From: SMTP%"pekka@cepheid.physics.utoronto.ca" 15-MAR-1998 13:28:07.27 To: TOBACK CC: Subj: Toronto Comments on PRD Draft of "Searches for New Physics..." Resent-Date: Sun, 15 Mar 1998 13:28:05 -0600 (CST) Resent-From: toback@FNALD.FNAL.GOV Resent-To: toback@uccdf.uchicago.edu From: Pekka K Sinervo Message-Id: <199803151927.TAA20744@cepheid.physics.utoronto.ca> Subject: Toronto Comments on PRD Draft of "Searches for New Physics..." To: toback@fnald.fnal.gov, frisch@fnalp.fnal.gov (Henry Frisch), mdshapiro@fnalp.fnal.gov (Marjorie Shapiro), yagil@fnalp.fnal.gov (Avi Yagil), kuhlmann@fnalp.fnal.gov (kuhlmann), rebcdf@fnalp.fnal.gov (blair) Date: Sun, 15 Mar 1998 14:27:56 -0500 (EST) Cc: pekka@cepheid.physics.utoronto.ca, cdf_toronto@physics.utoronto.ca (U of T CDF Group) X-Mailer: ELM [version 2.5 PL0b1] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear all- The Toronto group has reviewed the first draft of the PRD draft titled "Searches for New Particles in Diphoton Events...". We would first like to congratulate the authors on assembling such a comprehensive report on the studies prompted by the eegammagammaMEt event. We believe that this report is an important addition to the literature surrounding this event. ** Thank you. As this is a first draft, we will try to contain our comments to physics issues, and overall organization and presentation. We did not identify any major physics issues, given that this is following up on a blessed PRL. However, we found the draft to suffer from some organizational and presentation problems. ** This paper has been substantially revised. It was clear from the comments that a major revision was necessary, mostly in removing excess detail, references as well as making the text read more evenly and in the PRD style. Cheers, Pekka (for the Toronto group) General Comments ---------------- We found that the paper, though generally well organized, was uneven in its treatment of the various subjects. In particular, we believe that the level of detail presented on the one possibly anomalous event was too great for the typical (even well-informed) PRD reader to absorb or appreciate. Sections 4.8 and 4.9 are the culprits here, and we recommend a fairly heavy editing of these sections to reduce the level of detail that is not relevant to the issues of the events interpretation. ** We agree. This section has been substantially revised and greatly shortened. As one example, we note that the discussion of the SVX reconstruction is simply too detailed. A straightforward presentation of the efficacy of the reconstruction algorithm and the expected matching of SVX stubs with PEM clusters would suffice. ** We agree. This has been done. We also noted that the style of the draft alternated between reading like an internal analysis note, a typical PRD article and a Scientific American piece (an example of the latter is the last paragraph in the concluding section where we note "The eeggMEt event is tantalizing. Perhaps it is a hint of physics beyond the SM. Then again, it may just be one of the rare SM events that could show up in 10**12 interactions. Only more data will tell." We recommend we stick to a style more consistent with a PRD journal article. ** We have made the style more uniform and more consistent with a PRD article. We found that the analysis employed techniques or variables that in some places were not typical of other CDF analyses. In those cases, more description of the quantity and the motivation for it should be provided. Examples of these include the quantity sigma_CES, 3x3 trigger tower array and energy out-of-time. ** We have tried to include more about these variables, but found that doing so in more than a cursory manner would take away from the flow of the text. We hope that the text is better now. The paper also introduced symbols and variables without sufficient description or reference. An immediate example is in the second paragraph where N_1 and N_2 are introduced without any mention of what they represent. ** This has been fixed. Now all symbols come after the SUSY references. We did not explicitly define each symbol as the reader can infer them from the references. We found that the limits that were set in the paper were rather model-dependent and very specific to certain scenarios. Although attempts were made to give concrete results, the actual values quoted in the draft leave the reader with the sense that there is much more parameter space to explore. We wonder if a more model-independent method could be found to state our results (have no great ideas). ** Of course there is plenty of parameter space to be explored. This is SUSY. This is actually a fairly model independent analysis. There is only one parameter one could release which is mu. However, this is constrained if one assumes the MSSM (i.e, no other particles than defined in the MSSM). Even if one doesn't assume that, it doesn't change the results too much. These are the results the vast majority of the theory community want, and are most directly comparable to those of DZero and LEP. We also note that the the manuscript makes heavy use of references as a substitute for footnotes. We believe that this makes the document difficult to read, and suggest that information, where pertinent to the discussion, should be included in the prose, and references be limited primarily to referring to other publications.. ** We agree. This has been changed. The number of references has been reduced by almost 50%. Specific Comments: ------------------ Abstract: We start out badly by introducing missing Et in the first sentence. We suggest you use the term "weakly interacting or non-interacting particles. In addition, we should talk about candidates and not particles in describing the one event. ** This abstract is virtually identical to that in the PRL which the entire collaboration signed off on. We have left it as is. Introduction: The introduction needs to be carefully edited so that a reader is not immediately lost by the introduction of various new symbols representing SUSY particles. This section would be improved significantly if a general description of SUSY phenomenology was provided. ** We disagree. We have pointed the reader to the literature in a vary complete manner. This is not a SUSY paper, so we do not want to mislead the reader into thinking it is. We simply do not want to discuss general Technicolor and extended gauge groups here as well. CDF Detector: The description of the detector needs to be carefully checked for accuracy. For example, the performance figures that characterize the tracking and calorimeter resolution appear to be values specific to Run 1A and not 1B. We note that the standard selection criteria cited on Page 4 were largely developed for W, Z and QCD physics and the top group adopted them subsequently. We find the characterization of the end-plug as a "pole piece" to be imaginative but not accurate. There are many variables introduced here but not defined (eg. z_vertex). ** This section has been substantially shortened with most of the things you refer to having been removed. We think should take care of your objections. Data Selection and Photon