Dear authors of the GMSB susy search in diphotons,

the spokespersons publication review group, this time represented by Kevin
Burkett and Tommaso Dorigo, has read your first PRL draft and
congratulates you for bringing to publication this nice new result. 

==>Thank you.

Below
we offer a few general comments and specific corrections that we hope will
improve the text. Please note that Figure 1 appears wrong, please clarify.


General Comments:
-----------------

- The analysis is sound and produces a significant improvement
on the previous limit.  

==> Thank you.

- However, the discussion in the text
of the background contributions does not flow logically and
is hard to follow.

==> We have worked to make the background description easier to follow.
For more see below in the line-by-line comments.

- It would be significantly improved to add a few introductory
sentences describing the contributions as they are labelled in
Figure 1: SM with fake MET, EWK w/real MET, and non-collision.
That can include a brief explanation of the specific contributions
from fake photons, etc.  More details can be added in later
paragraphs, including the way that analysis cuts are used to
reduce those specific backgrounds.  Right now those cuts are
not introduced until the discussion of the optimization.

==> Done. For more see below in the line-by-line comments.

- a few remarks in the paper are obvious (eg P2 L59), others appear
unnecessary (eg P2 L43), and at least one is false (P1 L88).

==> Done. For more see below in the line-by-line comments.

- Important: the quoted backgrounds are inconsistent with what is shown in
Fig.1! Please check the figure. Suggest to start y axis at 10^-3.

==> See responses to P2 Figure 1 below in the line-by-line comments.

Line-by-line comments:
----------------------

P1 Abstract: Suggest "We set a 95% CL lower limit on the
chi^0_1 mass of 149 GeV/c2 for the chi^0_1 liftime tau=0,
as well as make exclusions in the chi^0_1 mass-lifetime
plane ...". Excluding points in the plane is already a statement about
reaching more than previous experiments, and we think we should not stress
with "world's best limit" a part of the investigated plane.

==> World-best limit is important. Simply saying that we exlude points 
doesn't mean that our limit is the world's best.
There were many previous searches and our limit significantly improves 
the results among them.

P1 L16: hyphen on "low-mass".

==> Done.

P1 L19: "We consider the scenario" --> "We consider a scenario"

==> Done.

P1 L21: "almost exclusively" is vague. Can we quantify ?

==> Alomst 100%. Even if the decay modes into the Z^(0) boson or the neutral
Higgs boson are kinematically allowed, they are highly suppressed by the
beta^(8) phase-factor. For m_chi10=140 GeV the branching ratio is 96.4%. We
have added that it is > 96% to the text to make it more explicit.

P1 L26: keV/c2

==> Done.

P1 L28: "to be larger than 100 GeV/c2" (spell out).

==> Done.

P1 L29: it is clearer to say "occurs primarily into" rather than "is
dominated by".

==> Rephrased the sentence.

P1 L30: pedices should have smaller fonts. This goes for all pedices
labeling particles in the text.

==> Done.

P1 L35: should use "Letter".

==> Done.

P1 L34,36: You say that different strategies are needed for
lifetimes above and below "about a nanosecond", and then two
lines later say you will search for GMSB with "tau<=2 ns".
These two statements seem inconsistent.

==> In line 34 we are emphasizing the order of lifetimes. 
What you are suggesting makes it look like the edge is a hard one and
there is clearly overlap in the sensitivity regions. We are extending beyond
what is given to provide extra coverage.

P1 L42: "This is ten times more data than our previous search" sounds a
bit awkward. Better "This dataset is ten times larger than the one used in
[8]".

==> Done.

P1 L47-50: Suggest "which allows us to considerably extend
the sensitivity of the search for large chi^0_1 masses
compared to other Tevatron searches.  We also exted the
search to consider lifetimes up to 2 ns."

==> Done.

P1 L75: "from our diphoton events" is awkward. Please refer to signal
events. Also, what does this mean ? That the trigger is 100% efficient on
signal after selection, or before any selection ?

==> After diphoton selection with ET>13 GeV, |eta|<1.1.
Sentence rephrased.

P1 L88: This does not appear to be true. We do not cut z<60cm to "maintain
the projective nature of the photon reconstruction in the calorimeter" in
any of our analyses. We do it for the other reason you list, by and large.
The fact that this time we get an added bonus is no justification for
twisting the fact: it looks like a deceiving statement.

