Thanks to everyone for carefully reviewing the PRL.
-- Comments from Sarah Eno ---
Q. Is the direct photon background calculated using PYTHIA?
A. Yes it is calculated using PYTHIA 5.7.
Q. Was the DDP background negligible?
A. No, it was calculated at 0.78 +- 0.32 events.
Q. Is Marc's limit setting code working properly?
A. Yes it does work properly. I checked several point's using a standard technique and the both methods produced the same results. Quickly, N = (xsect) . (Luminosity) . (Efficiency - 1S.D. in Eff.) The probability for zero observed events to fluctuate to three is 5%, thus 3 = xsect(95%c.l.) 101.2 (8.3-1.25) xsect(95%c.l.) = 0.42 pb and using Marc's method I get 0.41 pb. For all the values I calculated none differed by more the .01 pb.
--Hugh Montgomery ----
Q. Did you compare the gamma-gamma mass resolution to the Z-ee mass resolution?
A. Yes. M(gg,mc) = 90 GeV/c^2, dM(gg,mc) = 2.8 GeV/c^2 M(ee,mc) = 91.3 GeV/c^2, 3.0+-0.3 GeV/c^2
--Paul Grannis ----
Q. I thought the jet --> photon faking probability would be ET dependant?
A. If the energy isolation requirement on the photon candidate is 2 GeV, there is a strong ET dependance ( the flat 2GeV requirement is used by the photon analysis group). This analysis used a 10%xET isolation requirement, thus much of the ET dependance is lossed. A plot of dependance can be seen in Tom Fahland's Thesis.
--Stefan Grunendahl ----
Q. Has the acceptance for the fusion contribution been studied?
A. Yes. I used the same methods of MC generation and detector modeling as used for my HW/HZ acceptance and efficeincy calculations. The photon and jet Efficeincies for fusion where very similar to that of HW/HZ as they should, and the acceptance was lower than HW/HZ due to the forward radiated jets. All of the kinematic distributions seemed very reasonable, and nothing looked suspect.
--Jay Wightman ---
1) 1st person plural...
bryan
I will let the D0 style committee decide on this. I do not agree that this style is self-aggrandizing.
2) Final state cross section for ggjj channel...
bryan
Good observation Jay. The branching fraction plot was not correct, but the final cross section plot is. I corrected the BF plot, so that the two distributions are consistent.
Your example was the 100 GeV/c^2 Higgs.
xsect(HW/HZ) = .4 BF(W/Z -> jj) = .7 BF(H -> gg) = .3 xsect(fusion) = .041 , this is the Higgs fusion correction to the final xsect. Just need to fold in H -> gg BF.
xsect(limit) = (.4)(.7)(.3) + (.04)(.3) = .1 pb //
I further checked every point in the final cross section, and all are consistent with what is shown.
- Anomalous triple vector...
bryan
I think that it is clear from the abstract of the paper that this measurement is not made. The point of this sentence is that more can be drawn from this measurement than just the bosonic Higgs limit.
3) Eta definition with detector eta limts....
bryan
I extracted this method from previously published D0-prl's. I added a theta definition as well.
4) Photon's vs. electron conv....
bryan
I agree, I removed the electron conversion bit.
5) Trigger requirements...
bryan
I will add a sentence that a two photon trigger is used, and that it is fully efficeint at 15 GeV.
- Average size of photon or jet.
bryan
This is inappropriate for this paper. I quoted the standard D0 detector paper which covers this.
- How many photons have part of their shower in the ICD?
bryan
If part of the photon's shower is in the ICD, it rarely passes the EM requirements. For that matter, the photons would rarely pass the diphoton trigger. The photon has poor acceptance in general if it hits any calorimeter crack.
- Dijet Mass resolution...
bryan
dM(jj) ~ 20 GeV/c^2, but this can be misleading. The dijet mass distribution does not have a pure Guassian shape, like the Z ee resonance. The W/Z -> jj resonance has a long tail toward higher Mjj masses. Thus, I used a -2dMjj lower limit and a +4dMjj upper limit (roughly). It is well known that dijet mass resolutions are poor. For brevity, I left the previous out.
