STAR Flow, Chirality and Vorticity PWG

["Wang, Fuqiang" <fqwang AT purdue.edu>]

Hi Prithwish,

Let me add my comments below. 

Best regards,
Fuqiang



> -----Original Message-----
> From: Star-cme-focusgroup-l <star-cme-focusgroup-l-bounces AT lists.bnl.gov>
> On Behalf Of Prithwish Tribedy via Star-cme-focusgroup-l
> Sent: Friday, June 10, 2022 12:05 PM
> To: star-cme-focusgroup-l AT lists.bnl.gov; Takafumi Niida <niida AT bnl.gov>
> Cc: STAR Flow, Chirality and Vorticity PWG <star-fcv-l AT lists.bnl.gov>
> Subject: Re: [Star-cme-focusgroup-l] [Star-fcv-l] STAR presentation by 
> Yicheng
> Feng for SQM 2022 submitted for review
> 
> ---- External Email: Use caution with attachments, links, or sharing data 
> ----
> 
> 
> Hi Yicheng,
>     I would like to sign-off your talk and move it to star-talks. I am OK 
> to move
> with your current slides except for the one on R-correlator which needs
> polishing.
> 
> I find the linear correlations between 1/\sigma^2 and 1/Nch even after 
> scaling
> by shuffled event width in isobar data and with AVFD simulation of 
> 2203.10029. I
> believe there are some concerns that might lead to confusion. So, I suggest 
> we
> go through a couple of iterations on slide
> #17 and polish the wordings.
> 
> https://drupal.star.bnl.gov/STAR/system/files/SQM2022_YichengFeng_v14.pdf
> 
> 
> I’ll use “~” sign to indicate proportionality.
> 
> Left part:
> “Normalized by shuffled width, so the N_ch is already scaled out”.

I suggest we add a word "trivial" in front of N_ch.

> —But from both isobar data and the AVDF simulation of 2203.10029 we see that
> 1/\sigma^2 has a linear relation with 1/Nch.
> 

This doesn’t mean that the scaled quantity must be independent of Nch, just 
like N*Dgamma/v2 is not.

> “STAR has concluded that 1/\sigma^2 is approximately proportional to v_2”.
> — But the isobar data do not show that from your plot.

Again, this statement does not mean the data must show the exact 
proportionality. The statement means that algebraically/mathematically there 
is such a trivial dependence. Any deviation from such a dependence is 
possibly physics and the whole point of our research. 

> Also, I am not sure if STAR concluded this. In the isobar paper we say Ref. 
> [18]
> indicate an approximate scaling which is N* Δγ and based on that we quote a
> number 1.02.

Because N*Dgamma ~ v2.

> 
> “After v_2 scaling, the isobar ratio is even further below unity.”
> — From your plot we do not see 1/\sigma^2 to be proportional to v_2 so we
> cannot scale it. This is based on the same argument that we can’t scale by 
> 1/Nch.

Again, we scale Dgamma/v2 by 1/Nch, NOT because the data show Dgamma/v2 ~ 
1/Nch, but because it is mathematically encoded in the definition of Dgamma. 

> Right part:
> Top line says “an additional” which is not clear.
 
It refers to the fact that 1/Nch is already scaled in the Delta S' definition.

> 
> “The apparent 1/Nch dependence is not proportional, Why scale by 1/Nch ?”
> — This is not clear. They saw a linear dependence between 1/\sigma^2 vs 
> 1/Nch
> and applied a scaling. So are we saying they should have scaled by “a + 
> b/Nch”
> not by “1/Nch” ? Or are we saying they should have not done any scaling at 
> all?
 
Even if the data showed a precise proportionality, I would not scale it 
because it is not a mathematical given. We must understand it, instead of 
scaling it without understanding it. 

The statement on Yicheng's slide is meant to say: Okay, the data show a 
dependence and you may go ahead to scale it even though you don't understand 
it, why do you scale by Nch, not by a+b/Nch?

> 
> “The apparent 1/Nch dependence comes from 1/\sigma^2 ~ v_2, and v_2
> depends on N_ch”.
> — From isobar data we do not see 1/\sigma^2 ~ v_2 so how can we say this?
> I also have not seen an AVFD simulation plot showing 1/\sigma^2 ~v_2 — I 
> have
> requested this before.
 
Again, it is not about what the data show you. It is what you expect from 
general ground (algebra, math). 
Just for the sake of argument: if data showed us that Dgamma/v2 was flat as a 
function of Nch, I would not conclude that we should not scale it by Nch. 
Instead, we should scale it by Nch because that's just math, and then 
question what physics is making Nch*Dgamma/v2 ~ Nch so Dgamma/v2 is flat vs. 
Nch.