==> Rephrased the sentence. 

P1 L90: "...require the primary collision vertex position..."

==> Done.

P2 L2: please explain the need for the picking of the highest SumPt
vertex, if we anyway end up choosing the one with smallest missing Et. 
Either the sentence is wrong, or it is inconsistent.

==> Rephrased the sentences, following your suggestion.

P2 L21: no reason to include the citation for METMODEL again

==> Done.

P2 L42: how do triphoton events contaminate our sample ? Is it necessary
to mention this irrelevant source here ? What is the rate of such process
?

Triphoton events are an important background. 
Our study shows that wrong vertex and triphoton events are the dominant sources 
in event reconstruction pathologies in QCD type backgrounds. 
After final kinematic cuts the combined rate 
from wrong vertex and triphoton events is 0.06, which is about 13% of total SM
background with fake MET. See Section 3.2.2 and Table 20 in Section 6 of
cdfnote 9575.
 

P2 L53: "W's and Z's" sounds awkward, suggest "W and Z bosons".

==> Done.

P2 Figure 1: The number of background events quoted in the text (1.4) is
inconsistent with what this plot shows (bottom part), where above 200 geV
of Ht one can only account for at most 0.7 events. Where is the rest ? Is
this an error of normalization ? Furthermore, if one looks at the white
and grey bins above 200 GeV of Ht, one sees that the white ones (SM with
fake MEt) appear to have a larger normalization than the EWK with real MEt
ones. So not only does the whole normalization of the region above 200 GeV
look incorrect: it also appears that the two processes have wrong relative
fractions there.
Please plot Figure 2, bottom, with a y axis starting at 0.001 such that
the rest of the SM background shape can be seen. Does it flatten out or
does it die away after 350 GeV ? This is an important check and we expect
you to get back to us with a quick answer.

==>You are correct that the bottom part of the Figure 1 was plotted in a
misleading fashion and been fixed. A quick explanation is that the analysis
uses a global EWK scale factor calculated by integrating all events above an HT cut
since there are small statistics around the cut region (See Section 6 from cdfnote 9575). 
This is why there is a large error on the scale factor. 
Using a single scale factor is fine for the overall analysis, 
but is not fine for the plot since the scale factor can vary
significantly (again, taken into account in the analysis and optimization as
well as in the uncertainty). A better estimate of the bin-by-bin background
estimate for the HT plot in the bottom of Figure 1 is to use a bin-by-bin
correction to the EWK background. We have changed the plot to this format. In
the old plot the scale factor for all events was smaller than the scale factor
for HT>200 GeV. Thus, the plot contained a smaller scale factor which made the
estimate above 200 GeV appear smaller than it should have. Now that we plot
bin-by-bin the plot and the analysis are more consistently displayed.
As described in Section 6 of cdfnote 9575, this EWK global scale factor varies 
from 0.6 to 1.8 while other normalizations 
for SM with fake Met and non-collision backgrounds are fixed. 
This is why their fractions look different in the figure
To produce the bottom figure 1 with no cut on HT we applied a global scale factor 0.78 accordingly, 
but after the final cuts including HT>200 we have to apply a different global scale factor (1.41). 
Now we changed the plot with bin-by-bin scale factor and posted it at
http://hepr8.physics.tamu.edu/elee/research/n1_ht_data_2.6fb_bin.gif
To see why the scale factor is a function of HT see
http://hepr8.physics.tamu.edu/elee/research/n1_ht_egamma.gif which shows
the egamma data and MC comparison. The bin-by-bin ratio can be found at
http://hepr8.physics.tamu.edu/elee/research/ewk_sf_ht.gif. The scale factor
(the ratio of the integrated number of events above the HT cut) as a function
of cut can be found at
http://hepr8.physics.tamu.edu/elee/research/ewk_sf_ht_int.gif. As can be seen,
with no cut the scale factor would be 0.78+-0.02 (stat. only), and above our cut it is
1.4+-0.44. The plot in the PRL was with a constant scale factor 0.78
(http://hepr8.physics.tamu.edu/elee/research/n1_ht_data_2.6fb_pub.gif). A
version with a constant scale factor is shown at
http://hepr8.physics.tamu.edu/elee/research/n1_Ht_data_1.4.gif. The new
bin-by-bin corrected version can be found at
http://hepr8.physics.tamu.edu/elee/research/PRLfigure1_bin.gif
This is the version that is shown in the PRL.
 
