Hi,
	Here are the replies to your comments on the second posting
of the gamma-b paper.
			     Ray

>From dmacqueen@physics.utoronto.ca Mon Apr 23 10:30:33 2001
>Date: Mon, 09 Apr 2001 16:51:38 -0400
>From: Dan MacQueen <dmacqueen@physics.utoronto.ca>
>Subject: CDF Toronto comments on CDF/PUB/EXOTIC/PUBLIC/5450
>
>Dear Authors and Godparents,
>
>The University of Toronto group has read the second draft of the "Searches for
>New Physics in Events with a Photon and b-quark jet at CDF."  We have prepared
>a list of general and specific comments on your paper.  Again, we're sorry for
>the delay in getting these comments to you.
>
>We hope you find these useful.  Thanks again for your paper.
>
>Dan MacQueen,
>for CDF Toronto.
>
>=========================================================================
>
>General Comments:
>
>This paper improves on the first draft, although we still have a few comments
>and questions.  The idea of model-independent limits is better motivated in 
>this version than it was before. However, we would like to reiterate the 
>general comments on the appendix that we made after reading your first draft.
>
>========================================================================= 
>Specific Comments
>Abstract:								       
>1) We suggest "charged leptons" rather than "leptons (e, mu, and tau)"	       

    The reason for keeping it is that we frequently (technically incorrectly)
read "charged leptons" as "e or mu" since they are so more frequently 
analyzed.  The listing makes it clear.  We could drop the word leptons?

>1:									       
>
>1) First paragraph: Are there one or two references which could be added 
>here which discuss new physics in general at CDF?  We don't think that it 
>would be a good idea to add references for each possible type of new	 
>physics (as that would require another page of references at the end),	 
>but one or two references at the end of the paragraph might be useful.	 
>
	We're having trouble finding these referecnes.  Our SUSY review
is a good example, but only covers SUSY and is now 3+ years old.
We don't have any technicolor reviews.  There are recent talks with 
proceedings such as Ray's talk at DPF2000 which was "exotics searches" 
but those only cover a few recent results quickly.  Perhaps a reasonable
overview might be the PDG?


>2) p2: In the introduction, missing Et is written in roman type, although  
>it is in italics for the rest of the paper.				  
>
	fixed

>3) We suggest "leptons (e, mu, and tau)" rather than the more generic	  
>
>"leptons" or "charged leptons" here, in order to make clear that all	   
>three flavours of charged leptons are searched for.			   
>
	Is it OK if we keep it that way both places?

>4) "(massive) third generation quark":  the parentheses are redundant.	   
>
	removed.

>5) Eqn. 1 should have a period on the end, as it ends a sentence.	  
>
	fixed

>2.1:									 
>
>1) Paragraph 2:  We suggest you delete this paragraph -- the only
>important information, the definition of Et, could be put in reference
>[8].
>
	If we removed that sentence (units of GeV) we would either 
have to go through and put in all "/c"'s for momentum etc, or surely
suffer the request to put the sentence back in. (I think it was 
a response to a comment in the begining).  We would like to work
towards using only c=1 and no sentence, but that is probably pushing
it today.


>2.2:
>
>1) Paragraph 1: The first sentence is worded ambiguously.  It does not
>make it clear that there are two separate triggers used, and events must
>satisfy one of two possible trigger criteria.
>
	We don't understand this.  The sentence says "a pair or triggers"
and says that "one or both" must be satisfied.  We can't find the wrong way
to read it, can you send that?


>2) Paragraph 3: "A primary vertex ... is" -> "Primary vertices ... are".
>Since almost all events had multiple interactions, by the end of run Ib
>it would be better to re-phrase this to explain how generic primary
>vertices are reconstructed, without implying that there is only one per
>event
>
	it now reads:
"Primary vertices for the $\ppbar$ collisions are 
reconstructed in the VTX system and a primary vertex is selected as the 
one with the largest total $|p_t|$ attached to it, followed by adding 
silicon tracks for greater precision."


>2.4: 
>
>1) Paragraph 1: You say "the photon cluster can have tracks accidentally
>associated with it and could possibly be tagged; we remove those events."
>Isn't this a repetition of what the second paragraph in section 2.3 says?
>9 lines from bottom: It is not clear how a photon cluster can be
>accidentally "b tagged" when a photon cluster is allowed to have, at
>most, one track with p_t less than 1 GeV and two displaced tracks are
>required for a b tag.
>
	It comes in here since the tracking isolation is just loose
enough that a photon passing all the tracking isolation can still 
be tagged.  In a sense, this is additional isolation.  

>2.5: 
>
>1) First and second paragraphs: "(many) parameters", "is (arguably) not",
>and "(two-dimensional) vector" have redundant parentheses.
>
	We removed the paren on "(many)", left the "(arguably)" 
(we (arguably) think it adds meaning), removed the paren on "(two-dim)"


>2) First paragraph: "is serving as a sieve of" would be clearer as
>"serves to search."
>
	We think the word "sieve" better conveys the sense of a 
comprehensive but necessarily coarse search.  It is a refinement, do
you think the use is misleading or only flowery?  In the later
case we would like to keep it.


