From jkrane@fnal.gov Wed May 30 13:18:14 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f4UIIE625934 for ; Wed, 30 May 2001 13:18:14 -0500 Received: from fnal.gov ([131.225.226.198]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GE500FHMUUDF2@smtp.fnal.gov> for strauss@phyast.nhn.ou.edu; Wed, 30 May 2001 13:18:13 -0500 (CDT) Date: Wed, 30 May 2001 13:19:03 -0500 From: John Krane Subject: Re: approval of 630 geV photon results for EPS/LP and for thecollaboration To: strauss@mail.nhn.ou.edu Cc: "Jianming Qian, Univ. of Michigan, 734 936 1033" , Marek Zielinski , chopra@fnal.gov, dhiman@fnal.gov, yasuda@fnal.gov, "V. Daniel Elvira" , John Womersley , John Krane Message-id: <3B153997.A40D28C0@fnal.gov> Organization: Iowa State University MIME-version: 1.0 X-Mailer: Mozilla 4.76 [en] (WinNT; U) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en References: Content-Length: 4691 Status: RO I'm working from the 2-column version. I'm disappointed by the presence of all the passive voice, but I assume it came from people with more influence or persistence than I have, so I'm leaving it alone this time. Excluding that, please find my comments appended below. I did not actually go through the values in the tables. There are others in the collaboration with that particular strength anyway. Good luck Mike, I don't need to see the PRL again; just incorporate what you will of the following. - John pg 4, col 1, last full para, "...while EM3 is segmented into 0.05x0.05 sections, The electromagnetic..." Better make that comma into a period I think. --- Next para, first sentece, change to "D0 uses a three-tiered triggering system." What you have now is a bit clumsy and what you describe is not specific to photons until the final step. --- second column, last full para, insert new sentence at beginning: "Candidates within the acceptance region must satisfy four selection criteria." --- same para, middle. Although appently a D0 favorite, I do not like the CC (EC) convention. That same sentence starts with a symbol, which seems odd. You might consider changing two sentences "The efficiency of these last three...beyond 20 GeV." to these: "Simulation of these cuts determines their efficiency as a function of Et^gamma. In the CC, we find epsilon_s = 0.60 at 8 GeV and 0.88 at 20 GeV, and in the EC we find 0.75 and 0.90 respectively." --- same para, "tracks in the road due to tracks from the underlying event." Change to "tracks in the road due to particles from the underlying event". Or even "charged particles". Next sentence, delete "charged". --- Page 5, column 1, first para, "The relatively small probability...[all the way to] ...as our discriminant." Change to "Photons have a small probability to shower in the calorimeter cryostat or first absorber plate, and thus tend to deposit very little energy in EM1. Sensitivity to a doubling of this energy can help distinguish a neutral meson's two daughter photons from a single photon. Our studies show our best discriminant takes the form log log log log..." -- second para, you mention Pythia here by name but prior references to MC were not named. We should mention Pythia earlier (or Herwig or isajet if that was used). Be consistent. same para, "Three categories of fully simulated events are generated; ..." You want a colon here, not the semicolon. -- second column, "GEANT simulation". We have mentioned a detector simulation before, so this needs to be mentioned back there, not here. If you mention it by name, you should probably reference it. (I think we have in the past used a Moriond talk by Jonckheere for this.) -- next para, you should reference HMCMLL as being part of Cernlib. Or perhaps better, you could reference "fit" and have the bibliography cite HMCMLL, part of Cernlib. If this is in fact reference [10] then it should be moved to the thing you are referencing; right now it looks like a reference to constraining parameters to lie between 0 and 1. -- page 6, column 2 first full para, last sentence, "...good agreement between..." change to "...the deviations are not statistically significant and there exists good agreement between..." -- last para, 2nd sentence "Although some discrepancies exist..." I think you say all this better with "Despite some descrepancy between data and the prediction at low Et, there is good... -- Figure 1: Why use [ and { as the braces instead of the mathematical convention of using ( and [? I suggest you remove that from the caption entirely and use "Distribution of the discriminant for photon purity, where E1 is in units of GeV. Points..." That caption is overlong anyway. -- Fig 5, the "diagonal terms of the covariance matrix" is pretty clumsy