STAR Correlations and Fluctuations PWG

[傅川 <fuchuan AT impcas.ac.cn>]

Hi Hanna,

               Thank you for your comments.

Please find the response to your comments below:

https://drupal.star.bnl.gov/STAR/system/files/2-Comments-and-replies-for-PWG.pdf

If possible, please kindly let us know the time plot of PWGC preview.   

Regards,

Chuan for the PAs

-----原始邮件-----
发件人:"Hanna Paulina Zbroszczyk" <hanna.zbroszczyk AT pw.edu.pl>
发送时间:2023-07-13 01:55:59 (星期四)
收件人: "傅川" <fuchuan AT impcas.ac.cn>
抄送: "star-cf-l AT lists.bnl.gov" <star-cf-l AT lists.bnl.gov>
主题: Re: [Star-cf-l] CF PWG meeting, June 1st, 2023, 9:00 (BNL time)

Dear Chuan and PAs,

Thank you for including my comments. I am glad that your work is moving forward.

I have gone through your slides and analysis notes. I might have more comments when I reread them, but now I have the following comments:

Concerning details of your analysis: 

1. p-p

Systematics regarding single particle: I see you include the variation of DCA, NhitsFit, n\sigma, and the momentum smearing, but you check only the tighter option. What about looser ones? Please include them as well. If you release your cuts, you gain statistics, and Barlov’s test passes then. Please indicate where the systematic uncertainty you put is 0 due to limited statistics. 

Systematics regarding pairs: \delta \phi*  should be here. What about \delta \eta?. Please demonstrate why you skip it in your estimations. I don’t see any discussion regarding splitting and merging effects and how these effects affect your correlation function. You should use standard cuts introduced and tested in the past as working. For splitting, you should check Splitting Level (SL) (changing NhitsFit is not the correct way as you have to take care of two tracks that are very close to each other, NhitsFit works on a single track). For merging, there are 2 other approaches to check: AvSep or the fraction of merged hits (FMH). Please demonstrate to them that you are free from splitting and merging. 

The easiest way is to apply different SL and FMH/AvSep and to see how the correlation function changes.

For merging, \delta \phi* - \delta \ eta can work, but to check this try to include/exclude this condition and check what happens to your correlation function. 

TOF and TPC+TOF have different purities. 

How does the purity of p-p look as a function of k* for both cases: TPC only and TPC+TOF?

They should be different, so do you calculate two separate correlation functions for these two cases? Having different purities, they need to be corrected differently. Having two corrected correlation functions, you cannot add them as you can add only numerators and denominators separately. 

And, what is the uncertainty of the purity correction in your case?

For systematics, you should separate what accounts for the systematic error bar of each bin of the correlation function (as you just started this) and how this variation affects extracted parameters (r, f0, d0). 

The fitting range of the correlation function should also be checked to check the stability of your outcome. 

We need more discussion regarding the agreement between LL and CRAB. Please demonstrate that these two approaches give consistent results for the source sizes you extract, between 2 and 3 fm, for all considered systems (p-p, p-\La). Comparing data and transport model, the same approaches should be applied. If you extract some parameters for data, you should use the same for the model you use to build your correlation function. 

Please include your \Chi2 maps where you prove that all extracted parameters: r, f0, and d0 come from the minimum. Then, you can check the consistency with the CRAB for the best parameters.

Then, show the LL model and CRAB comparison with LL theoretical parameters. 

Do I understand correctly that you use CRAB to obtain the correlation function from the transport model and that you fit the LL model? Why? 

Please show the radii dependence as a function of pair transverse mass for all systems: p-p, p-\La, and p-\Xi. 

For the contours of f0-d0, please include theoretical value as a comparison. 

2. p-\La

Do you ensure that the proton already included in your analysis is not re-considered as the product of \La decay?

Can you please demonstrate there is no splitting and merging here? Each track (primordial or the product of the decay) can be split, and we don’t know how this would affect the correlation function. Similarly, there can be a merging between the primary proton and proton decayed from \La? These effects should be carefully checked as well. You could check how close these tracks are if you plot \delta \phi* - \delta \eta between primordial proton and phase from \La. 

