[Hadstruct] [EXTERNAL] Re: First GPD fits

[Dutrieux, Herve]

Hi Joe,

I have been trying to extract matrix elements from the correlation functions with a different method, and I have obtained quite different preliminary results. See attached a document with a description of the method and the result on various momenta for z = 0. I will be at WM this afternoon, or we can have a zoom chat if you want to discuss some elements in there. I have tried to also consider a bit effects of excited state contamination.

Cheers,
Herve




On Jul 13, 2023, at 11:05 AM, Joseph Karpie via Hadstruct <hadstruct at jlab.org> wrote:

Hey everyone,

As promised in the morning, yesterday I had done the first fit of the xi=0 GPD, and as I have told some, I was quite dissatisfied with the results. The chi^2/dof was 220 when I concluded O(alpha) correction.

I think the issue can be best seen in the inaccuracy of the form factor (z=0 for us). See attached plot of F_1(t)/F_1(0). First the range in t is pretty small we can't fit very meaningful t dependence. Second the point at the lowest |t| is at least 1 sigma BELOW some points at slightly larger |t|. This undesirable feature is reflected in all the GPD data with z!=0 which lead to terrible fit results where it's supposed to be monotonic. You can see how the GPD fits failed in the App C of https://urldefense.proofpoint.com/v2/url?u=https-3A__www.overleaf.com_7932724842pjgpxkdxnzsf&d=DwIFAg&c=CJqEzB1piLOyyvZjb8YUQw&r=phC8jtUHUVIl1OXaA25HKnc5L16EvLwhrqx4rXYGPWA&m=s_6frbucAbPNpyud7LY_-SumKWTgN2yFreR0H8DNLpyLyvbrrhFkRSVhOA1pLzFT&s=betf3IQa9MTAH1jPTcebts8k39yDxzaq7P610iFR0EM&e= 

There are some things I will do to move on, but please send any suggestions on these fits above or the next directions below.

First, Colin has given me xi=1 amplitudes which I will begin to look at next. In my notes, I briefly describe how I believe the A_5 amplitude maps to the D term explicitly. Interesting, if true....

Next, Colin has some xi=0 data with higher pz which I could add to this xi=0 analysis to give better range in nu. I will attempt to get the amplitudes from him. In his famous plot of our data spread in xi and t, there is not a great t coverage at xi=0, so there may not be much left to do prior to lattice without generating new data to clear up the situation.

Best,
Joe

PS: Overleaf is failing when I have too many plots (not quite 100). If I can't look at several dozen plots of data and analysis, is it even science? If you wish to see the appendices where I have each (pf,pi) plotted individually (or not see or to fix the too many plots error) then you must (un)comment the \iffalse and \if pair which surround the begin{figure} calls.

--
Joe Karpie
Pronouns: He/His/Him
Postdoctoral Fellow
Theoretical and Computational Physics Center
Jefferson Lab

<form_factor.pdf>_______________________________________________
Hadstruct mailing list
Hadstruct at jlab.org<mailto:Hadstruct at jlab.org>
https://mailman.jlab.org/mailman/listinfo/hadstruct

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/hadstruct/attachments/20230714/a2e12ce2/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Brouillon_VI___GPD_matrix_element_fits.pdf
Type: application/pdf
Size: 680072 bytes
Desc: Brouillon_VI___GPD_matrix_element_fits.pdf
URL: <https://mailman.jlab.org/pipermail/hadstruct/attachments/20230714/a2e12ce2/attachment-0001.pdf>