when you actually see it in print. How about, "shaded blocks indicate both bin width and the completely correlated uncertainty of each point." (Prediction: this shaded band will cause trouble in collaboration review. Your current phrasing is correct but clumsy, all attempts to change it will make it more elegant but wrong!) -- Table I, you don't actually have any columns labeled deltaU deltaC. They are delta-sigmaU, delta-sigmaC. Table II, do you really have to tell the reader (Ratio) (Theory)? I think not, myself. Also the column label thing. -- I hope the above changes shorten the document -- I think you are overlong. One remedy is to combine tables I and II. Use Et range, plotted Et, plotted xt, measured cross section, delta u, deltac, measured ratio, delta u deltac. You just might fit it in a single column. ------------------------------------------------------------------------ From strauss@phyast.nhn.ou.edu Thu May 31 09:18:55 2001 Return-Path: Received: from particle.nhn.ou.edu (particle.nhn.ou.edu [129.15.30.205]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f4VEHa623054; Thu, 31 May 2001 09:17:37 -0500 Received: by particle.nhn.ou.edu (8.10.2/8.10.2) id f4VEHZq02225; Thu, 31 May 2001 09:17:35 -0500 (CDT) Date: Thu, 31 May 2001 09:17:35 -0500 (CDT) From: Mike Strauss Message-Id: <200105311417.f4VEHZq02225@particle.nhn.ou.edu> To: strauss@mail.nhn.ou.edu, jkrane@fnal.gov Subject: Re: approval of 630 geV photon results for EPS/LP and for Cc: qianj@umich.edu, marek@d0mino.fnal.gov, chopra@fnal.gov, dhiman@fnal.gov, yasuda@fnal.gov, daniel@fnal.gov, womersley@fnal.gov, jkrane@fnal.gov Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-MD5: 0bcOT6X6+gq2Uj9bbOgi3w== Content-Length: 7183 Status: RO Below are my responses to John's comments. -Mike > I'm working from the 2-column version. I'm disappointed by the presence of all the passive voice, but I assume it came from people with more influence or persistence than I have, so I'm leaving it alone this time. Excluding that, please find my comments appended below. > > I did not actually go through the values in the tables. There are others in the collaboration with that particular strength anyway. Good luck Mike, I don't need to see the PRL again; just incorporate what you will of the following. > > - John > > -- > > pg 4, col 1, last full para, "...while EM3 is segmented into 0.05x0.05 sections, The electromagnetic..." Better make that comma into a period I think. Done > -- > > Next para, first sentece, change to "D0 uses a three-tiered triggering system." What you have now is a bit clumsy and what you describe is not specific to photons until the final step. I kept this as it was. Even though the trigger may be generic, I am describing how to trigger on photons. > -- > > second column, last full para, insert new sentence at beginning: "Candidates within the acceptance region must satisfy four selection criteria." Done > -- > > same para, middle. Although appently a D0 favorite, I do not like the CC (EC) convention. That same sentence starts with a symbol, which seems odd. You might consider changing two sentences > > "The efficiency of these last three...beyond 20 GeV." > > to these: "Simulation of these cuts determines their efficiency as a function of Et^gamma. In the CC, we find epsilon_s = 0.60 at 8 GeV and 0.88 at 20 GeV, and in the EC we find 0.75 and 0.90 respectively." Changed to " Monte Carlo simulation of the \D0 detector determines the efficiency for these three selection criteria, $\epsilon_s$, as a function of $E^{\gamma}_T$. We find $\epsilon_s \sim 60\% (75\%)$ in the CC (EC) at 8.0 GeV and $\epsilon_s \sim 88\% (90\%)$ above 20 GeV." I still use the CC (EC) convention (which I do like). > -- > > same para, "tracks in the road due to tracks from the underlying event." Change to "tracks in the road due to particles from the underlying event". Or even "charged particles". Next sentence, delete "charged". Done. (Changed to "charged particles." > -- > > Page 5, column 1, first para, "The relatively small probability...