From: SMTP%"MARTIN@hepmail.physics.yale.edu" 11-MAR-1998 16:37:48.84 To: TOBACK CC: Subj: comments on gamma-gamma PRD Date: Wed, 11 Mar 1998 17:41:16 -0500 (EST) From: MARTIN@hepmail.physics.yale.edu Subject: comments on gamma-gamma PRD To: TOBACK@UCCDFK.UCHICAGO.EDU, FRISCH@UCHEP.UCHICAGO.EDU, REB@HEP.ANL.GOV Cc: MDSHAPIRO@LBL.GOV, YAGIL@FNAL.GOV, STK@SGI6.HEP.ANL.GOV Message-id: <01IUJRAK8KAAD3YUNV@hepmail.physics.yale.edu> X-VMS-To: TOBACK@UCCDFK.UCHICAGO.EDU,FRISCH@UCHEP.UCHICAGO.EDU,REB@HEP.ANL.GOV X-VMS-Cc: MDSHAPIRO@LBL.GOV,YAGIL@FNAL.GOV,STK@SGI6.HEP.ANL.GOV,MARTIN MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Hi, I admit I haven't managed to read through the entire paper due to the volume of papers with comments due this week, however I have read enough of it to develop a few general comments. I think the physics is sound, however I have one question there. The analysis appears to have been thorough and attempts to deal with many possible theories as generically as possible. I haven't tried to include all of the usage and wording type comments, just those I feel stronger about. The comments follow below. ** Thanks. Andrew --------------------------------------------------------------------------- For the physics question, in section 2.6.3 I was concerned about the use of the PYTHIA Z->ee monte carlo in comparing with the ee sample with M_ee>30GeV/c^2. Do the data and monte carlo have the same mass distribution? If they are different, could it affect the efficiency and correction factor calculations? ** They are very similar. We have included the drell-yan component in the Monte Carlo. This is a very small effect and only affects the limits. The additional 5% systematic more than covers the problem and doesn't affect the limits. The biggest complaint I have with the paper has to do with the references. Many don't refer to anything at all but expand upon points in the text. I believe these should either be incorporated into the general text or placed as footnotes. There is at least one (91) which just refers the reader to a different reference. Why not eliminate the "middle-man" and just refer to ref. 18 directly? Additionally you should check the references carefully, on page 8 line 3 you give ref 33. I see no reference to a had/em cut in that paper (the CEM NIM article). Often you give a reference to another paper for something which is given in explicit detail here, I don't think those should even be there at all. One last complaint about the references, you tend to over quote them. Every time you mention a subject you don't need to reference it, just the first time. ** We agree. All these comments have been address and are much better in the new draft. I think the paper would stand on it's own much better if there was a section which described the standard cuts used to identify the various objects (ie. muons, electrons, b jets, ...). At this point you could make any necessary references (which shouldn't be many since you would then be defining the objects locally) once. This is the model used for example by the evidence for top PRD (50, 2966 (1994)). ** We disagree. This paper is already very long, and these cut are now standardized. This was not the case in the evidence PRD. There one was doing the standardizing. You need to make sure things occur in order. By this I mean definitions should occur the first time something appears in the body of the paper. For example, the coordinate system should be defined the first time it is needed (currently this occurs when you use |eta| in section 1.1) not at some random time later (in the coordinate system case the definition occurs when talking about the radii of the SVX layers). Also there is at least on ordering problem in the figures, figure 7 is referred to ahead of figures 4-6. ** These issues have been fixed as well. The following are more specific comments in the sections I managed to get through follow. Page 4, figure 3. The caption refers to the projective geometry of the calorimeter towers being shown, however I don't see any calorimeter towers in the figure. ** You are correct. That reference refers to an old figure which is no longer used. This has been fixed. Page 4, equation 2. I wondered why the P_T resolution is better for muons than what you quote on page 3 for generic charged tracks? (Also isn't there a multiple scattering term in the resolution?) ** This stuff has been removed. Page 5, the energy resolutions. Are these for generic jets or tau jets? It wasn't clear to me. ** Again, this stuff has been removed. page 6, 4th paragraph. You mention that there are 48 wedges in the calorimeter but neglect to mention that they each cover 15 degrees in phi. Without this the outside reader might think the towers are 7.5 degrees each. ** We now refer the reader to the width of a trigger tower earlier. page 7. Several places you redefine acronyms previously defined. ** Fixed. page 7, 4th paragraph. This is the figure 7 reference which should probably cause a figure reordering I mentioned earlier. ** Fixed. page 12. In the table you use a different notation for the monte carlo correction factors than in the text. You should try