For APS, all plots need to be/should be approved by February 8th. And because of this constrained timeline for getting results for talks for APS, Jan and I sat down and discussed what can be done/what needs to be done to accomplish this.

Firstly, there is very little chance we will have the final/paper ready results for APS because of two main reasons : the embedded simulation request and finalization/vetting of the analysis by the W-team and by the spin PWG that should occur. WIth this in mind we should try to just get preliminary results because we should be able to use non-embedded simulation for the preliminary results. If any of this does move along more quickly than we anticipate then we can re-visit the possibility of having final results by APS, but the rest of this entry/list is written from the perspective of going for preliminary results.

The four main results we are trying to get released are the vernier scan, the W cross section, the Z yield (possibly cross section), and the W AL (is there a reason to calculate a Z AL?). In order to get these results out there are a number of things that need to happen. Some of them can happen in parallel and some of them are dependent on others being completed before they can be finalized (though much of the work can proceed in parallel)

I'm sure the list is incomplete, so update it as you see fit.

  1. Vernier scan
    1. Need to resolve the fitting issues by xxx that are present right now. Reminder that this is the fact that having 1 Gaussian + 1 Gaussian does not describe the BTH3 vernier scan data well. There may be some physical motivation for going to 2 Gaussians + 2 Gaussians (need the reasoning from Angelika). Ross tried the fit with the 2+2 solution but found that it was unstable. Joe is now helping to explore the stability/convergence of the fit.
    2. The plan is to see if we can understand the instability in the 2+2 and if not we will regress to the 1+1 solution and assign a conservative systematic uncertainty by XXX.
    3. We have two vernier scans for the pp500 data. We hope that these are consistent with eachother within their uncertainties. Will assign a larger uncertainty if not. Should talk to Mike about the proper way to do this with this little data.
    4. Because of the variation of the amount of dead BHT3 trigger throughout various parts of the run, the cross section that will be quoted will be scaled such that it is for 100% working BEMC and then the proper dead area should be accounted for when calculating the luminosity with this cross section throughout the entire run.
  2. Walgo Finalization
    1. This should be decided on by ~January 8th to allow the efficiency studies to be completed in a timely manner
    2. There are a few outstanding questions
      1. How are we going to include the 1 awayside jet veto + 2 awayside jet veto? How will this change the subsequent background subtraction? Justin has some ideas. Joe will think about this and will write up his conclusions in the coming week+.
        Would be good to have final W-algo code in cvs by January 5th*.
      2. Do we plan on including the ETOW on the nearside energy cut? I believe the answer to this is yes, but some more thought should be put into this, especially when considering the "missing endcap" background subtraction. Joe will think about this.
      3. After the algorithm logic and style of cuts is decided on, Justin should implement this in the code and then Jan will run a quick study to try and optimize some of the cuts. Jan expects it will take 3-4 working days. Not a full cut-space search, just a simple tweaking around where the cuts are set now (which should be nearly optimal). Will be done with W-algo logic & cuts for the run 9 data. Just try to increased the S/B [or some other measure... should maybe try to optimize S/(2B/S+1)... good for optimizing when subtracting a background]
  3. Zalgo Finalization
    1. This needs a lot of work right now. But with the expectation that we will really only have a Z yield (and not cross section), the decision of the exact algorithm can be pushed to a later date, ~January 19th. This will be a combination of Ross and Joe.
  4. Walgo QA/Vetting - After Finalization
    1. The two main questions in the community about the result are the charge sign identification/mis-identification and the hadron/electron separation
      1. To support the charge sign identification we will use the Z peak result for the like sign and unlike sign tracks as well as the Q/pT plot being separated for the two charges at high pT.
      2. For the hadron/electron separation we should make a plot using the BSMD for e+,e- identified by W-algo and for hadrons (we don't have to worry about the absolute calibration). The current idea (should think about it more and refine it) is to look at the BSMD response for clusters with ET < 22 (probably hadrons..?)  and for clusters with ET>35 (probably electrons.. ?)
  5. Data QA
    1. Detector QA - This should be done from the great work by Hal et al. Is there any more we would like to do?
    2. Trigger QA - This needs some more work. The quantitiy of N(ET>25 reconstructed using Walgo)/N(BHT3 triggers) should be plotted versus run/fill/spin pattern to look for consistency
    3. Spin Pattern QA - Jan will do this.
  6. Calculation of AL
    1. There are 5 main things that need to be done in order to calculate the AL
      1. Polarization of beams - Because this is a preliminary result we will use the online average polarization value (or fill dependent online values -TBD).
      2. Relative Luminosity vs. bXing - assumes QCD events do not have SSA with magnitude relevant to AL from Ws (of 0.1).
        Currently it is based on the L2W-triggered events passing cuts rejecting beam background. Should be ok for the preliminary result, but perhaps we should use a different trigger for the final result.
      3. Spin sorted yields - The framework is in place. For some runs it fails after DB endTime was fixed - needs 1-2 days of work to investigate and fix.
      4. Estimation of background polarization - In principle there should be a spin dependence to the background so this needs to be thought about more. Joe wlll put some time into thinking about this.
      5. Energy scale - If there is a mis-calibration/lother effect/etc it will lead to an uncertainty because of the lower bound of the cut that we use to calculate the A_L. It should also fold in to the background subtraction. We should investigate using the Z peak and the new calibration because if there is a stark mis-calibration we won't even get a preliminary result.
      6. stability of AL for different data subsets (fill pattern) , W-algo cut changes, QCD model changes
  7. Calculation of W/Z cross section
    1. Lumi per run calculation - Is dependent on the vernier scan result, the same event rejection criteria needs to be applied.
    2. Efficiencies - These all need to be investigated better. It requires official M-C w/o (first) and w/ embedding (later). At this moment we have none.
      1. Vertexing - Should have a conservative error because we won't be using embedded simulation
      2. TPC tracking - Should have a conservative error because we won't be using embedded simulation
      3. Trigger - Has anyone calculated this yet?
      4. Algo reconstruction - Dependent on simulation, but easy enough to calculate.
    3. Unfolding of energy reconstruction - Should be investigated using simulation
    4. Background subtraction - will be finalized after the algorithm is finalized. How do we assign a systematic
    5. Background subtraction due to known and calculable sources (Z's, W->taus)
    6. Statistical uncertainties
    7. Systematic uncertainties (one for everything!)
  8. Data/MC Comparison
    1. For APS - Comparison of the W yield vs. ET (background subtracted) versus pythia expectation
    2. For Final - Need more and something more detector level like the calorimeter response in data and MC to high energy e+/e-
  9. Theory Calculations - Bernd
    1. RHICBOS - for AL/xsec
    2. Code from deFlorian - for AL/xsec
  10. Issues postponed after APS
    1. filtered QCD M-C to be used in embedding to show more realistic estimate of M-C QCD background.
  11. target dates added to calendar
    1. Feb 4 all plots shown to spin PWG for comments
    2. Feb 8 all plots approved to be shown at APS
    3. Feb 9-12 practice talks for spin PWG
  • No labels