[all the way to] ...as our discriminant." Change to "Photons have a small probability to shower in the calorimeter cryostat or first absorber plate, and thus tend to deposit very little energy in EM1. Sensitivity to a doubling of > this energy can help distinguish a neutral meson's two daughter photons from a single photon. Our studies show our best discriminant takes the form log log log log..." Changed to: "Photons have a small probability of showering in the calorimeter cryostat or first absorber plate, and thus tend to deposit very little energy in EM1. Sensitivity to a doubling of this energy can be used to distinguish a neutral meson's two daughter photons from a single photon. We use the function $\log_{10}[1+\log_{10}\{1+E_1{\rm{(GeV)}}\}]$ as our discriminant to determine the single photon purity." > -- > > second para, you mention Pythia here by name but prior references to MC were not named. We should mention Pythia earlier (or Herwig or isajet if that was used). Be consistent. This is the first time I mention the generator level, so I mention Pythia here. > -- > same para, "Three categories of fully simulated events are generated; ..." > You want a colon here, not the semicolon. Done > -- > > second column, "GEANT simulation". We have mentioned a detector simulation before, so this needs to be mentioned back there, not here. If you mention it by name, you should probably reference it. (I think we have in the past used a Moriond talk by Jonckheere for this.) Changed to: "The detector response is modeled using a detailed simulation with the energy response in $E_1$ corrected to match the data from $W \rightarrow e \nu$ events." > -- > > next para, you should reference HMCMLL as being part of Cernlib. Or perhaps better, you could reference "fit" and have the bibliography cite HMCMLL, part of Cernlib. If this is in fact reference [10] then it should be moved to the thing you are referencing; right now it looks like a reference to > constraining parameters to lie between 0 and 1. Changed to: " The fit is performed as a function of $E_T^\gamma$ using the CERNLIB fitting package {\textsc{HMCMLL}}\cite{mcmll}, with the fractions of signal and background constrained to be between 0.0 and 1.0. > -- > > page 6, column 2 first full para, last sentence, "...good agreement between..." change to "...the deviations are not statistically significant and there exists good agreement between..." Changed to: "Although the lowest $x_T$ points are systematically higher than NLO QCD predictions in both the EC and the CC regions, the deviations are not statistically significant and there exists good agreement between good agreement between the measured ratio and theory." > -- > > last para, 2nd sentence "Although some discrepancies exist..." I think you say all this better with "Despite some descrepancy between data and the prediction at low Et, there is good... Done > -- > > Figure 1: Why use [ and { as the braces instead of the mathematical convention of using ( and [? I suggest you remove that from the caption entirely and use "Distribution of the discriminant for photon purity, where E1 is in units of GeV. Points..." That caption is overlong anyway. Changed to exactly what you suggest. > -- > > Fig 5, the "diagonal terms of the covariance matrix" is pretty clumsy when you actually see it in print. How about, "shaded blocks indicate both bin width and the completely correlated uncertainty of each point." (Prediction: this shaded band will cause trouble in collaboration review. Your > current phrasing is correct but clumsy, all attempts to change it will make it more elegant but wrong!) Changed to "The error bars indicate the uncorrelated uncertainty and the shaded bands indicate the correlated uncertainty." > -- > Table I, you don't actually have any columns labeled deltaU deltaC. They are delta-sigmaU, delta-sigmaC. Fixed to be delta-sigmaU and delta-sigmaC > Table II, do you really have to tell the reader (Ratio) (Theory)? I think not, myself. Also the column label thing. Deleted (Ratio) and (Theory) and fixed columns. > -- > > I hope the above changes shorten the document -- I think you are overlong. One remedy is to combine tables I and II. Use Et range, plotted Et, plotted xt, measured cross section, delta u, deltac, measured ratio, delta u deltac. You just might fit it in a single column. > Currently still two Tables. I may have to reduce the length of the document and, if so, I will come back to the EB. I'm currently thinking of deleting some of the remarks regarding cuts, etc. and refering to the published 1800 paper. I may be able to combine the two tables, as well. Thanks for your comments. -Mike ------------------------------------------------------------------ From linn@fnal.gov Fri Jun 1 11:07:46 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f51G7j620768 for ; Fri, 1 Jun 2001 11:07:45 -0500 Received: from habana-clued0.fnal.gov ([131.225.224.252]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GE900LFUE4W13@smtp.fnal.gov> for strauss@phyast.nhn.ou.edu; Fri, 01 Jun 2001 11:07:44 -0500 (CDT) Date: Fri, 01 Jun 2001 11:07:44 -0500 (CDT) From: "Stephan L. LINN" Subject: Re: Pythia and Underlying Event In-reply-to: <200105291341.f4TDfiu11610@particle.nhn.ou.edu> To: Mike Strauss Cc: linn@scri.fsu.edu Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-Length: 