to use the same. ** This table has been removed. page 13, last line. You give the high threshold trigger efficiency as 100% with no error. There is at least a statistical error on this which is asymmetric and should appear. (According to my calculation it should read 100+0/-1.5 % given 121 out of 121 event.) ** You are correct, but we decided it simply makes things look more complicated. page 15, Figure 6. The plot is labeled as a ratio of efficiencies in the caption and as an efficiency in the actual figure. ** We think it is clear. pages 16-17, Figures 8-9. You show plots of the trigger efficiency as a function of photon E_T. However the points at or close to 0% and 100% appear to have no error bar even though points in neighboring bins have substantial error bars. Is this really correct, just a result of not knowing how to get PAW to plot asymmetric error bars, or not knowing how to calculate the correct error bars? ** The error bars are calculated correctly. The error get small for binomial statistics around 100% and 0%. We have made the asymetric errors symetric by picking the larger of the two. They are there, they are just small. Also, this is where most of the statistics are. pages 18-19. I had a hard time understanding what you were attempting to do with the missing E_T resolution. I think I figured out what you did, I'm not sure I understand the point of the exercise. You might want to say "up front" what you gain from doing this. ** We hope this is now clearer. Mostly we didn't open up the can of worms. Sorry for the confusion. page 19, 2nd paragraph. You claim to count jets the same as in the evidence for top PRD, however there the cuts are either E_T>10.0GeV & |eta|<2.4 in the dileptons or E_T>15.0GeV & |eta|<2.0 in lepton plus jets. Neither of these is what you use and I object to saying you use a method "identical" to that paper. ** The final dilepton jet cuts are Et>10 and |eta|<2.0. We also state them up-front. page 20, ref 44. Why are you using a thesis for the reference when we have a PRL on the topic? I think we would be much better served by at least including the PRL which is much more widely available. ** We have included the PRL. The thesis contains all the gory detail if the reader wants it. page 20, your ref. 45 (which shouldn't be a reference). I don't see the connection between the information in this "reference" and this paragraph. I think the paragraph stands fine without it. ** This reference has been removed. page 21, 1st paragraph. You claim to use the same isolation requirements as the top PRD. However there are 2 different types of isolation used there: calorimeter and tracking. Which did you use? ** We used both. page 22, Table 6. In the last sentence of the caption you should state the "electron emitted nearly all its energy". This would make it consistent with the text among other things. ** We have fixed this paragraph. pages 23-24, figures 13-14. Would it make sense to combine these two plots into one figure (two plots side by side)? Since they are essentially the same thing just using two different samples, that would seem reasonable. Similar to what you did with Figure 17. ** Done. pages 25-26, figures 15-16. Could these be made into one figure with 2 plots (perhaps side by side)? ** Done. page 27, ref 50. I really don't like this discussion at all. It looks to me as though we have no confidence in the analysis by claiming the e-e-gamma-gamma-met event is somehow not to be considered background because we don't like it in the background. It passes all the cuts applied and is thus a part of the sample even if it is highly unlikely to be background. ** We think that this is the fair thing to do, and we have great confidence in our analysis. We have a background estimate of from W's and Z's. This seems to describe the non-eeggmet events. The event does not look like the background, it simply isn't fair to compair it with the others. page 28, figures 18-19. Again could these plots be combined? ** Done. page 30, Table 10. I don't like the reference column in the table. It is completely unclear what the reference applies to. ** This table is virtually identical to that in the PRL. Most people found the references helpfull in identifying which cuts were used. At this point I skipped to the end and skimmed it as I was running out of time before the comments were due. If you like I will continue reading and send in more comments on the text. I think there should be a more definitive statement about SUSY in the conclusion. In the text we are able to set limits on portions of parameter space which are at least "competitive" to use your words. However in the conclusion you imply that we have some sensitivity to SUSY models and we can't conclude anything because of lack of data. These seem inconsistent to me and I prefer the stronger statement. ** We always welcome constructive critism. We hope you find the next draft better. On the SUSY issue. This, like almost all other searches, simply cannot rule out all of SUSY. We conclude what most people are able to conclude: SUSY isn't where we looked. That is what is said, and is what should be said. Thanks, Dave Toback