>3) Paragraphs 5 and 6 could be combined.  The first one refers to
>additional jets and additional b-tags. The second one only to additional
>photons. Couldn't all these additional objects be described in a single
>paragraph?
>
	You're right, this is better.

>4) In table 1, Et is used when $E_{t}$ should be.  As well, $P_{t}$ should
>be $p_{t}$.
>
	Fixed, thanks.


>3.1:
>
>1) p7, End of 2nd paragraph: You give some sense of the short-comings of
>the CES method. Why not include a similar benchmark for the CPR method?
>
	Added

>3.2:
>
>1) The title of Sec. 2.4 is "B-quark tagging". There could be other kinds
>of tagging (we have B-flavour tagging in the CP/mixing world). It seems a
>bit too much jargon to assume that all tagging, from here on in the
>paper, is going to be b-quark tagging. Couldn't you include "b" to
>qualify tagging in the title of this section and where you use it
>subsequently? We recognise that this will mean there are alot more "b"s
>included in the text, but we think they are important for clarity.
>

	Since we haven't defined tagging at that point, I changed the 
title to "B-quark Indentifcation".  At one point we went through all 
of the possibilities and chose this as the best by our judgement. 
"b-quark tagging" is a little wordy, btag might have worked 
as well, but I think we thought that was too informal.  On the 
other hand, if tag is real insider jargon, then it is neutral to an 
outsider and we can define it and use it as we wish.  If you think
it is very important, we can discuss it more.


>3.3:
>
>1) The sentences that starts "Readers familiar with CDF..." are too
>colloquial. It will discourage readers who are not familiar with CDF from
>reading further or trying to understand. This could easily be phrased
>more neutrally.
>
	We agree, it now reads:
"In the CDF top analysis \cite{top} one of the tagging background 
procedures was to apply the positive tagging probability to the 
jets in the untagged sample to arrive at a total 
tagging background estimate.
A similar procedure could be considered for our sample."



>2) p8, line 4: We were not able to reproduce the "simple" argument that a
>charge 1/3 quark radiates more (or less) frequently than the average jet.
>Clearly charge 2/3 quarks will radiate more. But many jets come from
>gluons that don't radiate at all. So the average probability to radiate a
>photon must come out not that far away from a charge 1/3 quark. At the
>very least this argument doesn't seem "simple" to us and so perhaps we
>could avoid characterising it that way.
>
	I think we intended "simple" to refer to the length of the 
argument that a significant effect may exist (quark charge), rather 
than the computation of the magnitude of the effect.  We removed "simple".

>3.5:
>
>1)Sec 3.5, last lines: We are not sure what "... the result is not
>statistically useful for..." means. Does this mean we can't make a 3
>sigma determination? A 5 sigma determination? We are about to ask the
>reader to take our quantitative limits seriously -- and compare their
>models to them. Please try to be more quantitative here.
>
	We took the characterization out:
"We find the method returns 100\% 
statistical uncertainties for samples of less
than approximately 25 photon candidates."

>3.6:
>
>1) p9, 2nd paragraph, line 2: Can you avoid using "prediction" twice in 6
>words?
>

	fixed


>2) p12,13 Figs 1/2: It is not clear to us why you multiply the SUSY
>predictions by a factor of 4 (40 in one case, 3 later in the paper). Is
>this because the MC prediction is wrong by this factor or just so you can
>see the SUSY shapes on the plots? If it is the latter then a factor of 4
>is not enough in 1a and 1c...
>
	For the latter.  We added "for display" to these captions.
For a) and c) we don't think the problem is worth redoing the plots 
and introducing another scale factor, although we might do it different 
if starting over.  

>4.2: 
>
>1) p15: 3rd paragraph: "The list of backgrounds in Section 3.1..." ->
>wrong section. You mean Section 3.
>
	fixed

>2)"multiply backgrounds estimates and fractions" -> "multiply background
>estimates by fractions".
>
	fixed

>3) p16: 2nd paragraph: the presentations of the systematics is unclear.
>Better say "50 %of the real photon fraction...". Where do you justify
>these 50%/100% numbers?
>
	These are rough estimates from judgement.  50% as a nominal large 
value.  100% from our experience with the met distribution in the 
Monte Carlo.

>4) Last paragraph: the use of 2.7 sigma is confusing and redundant with
>the 2.7 sigma referred to in the next paragraph (with a different
>probability).  We suggest to leave it out. Also, in the counting
>experiment, what does sigma stand for?
>

	The two paragraphs are explaining different measurements.
Each comes up with a probability for the difference measurement 
so they are not redundant.  We think the discussion clearly separates 
the effects so to understand the confuision you will have to be more 
specific.  The sigma refers to the standard conversion of a probability 
to a the area under one side of a Gaussian.  One gave 2.65 and the 
other gave 2.74, so they both round to 2.7.