192 Status: RO Mike, The underlying event associated with the hard interaction was "on" in Pythia. Any additional interactions and calorimeter noise were overlayed using the zero bias events. OK? SLL ----------------------------------------------------------------------- From yasuda@fnal.gov Mon Jun 4 10:04:19 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f54F4H625082 for ; Mon, 4 Jun 2001 10:04:17 -0500 Received: from d0nt75 ([131.225.225.135]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with SMTP id <0GEE004R4V756Y@smtp.fnal.gov> for strauss@phyast.nhn.ou.edu; Mon, 04 Jun 2001 10:04:17 -0500 (CDT) Date: Mon, 04 Jun 2001 10:04:22 -0500 From: Taka Yasuda Subject: Re: approval of 630 geV photon results for EPS/LP and for the collaboration To: Marek Zielinski , chopra@fnal.gov, dhiman@fnal.gov, jkrane@fnal.gov, marek@fnal.gov, strauss@mail.nhn.ou.edu Cc: qianj@umich.edu, "V. Daniel Elvira" , John Womersley Message-id: <001001c0ed07$a32aa990$87e1e183@fnal.gov> MIME-version: 1.0 X-MIMEOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 X-Mailer: Microsoft Outlook Express 5.50.4133.2400 Content-type: text/plain; charset=iso-8859-1 Content-transfer-encoding: 7bit X-Priority: 3 X-MSMail-priority: Normal References: <200105292050.PAA11835@d0mino.fnal.gov> Status: RO Content-Length: 821 Hi, I am very happy to hear that now we are ready to submit this paper to EPS/LP. I have one concern regarding this paper. Figure.2 shows the purity for two different eta ranges. In the upper plot of Fig. 2, there is one point clearly off from the all others. This point has the smallest error, also. I wonder how well we undrstand this plot. A less significant question is that in the lower plot, with the exception of the last point, all the points line up on a straight line. This is statistically very unlikely, unless the error bars are inflated. It is also true for the upper plot with the exception of the point with the smallest error. If we post this fiugre in the paper, referees will likely to ask explanations for at least the first question. Are we ready to defend our paper convincingly? Regards, Taka ------------------------------------------------------------------------ From jkrane@fnal.gov Mon Jun 4 10:59:18 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f54FxG624138 for ; Mon, 4 Jun 2001 10:59:16 -0500 Received: from fnal.gov ([131.225.105.97]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GEE00ECLXQQ82@smtp.fnal.gov> for strauss@phyast.nhn.ou.edu; Mon, 04 Jun 2001 10:59:15 -0500 (CDT) Date: Mon, 04 Jun 2001 11:14:40 -0500 From: John Krane Subject: Taka's Q To: strauss@mail.nhn.ou.edu Message-id: <3B1BB3F0.1040106@fnal.gov> Organization: Iowa State University MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; m18) Gecko/20010131 Netscape6/6.01 Status: RO Content-Length: 255 Hi Mike, It occurs to me (without have actually looked to see...) that Taka's questions might be assuaged if the error band of the fits were included in these plots for PRL. Just an idea. - John -- http://www-d0.fnal.gov/~jkrane/John_Krane.html -------------------------------------------------------------------- From strauss@phyast.nhn.ou.edu Tue Jun 5 08:59:02 2001 Return-Path: Received: from particle.nhn.ou.edu (particle.nhn.ou.edu [129.15.30.205]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f55Dld620248; Tue, 5 Jun 2001 08:47:39 -0500 Received: (from strauss@localhost) by particle.nhn.ou.edu (8.10.2/8.10.2) id f55DlcI07235; Tue, 5 Jun 2001 08:47:38 -0500 (CDT) Date: Tue, 5 Jun 2001 08:47:38 -0500 (CDT) From: Mike Strauss Message-Id: <200106051347.f55DlcI07235@particle.nhn.ou.edu> To: chopra@fnal.gov, dhiman@fnal.gov, jkrane@fnal.gov, marek@d0mino.fnal.gov, marek@fnal.gov, strauss@mail.nhn.ou.edu, yasuda@fnal.gov Cc: daniel@fnal.gov, qianj@umich.edu, womersley@fnal.gov Subject: Re: approval of 630 geV photon results for EPS/LP and for the collaboration Status: RO Content-Length: 537 Taka, et.al. I don't know if you ever got your question answered or not. The errors on Fig. 2 are the errors from the purity fit. They are not simply statistical errors. This means that they are inflated compared to purely statistical errors. I agree that the one point in the CC with small error bars looks suspect, but in every analysis method I have used, I get the same qualitative feature. The caption to the plot could indicate that the errors are the errors on the fit and are larger than statistical errors alone. -Mike ------------------------------------------------------------------- From jkrane@fnal.gov Tue Jun 5 08:56:19 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f55DuJ616324 for ; Tue, 5 Jun 