[Olympus] Fwd: Re: Revised PRL paper

[Belostotski, Stanislav]

Dear Douglas and colleagues,
I wonder what is the status of PRL draft.
Are we in second run of mailing exchange with referees?
I have to apologize for phrasing  it.3 in my last mail.
Saying "weighted mean" I had in mind the result and not the error bar of 
course. As I understand all four analyses performed using one data set. 
If so the stat. error bars must be close one to another. On the other 
hand the systematic error might be estimated differently in each 
analysis depending on method used.
  I feel lack of information on how the data had been obtained. There 
needs a detailed report on that. For example, to show the results of all 
four analysis (internally of course) would be a good help.
With best regards stanB


-------- Original Message --------
Subject: Re: [Olympus] Revised PRL paper
Date: Sat, 10 Dec 2016 23:34:40 +0300
From: Belostotski, Stanislav <stanislav.belostotski at desy.de>
Organization: Desy
To: Douglas K Hasell <hasell at mit.edu>
CC: olympus <olympus at mit.edu>

Dear Douglas and all,
The paper now looks better but  some things must be still clarified.
1.First about MIE. It is now specified  which value is used for
monitoring: ratio of what to what. However the statement
" By extracting luminosity from a ratio of
rates, the MIE analysis exploited cancellation of several
systematic uncertainties, reducing the uncertainty in the
relative luminosity between beam species to 0.36%."
    must  be  clarified   by giving at least  one example of mentioned
above cancellation and clear indication of what is the main source of
the systematic uncertainty advertised.
2. Fig.3  I am afraid misleading, Effect is small and for sure
comparable with uncertainty in Blunden curve.The calculated TPE  is
function BOTH epsilon and Q**2. Then we must show  the similar  plot for
   Q**2. Why should we show the comparison versus epsilon and should not
versus Q**2?
3.When calculating the statistical error bar unifying a pair (or
several) independent analyses, it should not be taken as average. If we
believe in the results and trust in all variant of the analysis
it must be calculated as weighted mean. As for systematic uncertainty,
this is always difficult as there is no  definite prescription for that.
Cheers,  StanB



On 08.12.2016 7:42, Douglas K Hasell wrote:
> Dear Colleagues,
>
>          The attached PDF is the revised paper after responding to the concerns of the PRL referees.  I also attach the letter I would propose sending back to PRL to address the referee’s questions.
>
>          Please send me any comments or suggestions. as soon as possible or by Monday, December 12 at the latest.
>
>          Thank you.
>
>                                                       Cheers,
>                                                               Douglas
>
> 26-415 M.I.T.                                  Tel: +1 (617) 258-7199
> 77 Massachusetts Avenue                        Fax: +1 (617) 258-5440
> Cambridge, MA 02139, USA                       E-mail: hasell at mit.edu
>
>
>
>
>
>
>
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: http://mailman.mit.edu/pipermail/olympus/attachments/20161207/b66ed71f/attachment.html
> -------------- next part --------------
> A non-text attachment was scrubbed...
> Name: tpeprl.pdf
> Type: application/pdf
> Size: 344981 bytes
> Desc: tpeprl.pdf
> Url : http://mailman.mit.edu/pipermail/olympus/attachments/20161207/b66ed71f/attachment.pdf
> -------------- next part --------------
> An embedded and charset-unspecified text was scrubbed...
> Name: referee_round1_with_answers.txt
> Url: http://mailman.mit.edu/pipermail/olympus/attachments/20161207/b66ed71f/attachment.txt
> _______________________________________________
> Olympus mailing list
> Olympus at mit.edu
> http://mailman.mit.edu/mailman/listinfo/olympus
>
_______________________________________________
Olympus mailing list
Olympus at mit.edu
http://mailman.mit.edu/mailman/listinfo/olympus