Identification: Some of the content is better placed in the detector description section (eg, bottom of first paragraph, description of CES). Some of this is also repetitive with earlier descriptions. ** We agree. Some has been moved to the detector description and is no longer repeated. We note here an extensive use of references as footnotes. We find that this style is distracting, and in particular with a PRD unnecessary since we can say what we need to say in the article without referring to some other publication (at least we can give brief descriptions to make the manuscript more self-contained). We also note that the use of a reference to refer to another reference seems to be taking the concept of linking to the extreme ([32] is the most over-the-wall example of this). ** Again we agree. We have removed most of the references and made the text tighter. The variable sigma_CES has to be better defined, as does the use of the term trigger tower. An assertion in 2.3.1 is that above 22 GeV, photons leak enough energy into the hadron calorimeter to make the isolation cut inefficient. This is certainly not true. The trigger descriptions in 2.4 need to be made less confusing. ** We hope you find the new definitions and descriptions better. However, we should point out that the statement about the isolation cut is most certainly true. Most people don't realize it, but box isolation includes the hadronic tower behind the electromagnetic tower. Thus for high Et it becomes the equivalent to a Had/EM cut. For Et>50 the cut is more restrictive that Had/EM<0.05 and getting worse. In section 2.6.1, the discussion concerning the vertex efficiency is redundant. It is recognized that the data measurement does not yield an accurate efficiency measurement, so we in fact rely entirely upon MC to make this measurement. Couching it in the way it is done leads the careless reader to the conclusions that we are somehow using the data measurement! We suggest removing the discussion of the data measurement altogether. ** We have fixed this section. We believe it now satisfies your objection. In 2.6.2, we justify using the data because the we do not simulate energy out-of-time in the MC? We suggest looking for a better motivation for the measurement technique. In 2.7, we describe our triggers as "unusual" in CDF. This is largely a content-free statement to the non-CDF reader. We suggest just stating what are the characteristics of the trigger. If you want to admit to the fact that we did not sharpen the trigger threshold in L3, that's fine, but we don't think that is something all that noteworthy. Uncertainty on the high trigger threshold? What about the systematic uncertainties, which don't appear to be mentioned? ** We agree. All these comments have been taken into account in the text. In 2.8, we believe the discussion of the technique used to determine the purity should be summarized and not just referred to. The fact that the sample appears to be dominated by sources other than real di-photon events should be elaborated upon. ** This was discussed in great gory detail in the PRD referenced. We believe the results should simply be summarized. Searches for Anomalous Events: We recommend a careful editing of this section to identify the introduction of new variables and make sure appropriate definitions are provided. Is the missing Et resolution consistent with what we have stated in other analyses? The meaning of "intrinsic MEt" needs to be clarified. ** Done. The relationship in equation (10) needs to be properly discussed (it is not sufficient to characterize it as "having been known for some time"). It was originally introduced theoretically (Giele, we think), and then first confirmed by D0 in their top analysis of Run 1A data. We subsequently confirmed it ourselves. ** We do not find the discussion particularly troubling. This is well covered ground in QCD and we are not making a QCD measurement. To give this special treatment would make the text uneven. We also note that the structure chosen for this section makes it somewhat choppy. We talk about our data for one particular signature, and then discuss the SM sources of background. It might read better if we identify the general characteristics of the events we characterize as anomalous, based on the predicted SM rates, and then look in the data. Does this make sense? ** We have fixed the unevenness of the text. Thanks. The discussion in the first paragraph of 3.5 regarding b quark tagging needs a good rewrite. It should be somewhat independent from our b-tagging discussions in the top paper, to the extent that we have a number of algorithms on record and we should be clear as to which one we employ. ** Again, this is a standard algorithm well documented in the literature. We are not going to repeat it here. The eeggMEt Event: The 11 subsections of this section are very poorly balanced, with far too much detail in 4.8 and 4.9. Many of the earlier subsections can be compressed into a smaller number focusing on the characteristics of the event. This section also generates over 20 references, almost all of which are parenthetical comments that should either be incorporated into the text or just dropped altogether. ** You are correct. This has been greatly fixed. It seems that the philosophy hear is to do a literal core-dump of the event, hoping to convince the reader that everything up to Section 4.8 is pristine. We think that the reader will either believe our interpretation of the "standard" features in the event or not, and the detail written in this manner is not convincing. We particularly note that the tau interpretation does not yield a specific probability ("conservatively a few per cent"). If a reasonable estimate is in the percentish range, it is not clear one can even be so confident about ruling it out. ** Again, we agree. These sections have been greatly reduced. Table 22 and 23 seem to be overkill. Given that we have provided four vectors, of what benefit is the list of all possible transverse mass combinations? Finally, we note that section 4.9.6 and 4.11 are very repetitive. ** We think these tables are important. People have used them over and over again to discuss various mass combinations. We have decided to leave them as is. We have cleaned up the the repetition. Estimating the Number of ...": This section is quite convoluted and filled with intermediate calculations and estimates. It could be condensed significantly if one removed most of the discussion of the subsidiary event rates, taking a global approach to the problem by identifying a few key background processes, estimating their rates, and then summarizing the rest of the calculation. Otherwise, the reader gets totally lost. ** We agree. This section has been cleaned up. Setting Limits...: This section attempts to quantify the type of limits that can be set assuming certain SUSY models. It would be nice to see how much more general these limits can be made. ** Sorry. This is the analysis we did. It is our job to report on it. Conclusion: Again, we note the colloquial style of the last paragraph and recommend that it be tightened up. ** Done. Thanks again, Dave Toback