June 26/25 PionLT/KaonLT Analysis Meeting Notes ----------------------------------------------- (Notes by GH) Today: KaonLT will be discussed first Please remember to post your slides at: https://redmine.jlab.org/projects/kltexp/wiki/Kaon_LT_Meetings Present ------- Regina - Garth Huber, Ivan Zhenchuk, Nacer Hamdi FIU - Pete Markowitz Ohio - Julie Roche York - Stephen Kay Virginia - Richard Trotta CUA - Tanja Horn Richard ------- High Q2 KaonLT LT-separations - found a bug affecting the pion subtraction undertainty - uncertainty propagated correctly now - pion subtraction also improved a bit, some bins were over-subtracted, which explained some bad phi bins in ratio plots - made a new method of doing Rosenbluth fits - was hitting fit bounds before - earlier fits imposed a boundary condition sigL,sigT>0 - now derived a better constrained fit using Cauchy-Schwarz inequality - baked the positivity constraint into the fit itself - model is now pushed out of unphysical local minima - uses Drechsel-Tiator J.Phys.G paper available at https://maid.kph.uni-mainz.de/JPhysG18/JPhysG18.pdf - uses rhoLT=|sigLT|/sqrt(sigT*sigL)<1.0 rhoTT=|sigTT|/sigT<1.0 - copy of Richard's derivation of relation at https://redmine.jlab.org/attachments/2878 - now using much simpler functional forms for Q2=4.4 - sigL=(p1*ft)*exp(-p2*|t|) sigT=p5/(|t|^p6)*exp(-p7*|t|) sigLT,sigTT also simpler, have no sin(theta*) terms yet - did 20 iterations with new forms - ratios much closer to 1 now, some evidence of oscillations so might restore the sin(theta*) in LT,TT - Q2,W,t,MM plots - MC:data disagreement worse for left SHMS than center at low epsilon similar is true at high epsilon - initial results: sigL~20%*sigT sigLT~sigTT - investigating other settings - need to adjust pion subtraction - Q2=3.0,W=3.14 left SHMS SIMC at both low and high epsilon has a shift in MM peak compared to data - still has to do HGC PID pt-pt systematics, currently estimates ~3.6% unc - mentions some difficulty running SIMC for very large #events - Garth: you can modify SIMC to use a random seed that you specify rather than the default one, in which case you can break a big SIMC job into smaller parts, provided eacn one uses a different random seed - Nacer asks for a copy of Richard's sig_unsep files, would like to check if he gets similar Rosenbluth fit to Richard - Richard's data at /lustre24/expphy/volatile/hallc/c-kaonlt/trottar/OUTPUT/Analysis/KaonLT/cache_transfer/kaon/Q4p4W2p74/Trial_20/ - this is where all runs are for Q2=4.4 for the plots. Run 2025June26_H07M55S23 (i=20) are where the plots are from. In 2025June26_H07M55S23/xsects are the separated and unseparated cross section values - 2025June26_H07M55S23/parameters/par.pl_Q44W274.dat is my parameterization and 2025June26_H07M55S23/functions/Q4p4W2p74.model is my model Nacer ----- Q2=0.5 KaonLT LT-separations - setting up framework - for Rosenbluth fit, using python curve_fit https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html - compared to GH's L/T/LT/TT values obtained with PHYSICA, gets same central values, hasn't compared the errors yet - sigL is negative for most t-bins, but no iterations have been done yet, this was with default Koltenuk model and ratios R~2 - working on parameter iteration script Ivan ---- KaonLT BSA analysis - added dummy subtraction - updated t-binning based on sum of Left, Center, Right SHMS settings, not just Center - now has ~60k events/bin - simplified fit clearly makes more sense than full fit, which overfits the fluctuations in the data - extracted mean kinematics per t-bin - *NB* Garth suggests to cross-check both the mean kinematics and the yields per t-bin with Nacer's, in the region where they overlap - also looked at higher Q2 K+ BSA - Q2=2.1 and Q2=3.0,W=2.32 have good statistics, BSA looks fairly promising - Q2=4.4,5.5 statistics too poor to have much of a result - hasn't looked yet at Q2=3.0,W=3.14 setting Next Meeting ------------- - Thur July 3 @ 15:30 Eastern/13:30 Regina - PionLT will go first