> 
> Also, “v_2 depends on Nch”
> — this is correct. But its more like "v_2 ~ 1/√Nch.” So how can we say
> "1/\sigma^2 ~ v_2 ~ 1/Nch” ?
> 
 
If you look at 1/R^2 vs 1/Nch, you see a linear trend but if you factor in 
the (0,0) then it's probably close to sqrt(1/Nch). Similarly for v2 vs 1/Nch. 
And this results in the 1/R^2 vs N*Dgamma dependence which is not exactly 
proportional but not far. In any case, all these are in the data. We need to 
understand the data. We do not arbitrarily scale data with whatever we 
observe in data. 

> 
> “From Zr+Zr to Ru+Ru 1/Nch decreases and v_2 increase.”
> — This is not clear what we’re saying here. I guess we are saying that:
> 
>   1/\sigma = f(v_2, Nch) that is the width is probably a function of both
> v_2 and Nch but then the argument is not clear.
> 
> 1/\sigma (Ru)/ 1/\sigma (Zr) =  f(v_2, Nch) (Ru) /  f(v_2, Nch) (Zr)
> 
> And we know  v_2(Ru) > v_2(Zr)  and also 1/Nch(Ru)< 1/Nch(Zr) but then what 
> ?
 
If you scale with 1/Nch you get a ratio >1, and if you scale with v2, you get 
a ratio <1. 
The point is: you don't scale things arbitrarily. You have to understand the 
reason why you do such and such scaling.
We're comparing two very close numbers, and you're scaling them with a gross 
function and try to conclude on the resultant subtle difference? 

> 
> “Lacey et al. conclusion by 1/Nch scaling contradicts STAR’s conclusion”
> — Not clear which STAR conclusion is being contradicted. The conclusion of 
> the
> STAR paper was “no predefined criteria for CME was observed in blind 
> analysis
> which includes 1/\sigma^2 ratio to be less than unity”.
> We’re not saying that here. In the post-blind section we have a baseline for
> 1/\sigma^2 which is N* Δγ = 1.02 for 20-50%. However, it is based on the 
> scaling
> relations extracted in Ref. [18]. So are we saying that the baseline 
> extracted in
> the post-blind section is contradicted. 
 
Yes.

What if someone says it was a naive
> baseline and 2203.10029 presents an updated baseline based on AVFD. This
> needs some polishing.
 
That's the point of the debate, right? We don't just take what people say. We 
ask question why such scaling, is it credible? We exercise our own reasoning 
and form judgement, etc. and that's how we progress.

> Best,
> Prithwish
> 
> 
> 
> On 2022-06-01 22:19, Yicheng Feng via Star-cme-focusgroup-l wrote:
> > Hi All,
> >
> > I have uploaded a new version of slides to the same node:
> > https://drupal.star.bnl.gov/STAR/node/59776
> >
> > where one new slide (page 17) is added for the R observable.
> >
> > Sincerely,
> >
> > Yicheng
> >
> > -------------------------
> >
> > From: Star-fcv-l <star-fcv-l-bounces AT lists.bnl.gov> on behalf of
> > webmaster--- via Star-fcv-l <star-fcv-l AT lists.bnl.gov>
> > Sent: Monday, May 30, 2022 5:36 PM
> > To: star-fcv-l AT lists.bnl.gov <star-fcv-l AT lists.bnl.gov>
> > Subject: [Star-fcv-l] STAR presentation by Yicheng Feng for SQM 2022
> > submitted for review
> >
> > Dear star-fcv-l AT lists.bnl.gov members,
> >
> > Yicheng Feng (feng216 AT purdue.edu) has submitted a material for a
> > review, please have a look:
> > https://drupal.star.bnl.gov/STAR/node/59776
> >
> > ---
> > If you have any problems with the review process, please contact
> > webmaster AT www.star.bnl.gov
> > _______________________________________________
> > Star-fcv-l mailing list
> > Star-fcv-l AT lists.bnl.gov
> > https://lists.bnl.gov/mailman/listinfo/star-fcv-l
> > _______________________________________________
> > Star-cme-focusgroup-l mailing list
> > Star-cme-focusgroup-l AT lists.bnl.gov
> > https://lists.bnl.gov/mailman/listinfo/star-cme-focusgroup-l
> _______________________________________________
> Star-cme-focusgroup-l mailing list
> Star-cme-focusgroup-l AT lists.bnl.gov
> https://lists.bnl.gov/mailman/listinfo/star-cme-focusgroup-l