Oct 17/24 PionLT/KaonLT Analysis Meeting Notes
                 ----------------------------------------------
                                (Notes by GH)

                     Today: PionLT will be discussed first

Present
-------
Regina - Garth Huber, Alicia Postuma, Muhammad Junaid, Nathan Heinrich,
   Vijay Kumar, Ali Usman, Nacer Hamdi, Argha Das
JLab - Dave Gaskell
CSULA - Konrad Aniol
CUA - Tanja Horn
Virginia - Richard Trotta
Ohio - Julie Roche

Nathan
------
- working on random subtractions on PionLT Coin Lumi data, no report yet

Junaid
------
PionLT pi+ PID study, 5.9 GeV setting
- applied Aero>3.0pe and 1.2<RF<3.4 cuts
  - Dave: prefer to use fewest #cuts possible, as each one needs an efficiency
    to be evaluated, and possibly larger systematic uncertainties
  - Garth: but if extra cut means that the applied cuts can be looser, this
    results in smaller cut dependence, and lower systematics.  So it depends on
    exactly what cuts are needed for the data in question

  - Dave: is the path length correction implemented in this RF time?
    - no, still looking into that
    - *NB* Stephen had done some initial implementation, best to consult him

- shows pi+n MM peak under different cuts, estimated #counts in peak with
  background
  - difference in #counts between Acceptance only, and Acceptance+CoinTime+Aero
    is very large, 21k vs 14k
  - *NB* need to understand which cut is causing such a large drop in #counts,
    seems unlikely to be random subtraction since the counts are all in pi+n MM
    peak
  - Ali: *NB* suggests to try a wide CoinTime cut in the Acceptance only plot,
    wide enough to not remove p, K+ but still exclude randoms

- will repeat stdy for 7.9 and 8.5 GeV settings from 2022 data

HMS Matrix Elements
- implementing new HMS ME for 6.6 GeV/c setting

Richard
-------
KaonLT trying different sigTT parameterizations for Q2=3.0, W=3.14 data set
- the different paramterizations tried:
  - Marco's thesis sigTT form
  - Marco's sigTT with sin^2(theta) and sin^3(theta) added
  - Vijay's thesis sigTT form, with pole term removed
  - Richard's earlier function:
    (-p13*|t|+p14)*|t|**(Q^2/p15)-p16*Q^2*sin^2(theta)
- all use 1/(W^2-mp^2)**n with n=2.25

- Data/MC ratio plots vs phi for all parameterizations, all after 5 iterations
- Marco's thesis form gives slightly better fit than the others
  - Marco with sin^2(theta) is similar
  - all 5 have similar oscillations in ratio vs phi, but a bit more damped with
    Marco equation
- will do more iterations w/ Marco's sigTT and look at other settings

- Garth: the calculated chi-square for your plots seems too high in
  comparison with the fluctuations in the plots
  - *NB* need to recheck chi-square calculation, suspicion is that statistical
    errors are not taken into account correctly

Vijay
-----
Replaying Q^2=0.42 PionLT data
- running into some problems adapting replay scripts for Alma9
  - getting a STE_LAUNCH_FAIL error, indicating incorrect batch queue
    parameters
  - Dave: are you using hcswif script which was hardwired to CentOS7?  If so,
    you will need to modify it

Alicia
------
BSA revised manuscript has been submitted to PRL, waiting w/ editor

Ali
---
Back to thesis writing after returning from DNP meeting in Boston
- next analysis step will be to evaluate systematic uncertainties on piDelta
  BSA

Nacer
-----
KaonLT Heep-COIN analysis
- using new SHMS-3/4 effs after Mark's code improvements (more restrictive
  selection criteria)
  - effs now >99.5%, vs 92-97% before
- shows plots of Data vs SIMC under various conditions
  - no cuts
  - CoinTime, HCer, HCal, MMp(exclude pi0)<0.1 GeV
  - after cuts, EM, PM, W data distributions are much closer to SIMC prediction
  - Jule: there is a deviation "small extra bump" of data around PM~0.1 GeV/c
    not seen in the simulation
    
- Dave: Mark Jones recently reminded that in his VCS analysis, there was a
    low W facture anomaly that he resolved in a modified tracking algorithm
    - changed how clusters are defined, more candiate tracks per event, better
      selection of "best track" in event
    - There are four main changes. Each can be turned on/off separately by
      parameters. In the end, the VCS analysis used them all.
    - *NB* see Mark's presentation posted on RedMine: MJones-HCANA_DC-update.pdf
    - Mark added output to the root tree that makes it easier to understand the
      clustering and tracking which is not documented in the slideshow
      (though hinted at it at the end)
    - *NB* GH will see if Mark can give a summary of it at our next meeting

- Data/MC ratios w/o proton absorption:
  - 0.992-1.011 +/- 0.007, almost no beam energy dependence
  - Garth: Nacer had looked earlier at the SIMC elastic model dependence, which
    is 2-5%, depending on Q^2

KaonLT Heep-SINGLES analysis
- HMS electrons: 3.8 GeV, P_HMS=2.026 GeV/c
  - Dave: PMX,Y,Z,PM,EM not defined in hcana for singles events
  - W shows an enhancement at the Delta mass in data, apply 0.9<W<1.05 cut to
    eliminate
  - agreement between Data and MC for HMS Focal Plane variables very good
  - Data/MC ratio 1.037+/-0.001

- SHMS electrons: run 6629
  - W data plot is empty
  - *NB* this works for Vijay, so check with him on what's wrong

Next Meeting
-------------
- Thur Oct 24 @ 16:30 Eastern/14:30 Regina
  - KaonLT will go first