Dear all,
Please, have a look at combined comments from Matthias, Jeroen, Gerco,
Marcel and myself before Tuesday 20 h.
There is a remark about MeV which I did not understand myself (?????? general)
and a complicated remark from Gerco which needs our detailed
instructions to ask
for an improvement of the text in the paper (92-94).
Best regards,
Tjeerd
----- Doorgestuurd bericht van i93(a)nikhef.nl -----
Datum: Mon, 31 Mar 2014 11:57:36 +0200
Van: Marcel Merk <i93(a)nikhef.nl>
Onderwerp: Re: [Bfys-physics] Comments PAPER-2014-014
Aan: Tjeerd Ketel <tjeerd(a)nikhef.nl>
Tjeerd,
Hartelijk dank voor het verwerken van de commentaren.
Mijn comments zitten er al in.
groeten,
- Marcel
On 31 March 2014 11:29, Tjeerd Ketel <tjeerd(a)nikhef.nl> wrote:
> Beste Marcel,
> Waarschijnlijk wil je nog wat toevoegen.
> Stuur het dan vandaag nog rond aan bfys-physics.
> Groeten,
> Tjeerd
>
> Comments for LHCB-PAPER-2014-0014:
>
> General comments:
>
> - The paper is very difficult to read without having read
> the BaBar/Belle publications first. This shows a lack of
> interest to who will be reading it.
>
> - The "Z(4430)-" symbol should be following by a noun
> (i.e. Z resonance, Z peak, Z particle, Z parameter).
> - Similarly "Belle" should be followed by "collaboration, data or
> publication".
> - Use the LHCb template symbols style for MeV in line 13, 52 and other.
> ???????
>
> Detailed comments:
>
> - Abstract:
> - Replace "3 fb-1 of data" by the EB standard expression. See EB rules.
> Also used in line 18,
>
> - Replace "a mass (width) of 4475 (172) MeV" by
> "a mass of 4475 MeV and a width of 172 MeV".
>
> -The abstract should be more concise and several parts can be
> removed without missing important points:
> Therefore, remove "as also demonstrated using a model-independent
> approach",
> "and more precise than", "the significance of the Z(4430)- signal is
> overwhelming, at least 13.9 sigma, confirming the existence of this state"
> (This statement is superfluous as Belle has shown more than 5 sigma),
> "in agreement with the 3.4 sigma indication from Belle".
>
> - Replace "more precise than the previous Belle measurement" by
> "more precise than the measurement of the Bell collaboration".
> It is not clear if "previous" refers to the last measurement of Belle
> or to the one-but-last measurement of Belle.
>
> - Replace "significance is overwhelming" by
> "significance is high".
> "Evidence is overwhelming" is also possible.
> Also used in line 166.
>
> - Line 12 or another place:
> Mention that natural units (with c=1) are used.
>
> Line 21-22:
> - Replace "one fewer charged pion" by "one charged pion less".
> - Replace "to minimize mismodeling" by "to improve modeling".
> - Replace "around the Dalitz plot boundary" by
> "near the Dalitz plot boundary".
>
> - Lines 27-32:
> How is this efficiency correction supposed to work?
> If I integrate over phase space to obtain the normalization,
> the best I can hope to achieve is that the average normalization
> is good.
> This does not (at first glance) seem to help
> with efficiency variations over phase space...
> Probably, most readers will not get the point without the Babar/Belle
> publications...
>
> - Figure 1:
> - Move the figure after its first reference in line 72
> or refer to it earlier, preferable with a description of the background
> distribution subtraction and efficiency distribution correction method.
> - Describe in the caption the superimposed curve as the distribution we
> expect
> without resonances in phi'pi. The present text is not suited for in a
> caption.
> - Make the vertical scale arbitrary by using a scale between 0 and 1.
>
> - Line 49:
> Replace "freedom (ndf) equals Nbin-1 minus the number of ..." by
> "freedom, ndf = Nbin - Nvar, with Nvar the number of ..."
> Omit -1 here.
>
> - Line 47:
> Replace "confidence level (CL) of the fit" by
> "p-value of the fit".
> We do not refer to a confidence region here.
> Also correct at other places, e.g. in line 62, 64, 77, 120, and 153.
>
> - Line 48:
> How is it possible that the p-value distribution from simulated data
> is not consistent with flat? Please, explain better.
> This suggests a systematic effect which is apparently NOT simulated.
>
> - Line 46-50:
> Clarify in the text for what reason toy study is done.
> And specify exactly which assumption is tested here.
> The text suggests only that we did not introduce a bug.
>
> Figure 2:
> Replace "the fit variables (black data points)" by
> "(top) the reconstructed invariant mass squared and (bottom) the
> angles theta_phi' and phi".
>
> - Line 83:
> Replace "suggests neglecting D-wave decays" by
>
> "suggests that D-wave decays can be neglected".
>
> - Line 86:
> Replace "p-value" by "$p$-value".
>
> - Line 93:
> Please, clarify:
> Do we know what DeltaL distribution we expect?
> If not, do we fit the expected shape with a chi^2 probability
> distribution (with ndf floating) from toys, and use that to quantify
> significances in data fits?
>
> - Lines 92-94:
> "We find that the simulated ?L distribution follows a ?2 PDF with
> ndf = 7.50 ± 0.11 (statistical error)? :
> what distribution is EXPECTED?
> IS the chi2 distribution chosen more or less phenomenologically?
> What is the meaning of ndf in this case?
> Is it just some parameter used to parametrise the ?L distribution,
> or is it related to something?
> In the next sentence the suggestion is made that it is somehow related
> to the number of free parameters in the Z- fit.
> Also, in what sense does taking ndf=8 result in a CONSERVATIVE result?
> Lastly, the distribution in fig. 3 doesn?t look like a chi2 distribution
> AT ALL!
>
> - Line 95:
>
> You have done toys and concluded that the effective ndf value is 7.5.
> Assuming 8, this is 2x the number of free parameters.
> But what does that mean? Is this a problem with your model?
> Please make this clear in the text.
>
> Line 105:
> Remove "even without considering systematic uncertainties".
> This is not how "consistent" should be understood.
>
> Line 107:
> Rewrite "float all K* masses and widths while constraining" and explain
> how the masses are constrained.
> For example, "by assigning a prior (Gaussian ?) distribution
> to them according to the PDG values".
>
> - Line 117:
> Replace "term in Equation (2)" by "term in Eq.(2) of that paper".
>
> - Figure 3:
> If space for explanations is a problem in PRL, then this trivial figure
> is a luxury which we could move to the added material.
>
> - Lines 122+122:
> Change the differential "d" in the integrals to roman style (2x).
>
> - Figure 4:
> Add the corresponding mass (or mass2) range of the 6 data points.
>
> - Figure 5:
> - Please give the two Z's different colours and symbols.
> - Also add a legend in this plot.
> - Change "candidates" to "number of counts" at the axis.
> - Replace "data (black points) for m2phi'pi in 1.0<" by
> "(black points) reconstructed m2phi'pi for B0->phi'piK+ data
> with 1.0 <".
> - Replace "veto region" by "excluded".
>
> - Line 160:
>
> Write "Figure 5", since it starts a sentence (EB rules).
>
> - Line 243:
> Change "Babar" by "BaBar".
>
>
>
>
----- Einde doorgestuurd bericht -----