2001 08:56:19 -0500 Received: from fnal.gov ([131.225.105.101]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GEG0022WMPSZV@smtp.fnal.gov> for strauss@mail.nhn.ou.edu; Tue, 05 Jun 2001 08:56:18 -0500 (CDT) Date: Tue, 05 Jun 2001 09:11:46 -0500 From: John Krane Subject: Re: approval of 630 geV photon results for EPS/LP and for the collaboration To: Mike Strauss Message-id: <3B1CE8A2.5000403@fnal.gov> Organization: Iowa State University MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; m18) Gecko/20010131 Netscape6/6.01 References: <200106051347.f55DlcI07235@particle.nhn.ou.edu> Status: RO Content-Length: 799 Mike, Each point results from a fit to a distribution, right? So that usually means that the errors are *smaller* than pure statistics would indicate. - John Mike Strauss wrote: > Taka, et.al. > I don't know if you ever got your question answered or not. The > errors on Fig. 2 are the errors from the purity fit. They are not simply > statistical errors. This means that they are inflated compared to purely > statistical errors. I agree that the one point in the CC with small error > bars looks suspect, but in every analysis method I have used, I get the > same qualitative feature. The caption to the plot could indicate that > the errors are the errors on the fit and are larger than statistical > errors alone. > > -Mike ------------------------------------------------------------------------ From strauss@phyast.nhn.ou.edu Tue Jun 5 09:09:40 2001 Return-Path: Received: from particle.nhn.ou.edu (particle.nhn.ou.edu [129.15.30.205]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f55E9b617422 for ; Tue, 5 Jun 2001 09:09:37 -0500 Received: (from strauss@localhost) by particle.nhn.ou.edu (8.10.2/8.10.2) id f55E9aO07251 for strauss; Tue, 5 Jun 2001 09:09:36 -0500 (CDT) Date: Tue, 5 Jun 2001 09:09:36 -0500 (CDT) From: Mike Strauss Message-Id: <200106051409.f55E9aO07251@particle.nhn.ou.edu> To: strauss@mail.nhn.ou.edu Subject: Reply to Krane Status: RO Content-Length: 519 That is true, in general. However, here we are trying to determine the fraction of the distribution that comes from photons. The error is the error on the fraction. Suppose that the signal and background had identical shapes. Then no matter how much data you had (infinite statistics), the error on the fraction would still be 100%. Because the difference in signal and background is primarily in the tail of the distribution, the error here is actually larger than statistica alone. Does that make sense? -Mike ------------------------------------------------------------------------ From jkrane@fnal.gov Tue Jun 5 10:15:39 2001 Return-Path: Received: from fnal.gov (heffalump.fnal.gov [131.225.9.20]) by phyast.nhn.ou.edu (8.10.2/8.10.2) with ESMTP id f55FFd615664 for ; Tue, 5 Jun 2001 10:15:39 -0500 Received: from fnal.gov ([131.225.105.101]) by smtp.fnal.gov (PMDF V6.0-24 #37519) with ESMTP id <0GEG002N2QDYYW@smtp.fnal.gov> for strauss@mail.nhn.ou.edu; Tue, 05 Jun 2001 10:15:36 -0500 (CDT) Date: Tue, 05 Jun 2001 10:31:04 -0500 From: John Krane Subject: Re: approval of 630 geV photon results for EPS/LP and for the collaboration To: Mike Strauss Message-id: <3B1CFB38.30105@fnal.gov> Organization: Iowa State University MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; m18) Gecko/20010131 Netscape6/6.01 References: <200106051408.f55E8qH07248@particle.nhn.ou.edu> Status: RO Content-Length: 1357 Mike, yes this makes perfect sense. My difficulties arise because I'm used to fitting this set of points to a functional form and displaying the resulting error on this Fit (let's say "capitol F": Fit). This contrasts to the way you get each individual point and error bar, with a little-f fit. My knowledge that we had all agreed to no Fit, to instead use bin-by-bin information, got suppressed by a long-ingrained instict of mine to minimize uncertainties whenever possible. So when Taka asked about the error bars, I suggested showing the Fit errors. You replied, to my confusion, that the little-f fit errors were already displayed. My recent correspondence with you sort of went downhill from there! Sorry. - John Mike Strauss wrote: > That is true, in general. However, here we are trying to determine the > fraction of the distribution that comes from photons. The error is the > error on the fraction. Suppose that the signal and background had > identical shapes. Then no matter how much data you had (infinite > statistics), the error on the fraction would still be 100%. Because > the difference in signal and background is primarily in the tail of > the distribution, the error here is actually larger than statistica > alone. > > Does that make sense? > > -Mike