Halld-tagger@jlab.org</a>
<a class="moz-txt-link-freetext" href="https://mailman.jlab.org/mailman/listinfo/halld-tagger">https://mailman.jlab.org/mailman/listinfo/halld-tagger</a></pre>
    </blockquote>
    <br>
  </body>
</html>