- Selection criteria, reason for their choice...
bryan
I will add a sentence here. Regarding the rejection vs. background, again I think brevity is important.
- MI issue...
bryan
I just take the primary vertex.
- Charged track issue...
bryan
The events have two high ET jets which are used for reconstructing the vtx.
6) Momentum fraction of meson....
bryan
The reference covers this thoroughly. This method was used in all photon xsect papers. I think one paragraph is enough.
7) "true" direct photon back.....
bryan
OK, I removed the "true".
- W->ev+g processes...
bryan
I use all processes, this background is sharply rejected by the tracking requirements.
- misvertexing issues...
bryan
Misvertexing issues have been thoroughly studied.
8) Page 4, 1st column, last paragraph
"Both EM clusters are required to have more than 90% of their energy in the EM section..." "...at least one of the two EM clusters is required to fail the photon quality criteria (isolation, shower shape, or EM fraction)." The clusters being considered have already failed the EM fraction requirement. You do not impose the tracking requirement at all? How does the efficiency for this relaxed set of requirements compare with the tight set? You should emphasize the fact that mgammagamma < 60 GeV/c2 is the mass range where there is no expected bosonic Higgs contribution.
bryan
Read this more carefully. I stated that the background photon candadates must be PPHO's at least (EMF>90% and no recoed track). The background photons to my signal sample photons scale similarly in all kinematic distributions I studied. Thus, their efficeincies scale similarly. I added your last comment to the normalization sentence.
9) Page 4, 2nd column, 1st paragraph
Your direct photon samples are 1 photon with 3 jets and 2 photons with 2 jets. You should be explicit about this.
bryan
This is mentioned thoroughly in the 2nd sentence of this paragraph.
10.)Page 4, 2nd column, 2nd paragraph
"The total expected background is shown in Fig. 2 as the expected number of background events with errors per 10 GeV/c2 bin." How can the total number be expressed as an expected number per bin?
"If we increase the gamma..." How can the same background estimation give the expected number of background events in this case? The same technique may do so, but the same estimation cannot. I might suggest the following wording for here: The background calculations are verified by relaxing the cuts to ... and finding 38+/-10 events predicted with 39 events detected.
Bryan
The total number of events is the sum of the binned distributed events. I added technique to the sentence.
11) Is it important to state how many events in each sample? More importantly, the statistical uncertainty on 5000 events is not 3% as quoted on the following page.
"The events are processed with the detector simulation..."
Bryan
Statistical uncertainty is 3% because not all 5000 events pass into the final sample. More like a thousand events do, thus ~3%.
12.)Page 5, 1st column, 2nd paragraph
In your Bayesian limit did you include all correlations? As I recall your original limits did not include any correlations.
"...using PYTHIA and corrected using a K-factor of 1.25[8]." What is the dependence of this factor on mHiggs? Are there any other dependencies of the K-factor? What is the K-factor trying to simulate?
Bryan
I added a sentence stating that correlations are not included. The K-factor is a NLO correction factor. The reference states that the K-factor has little kinematic dependance.
13.)Fig. 1
I think the two plots should have the same vertical axes; they already have the same horizontal axes. What is the minimal model Higgs? If you mean Standard Model Higgs, then state it as such.
"...where V = gamma, W, or Z."
Bryan
The hard-drive crashed so this is a difficult fix. I will leave it for now. I am not sure if I agree with you about making them the same...what does it matter. I added standard to statement. I fixed V.
14.)Fig. 2
Is the binning in mjj consistent with the jet mass resolution? If not, you are inviting the reader to extract more information than is actually present in your plot.
Bryan
The binning is 10 GeV. Not sure if it matters, the important distribution is Mgg. How much can you draw from 4 events?
---Ken Johns---
1)are the efficiencies of the background sample the same as for the signal sample?
Yes, they are. The background and signal sample have the same kinematic dependance, thus the same efficiency. Since, the efficiency is the same, I am normalizing their cross sections.