What are the uncertainties of the purity correction?

What are the uncertainties of the resolution smearing?

In the experiment, you construct \La from decay products. Do you treat protons and pions individually in case of the smearing of the momentum of \La?

Why, for p-\La case, the feed-down fraction is ~30% and for p-p only 2%?

And, as you apply the Barlov test, which contribution do you eliminate due to limited statistics?

For systematics, you should separate what accounts for the systematic error bar of each bin of the correlation function (as you just started this) and how this variation affects extracted parameters (r, f0, d0). The fitting range of the correlation function should also be checked to check the stability of your outcome. 

Regarding the comparison of data and model: again, as for the p-p case, the same approach should be used. Please demonstrate the agreement for LL and CRAB in the case of p-\La (for given source sizes ranges).

3. p-\Xî:

I see systematic included only from single tracks. Two-track effects are not checked / included (splitting and merging, here, the same comments are applied as in the case of p-\La.

Please similarly introduce residual correlations correction for all three systems: p-p, p-\La, and p-\Xi. 

What are systematic uncertainties of purity correction, resolution smearing, and residual correlations? 

Different ranges of fitting should be included as systematic uncertainties as well. 

Thanks,

Hanna

Hanna Paulina Zbroszczyk

PhD DSc Eng, Professor WUT

E-mail: hanna.zbroszczyk AT pw.edu.pl

Tel: +48 22 234 5851 (office)

Address:

Warsaw University of Technology

Faculty of Physics

Nuclear Physics Division

Koszykowa 75

Office: 117b (via 115)

00-662 Warsaw, Poland

Wiadomość napisana przez 傅川 <fuchuan AT impcas.ac.cn> w dniu 10.07.2023, o godz. 06:02:

Dear Hanna,

                   Thank you for your suggestions and comments.

We carefully considered your suggestons and revised the slides for PWGC preview.

Please find the reply (from page 1) below:

https://drupal.star.bnl.gov/STAR/system/files/Comments-and-replies-for-PWGC-slides.pdf

If possible, could you please add our PWGC preview to the schedule?

Thanks,

Chuan for the PAs.

-----原始邮件-----
发件人:ChuanFu <fuchuan AT mails.ccnu.edu.cn>
发送时间:2023-07-10 11:45:16 (星期一)
收件人: fuchuan <fuchuan AT impcas.ac.cn>
抄送:
主题: Fw:Re: [Star-cf-l] CF PWG meeting, June 1st, 2023, 9:00 (BNL time)

 

 

------------------ Original ------------------

From:  "Star-cf"<star-cf-l AT lists.bnl.gov>;

Date:  Fri, Jun 2, 2023 11:54 PM

To:  "Star-cf"<star-cf-l AT lists.bnl.gov>;

Subject:  Re: [Star-cf-l] CF PWG meeting, June 1st, 2023, 9:00 (BNL time)

 

Dear Chuan and other PAs,

As already mentioned, I am glad this analysis is moving forward.

Following our yesterday's discussion, I have the following comments and requests regarding your paper proposal (some of them may overlap with Grigory's ones):

General comments: I encourage you to make stronger physics motivation. I understand that it relates to studies of strong interaction, so I suggest building your story this way.

Another essential aspect is that before we move forward, we need to know every detail of your analysis. Please include all analysis details, ensuring you discuss (even as a backup if you don't want to significantly extend your presentation's main part).

- make sure you list all your cuts applied at the event, track, and pair selection level.

Mention all their default values (and later, how you change them to determine systematic uncertainties).

It is essential for all systems you discuss: p-p, p-\Lambda, p-\Xi.

- discuss two-track effects in detail: splitting, merging (all systems you discuss: p-p, p-\Lambda, p-\Xi)

- mention and discuss all corrections: purity, resolution, feed-down contribution, side-band, and all you included in your studies (all systems you discuss: p-p, p-\Lambda, p-\Xi).