Another point: the non-collision background clusters at 80 GeV of Ht.
Those events correspond to the ones at significance of 10 in the upper
plot (except for the absence of a DP cut there). Question (mainly a
curiosity): how does a missing Et not exceeding 40 GeV (for a 80 GeV Ht)
get such a large significance ? Is it due to the absence of any hadronic
energy in the calorimeter for these events ?

==> Yes. These have very large MetSig because there is very little hadronic energy in
the event. It isn't surprising that these events which have lots of real MET
(since at least one of the photons is fake) that it is measured to be
significant. Another way of saying "significant" is saying "not due to typical
calorimeter measurement fluctuations". This is to be expected and consistent
with expectations. 

P2 L58: "and taking into account ...": now this specification is really
not needed. It is obvious to non-experts, and while experts might for a
second wonder whether we do include tau decays, they certainly conclude we
do, otherwise we'd say we do not.

==> We have made it explicit since it's small amount of text. 
We removed "of the bosons" at the end to shorten as you suggest.

P2 L62: "ISR/FSR simulates W/Z+gamma gamma events" -->
"ISR/FSR produce W/Z+gamma gamma events"

==> Done.

P2 L66: "MC-data".

==> Done.

P2 L75: cosmic-enhanced, beam-related need hyphens.

==> Done.

P2 L83: this sentence would sit better in a separate paragraph.

==> This is thoughtful, but we prefer not to have a single sentence paragraph
as stylistic preference.

P3 L18: "limit with the no-signal assumption without" -> suggest
"limit using a no-signal assumption, without". Mind the comma.

==> We have fixed this sentence.

P3 L21: "acceptance, and". Mind the comma. Same, on line 26 after "jets".

==> Done.

P3 L27-42: This discussion is important and should be moved
earlier in the paper, rather than being treated as a detail
of the optimization.

==> The backgrounds overview has been changed as you suggested. 
However, these cuts are a part of optimization and would look out of place earlier. 
We did indicate that the HT cut was coming earlier on page 1.
We prefer to leave as it is.


P3 L30: "high Et, light final state" -> "high-Et light final state". No
comma here.

==> Done.

P3 L31: Suggest "The GMSB".

==> Done.

P3 L35, 39: Please add hyphen to "high-Et".

==> Done.

P3 L45-48: confusing sentence. Colon makes no sense. Please rewrite.

==> Done.

P3 L50: Please add comma after "uncertainties".

==> Done.

P3 L55: Saying "events with 0.9+-0.4 events from ew sources" is less clear
than saying "events, 0.9+-0.4 of which from ew sources" which we suggest.

==> Done.

P3 Figure 2:  Throughout the paper you say you consider
lifetimes below 2 ns, so why does the plot extend to 3 ns?

==> That way it is more clear we can exlude lifetime up to 2 ns. It is just
included for completeness. We have changed the text earlier to say
approximately equal to 2 ns rather than as a hard cutoff.

P4 Figure 3: the observed exclusion region is too light-shaded, does not
show up in BW printouts. Also, the favored region should be hatched in a
totally different style to stress it is not an exclusion and ease the
understanding of the plot.

==> This is the same color scheme as in delayed photon PRL/PRD. For this
reason we would like to leave it as is.


P4 L1-2: Suggest dropping "by the end of run 2..."
Suggest simply "With an integrated luminosity of 10 fb-1,
we estimate a mass reach of ..."

==> "By the end of run II" tells why we picked 10 fb-1. We prefer to leave it
as is.

P4 L23-P5 L9: Please follow AIP rules for citations.

==> Done as noted below.

P4 L23: add ", and" in short list of names. This goes for all the
reference list. Also, add comma before "and" in several instances in a
list of authors.

==> Done.

P4 L26: remove "and" after colon. Sure about "Rept." ?

==> Done. Rept. should be Rep.

P4 L35: missing transverse energy should be spelled in full here, when
defining it. "Missing transverse energy is defined as...".

==> All previous three papers(ref [8] and [16]) use the same way, 
not spelling it out this way in the reference. We have left this as is.

P4 L69: incomplete title and reference, please fix.

==> Done.

P5 L3: "The initial and final state radiations" is awkward, suggest
"Initial and final state radiation".

==> Done.

That is all, thank you
Tommaso for the SPRG