>5) p17: We suggest that you leave out the last paragraph, which brings
>out more confusion than anything else.
>
	We disagree - in evaluating the significance of a 
fluctuation it is relevant to note the trials factor.
We are stating an important fact, that the significance is less than
the numbers mentioned due to an effect that we have not addressed 
to this point.  We don't understand what information is not useful here.

>6) Figure 3: The body of the figure is too crowded. We suggest that you
>leave out the minimum and maximum probability (besides, the maximum
>probability isn't quoted anywhere).
>
	We are not worried about the crowding, it looks like a typical
blessed plot to us.  We often have more writing explaining the SUSY model,
for example.  Due to the potential for back and forth on this we 
would rather not.

>7) Figure 4: This figure is not referred to anywhere.
>
	It is referred to at the bottom of page 14 at the same 
time as figure 3.

>4.3:
>
>1) p21:  Using negative number of events as predictions can be very
>confusing, especially when looking at figure 6: it seems that there is an
>excess for the 3 jets events in figure 6 b). Maybe you should start the
>axis at -10, if you keep negative numbers in table 8.
>

	We realized we were being inconsistent with reporting 
a background with the CES/CPR method and less than 25 events so most
of these are removed.  There does remain one negative background
which the godparents discussed at some length.  We decided to leave 
it as is - a non-zero scale is also misleading at a glance.


>2) 1st paragraph: "...photon background, When" -> "...photon background.
>When"
>

	fixed

>3) 3rd paragraph: "E_t > 30 GeV.^{2}" -> "E_t > 30 GeV^{2} ."
>
	fixed

>5.1:
>
>1) When reporting the limits on this specific model, your choices of
>parameters (such as tan beta = 1.2 and stop mass = lightest neutralino
>mass + 5 GeV) might seem less arbitrary if you explained in more detail
>why those values were chosen and how sensitive your results are to your
>particular choices.
>
	We added a sentence saying that we are optimizing 
for acceptance, which was not clear.  As for the individual 
parameters, we state what we are doing and why.  The question of
sensitivity is a good one.  It would take a long time to quantify, but 
qualitatively we are sensitive to the parameters.  We added:
"In addition, these models can show sensistivity to small changes in 
the parameters."

>5.2:
>
>1) Since 5.2, 5.3, and 5.4 discuss the model introduced in 5.1, they
>might be better numbered as 5.1.1, 5.1.2, and 5.1.3.  Thus, 5.5 and 5.6
>would become 5.2 and 5.3.
>

	We did a similar reorganization.  5.1 is now all the 
N2->gamma N1 model, with direct production, squark and gluinos being 5.1.1
and 5.1.2 etc.  The GMSB model is 5.2.


>5.6:
>
>1) 5th paragraph: "The first point is overly efficient..."  This is a
>confusing sentence -- does is refer to table 14?  If so, it should
>probably say "first entry" instead.  Also, it is not clear what "overly
>efficient" means.
>

	We changed it to 
"The first point appears to have an unusually large efficiency 
because of other sources for $b$ quarks 
which are not reflected in the definition of the 
signal branching ratio."


>6.1:
>
>1) (85 +/- 3.4) pb^-1 should be (85 +/- 3) pb^-1.
>
	fixed

>Appendix:
>
>1) This section needs to be reread carefully to ensure its readability.
>
>2) Your 6-point list of advantages to the model-independent approach
>might read better as a paragraph.
>
	We have discussed this with godparents and prefer this solution.  
Let us know if you feel strongly.


>3) Point 2 might read better as "No a priori judgement is necessary to
>determine what is an interesting model."
>
	done

>4) Second last paragraph: This sentence reads awkwardly, and is not
>neutrally phrased (see our earlier comment on the phrase "Readers
>familiar with CDF.)  We would suggest "In this appendix, we present
>three methods to calculate A\epsilon.  The model-independent limits
>presented in this paper set limits on the model and parameter of choice."
>
	How about the following:
	"In this Appendix, we present three methods to calculate 
$A\epsilon$.  These results together with the model-independent limits
can be used to set limits on most models 
that predict events with the $\gamma b \met$ signature."


>References:
>
>1) When reading your previous draft, we complained that in some
>references to multiple papers with the same authors, such as Ref. [1],
>[11] and [13], the authors' names weren't repeated.  According to the APS
>style guide, this is acceptable.  However, in references [4] and [14],
>you do repeat the authors.  This should be done consistently.
>
	fixed


>Cheers,
>
>------------------------------------------------------
>Dan MacQueen -- Ph.D. student -- University of Toronto
>--- (416)978-6632 -- dmacqueen@physics.utoronto.ca ---
>------------------------------------------------------