-Figure 3: consider how to report your results as dashed lines suggest they are fitting curves which is not the case.

I suggest extracting the source size (together with parameters of interactions) from fitting and, as a separate figure to show model prediction.

- You use LL mode to extract source sizes and interaction parameters, but for the UrQMD model, you calculate correlation functions using the CRAB approach.

CRAB already has included some assumptions about interactions. Make sure you use the same approach (the same parametrization) in the case of experimental data and model.

- Figure 4 shows very modest uncertainties of source sizes. Can you explain how you calculate uncertainties in the case of radii and parameters of the strong interaction?

- You extracted only one set of parameters (f0, d0) for each system. What about singlet and triplet? What are the fitting results in the case of singlet and triplet states considered separately?

- discuss more details about extraction of the parameters: source sizes, f0, d0, discuss shapes of the countours.

- Slide 18 confuses me. I understand you used CRAB to calculate correlations for UrQMD data. What did you do next? Did you extract interaction parameters (and source size) with the LL model? We need to understand what parametrization is included in CRAB you used.

- Discuss your systematics in detail.

Thanks,

Hanna

_

Hanna Paulina Zbroszczyk

PhD DSc Eng, Professor WUT

E-mail: hanna.zbroszczyk AT pw.edu.pl

Tel: +48 22 234 5851 (office)

Address:

Warsaw University of Technology

Faculty of Physics

Nuclear Physics Division

Koszykowa 75

Office: 117b (via 115)

00-662 Warsaw, Poland

Wiadomość napisana przez Grigory Nigmatkulov via Star-cf-l <star-cf-l AT lists.bnl.gov> w dniu 02.06.2023, o godz. 12:29:

Hi Chuan and PAs,

First, I would like to congratulate you with these brilliant and important results!

Please find some questions for your nice study.

General questions:

- Did you take the influence of the momentum resolution of Lambda and Xi, not only p?

- Did you check that protons are not daughter tracks of Lambdas from decays of Lambda->p and Xi->Lambda->p?

- Did you check the two-track effects (merging and splitting) affects daughter protons? How did you do that?

nHits variation and delta-eta vs. delta-phi (not phi*) cuts are quite insensitive to track-merging and track-splitting effects.

- Could you demonstrate the effects of the track-merging and splitting removal procedure?

- When studied pLambda and pXi interactions have you looked at the CFs when hyperons

are selected outside of the signal peak (sidebands)? Could you demonstrate it?

- For final state interaction parameters:

  -- For p-p: I suggest you do add a plot with the world systematics (STAR at 200 GeV, ALICE, others)

  -- For p-Lambda: Please add world systematics (STAR and results from CEBAF)

Abstract:

Please proofread the text. For example:

parametrs -> parameters

remove comma at the of the first line,

etc.

Please make clear what you do with the UrQMD + CRAB and LL model.

slide 3:

Track selection says that transverse momentum criteria is [0.15, 10] GeV/c. Is that for all particles?

Is it true for protons?

slide 5:

Why do you think that the listed values are the sources of systematic uncertainty (especially nsigma values)?

Do you apply purity correction to the correlation functions?

slide 9:

As we discussed yesterday, please update the figure. For example, one could change MC curves with

the markers. Usually curves represent the fit of the function to the data.

Could you show the fits to the data?

pXi should be redrawn because it is completely misleading. Curve does not represent the fit.

slide 10:

Right figure:

- Please explain what pLambda(t) and pLambda(s) are.

- The label says Model. Do you mean UrQMD? If yes, please replace Model with UrQMD

- Why no uncertainties on d0 and f0 are shown for pp pairs?

slide 11:

Please check the uncertainties carefully and write those according to the rounding rules.

Are those uncertainties statistical only? What are the systematic uncertainties on the values?

slide 17:

It would be good to add it to the main figures.

Can you explain why some of the contours  have a "funny" shape? For example, black contour on the right figure?

Cheers,

Grigory

_______________________________________________
Star-cf-l mailing list
Star-cf-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/star-cf-l