2)when you do subtraction of the direct photon backgrounds do you take into account all efficiencies (e.g. trigger)?
Yes, I do.
3)in the sentence (the efficiency for events to pass the photon...using z to e+e-). how do you get the efficiency for tracking photons using e+e- events? i see how for the others you can say photons and electrons are the same but not for tracking. i mean it could be the effic. for tracking electrons is 1 and the efficiency for tracking photons is 1 (if you had a really noisy system).
Recall, that I construct "virtual" photons within Z to ee events. I constructed them at the perpindicular bisector point between the two parent electrons. Therefore, I could calculate the tracking quantities for a high pT event (i.e. have the proper underlying event noise, etc.).
4)you say that the expected background agrees with the observed number of events. isn't that by construction? i mean you normalized the background sample to your observed sample (for m(gg) < 60). so i don't see why this "agreement" is noteworthy. you set it up that way.
Quick note, I normalize my QCD multijet back. to the number events with a Mgg mass less than 60 MINUS the direct photon content.
N(jjgg-back,m(gg)<60) = N(signal sample,m(gg)<60) - N(direct photon,m(gg)<60)
I normalize the background for the loose or initial-selection event requirements. I then use this same normalization factor for my tight requirements. By doing this, I see that my two samples, signal and background behave similarly through these requirements, which is noteworthy. Since the two samples scale similarly through the optimization process, the two samples are representative of each other, i.e. the background sample is a good representative of my signal sample.
5)when you give the number of qcd and direct photon events should you say over what mass range that is? i assume it's over the full mass range.
I state in the sentence, "The total number of QCD multijet events exp..."
6)can you please check (again) that the theory curve in fig 3. is correct?
Yes, I just quadruple-checked it for Jay Wightman...All is well.
7)why do the exclusion contours fall so precipitously between 60 and 70 gev/c**2 and then flatten out? i think i asked this once before but don't remember the answer.
Because a few observed events fall in limit calculation for the lower mass range.
8)you said in your note to stefan that the acceptance for the fusion channel is lower than for associated channels. is this corrected for in the exclusion contours? it's probably a small effect if it's not.
The acceptance is lower, and I do correct for this in the final state cross section plot.
9)why is the luminosity error singled out? why not just put that in the list of systematic errors?
Well, the uncertainty in integrated luminosity is not associated with the uncertainty in signal efficiency. The luminosity uncertainty is treated seperately in the limit calculation as well.
1)i am not thrilled with the abstract. do you want to say a search for a boson higgs in the channel pp to hw to ggjj has been carried out for the first sentence?
Not really, because I do include pp to Hjj to ggjj events, i.e. fusion. So to make the abstract simple, and precise I think it is best to leave it in the pp to ggjj form.
2)the second sentence i don't think is correct english. maybe with the requirement of ... or with selection criteria of
corrected
3)shouldn't standard model be Standard Model everywhere? i don't know but i seem to always see it capped.
You are correct, I will make sure that all are capped.
4)i don't know what a non-standard higgs is? isn't that slang? maybe say we refer to such higgs with suppressed couplings to fermions as are the ;'s correct in the bottom of the first column? could/should commas be used instead?
I will add the definition of non-standard in the following sentence.
The semi-colons are correct, it is an itemized list.
5)p3. are the selection criteria for 2 photons and 2 jets for at least 2 photons and at least 2 jets.
You are correct, at least two....I am not sure if I should put this in though....it might add some confusion.
6)p3. shouldn't one have et> 30 gev and |eta| < ... it's a list with 2 objects so needs an and not a comma?
fixed.
7)p4. is dataset a word? or is it data set...
not sure here.
8)p4. do you have to reference the calculated direct photon cross section or does that just come out of pythia?
The direct photon cross sections are used as a consistency check, even- though I use Pythia for all of my calculations.
9)p4. i would write "the direct photon background...section. The dominant systematic... (i.e. get rid of the ; (clearly ; scare me).
I removed the semi-colons.
just j