Message-Id: <200007110033.JAA04772@neutrino.kek.jp> Subject: SciFi summary (Jun99-Jun00) Date: Tue, 11 Jul 2000 09:33:20 +0900 From: Makoto Sakuda <sakuda@neutrino.kek.jp> Hi, I am forwarding you an e-mail sent to Conveners. Please comment. Regards, M.Sakuda -------------------------- Dear Conveners, As announced by Ishida last Friday, SciFi summary is shown in http://neutrino.kek.jp/internal/work_in_progress/fgd/ (Scifi analysis). This shows 4 figures: 1. data2mc.ps Data/MC of scifi data from Jun99 to Jun00 ( See also http://neutrino.kek.jp/~sakuda/conveners/data-mc1.eps for comparison of Data/MC (scifi) with Data/MC (1kton/MUC).) 2. expected-rate-scifi.. All the numbers are contained in this table: CT, efficiency, Data/MC, SK-expected (3. reduc.detail.ps This shows the data reduction. A detailed info. ) 4. Stability.vetm.muc3d Upper figure is the rate for events containing a scifi-track matched with downstream veto-counter, Scifi*Veto. Lower figure is the rate for our official events (SciFi*MUC-3d), Scifi*veto*MUC. The rate at each period is shown. In addition, 1. Changes from 2000.06 Official plots will be explained at the convener meeting. 2. Typical distributions (E_mu, mu_angle and vertex) are shown. Regards, M.Sakuda ------- End of Forwarded Message To: conveners cc: ishii, kate, hayato, sakuda Subject: ICHEP and Nu2000 policy Date: Mon, 10 Jul 2000 10:24:35 +0900 From: Makoto Sakuda <sakuda@neutrino.kek.jp> Dear Conveners, I would like to comment on FCin and FCout events. FCin = Fully contained events in the fiducial volume (22.5kton). FCout = Fully contained events in outer volume. I show some concrete numbers and show: Even if we keep the policy at Nu2000 for ICHEP, ie even if we show all numbers for FCin and FCout, there is no inconsistency in the conclusion, shown at Nu2000 and at ICHEP. Let me summarize FCin and FCout events from Jun99 to Jun00. (There may be a small mistake in the numbers, which will not change the whole feature.) 1) FCin events - - - - - --------------- FCin Month CT*10^18 Data Expected R=Data/(No Osc) /Year (No oscillations) R errors Jun99 3.1 1 5.01 0.20 +0.35-0.13 Nov99 3.5 2 6.18 0.32 +0.37-0.21 Jan00 2.2 2 4.09 0.49 +0.55-0.31 Feb00 4.0 6 7.29 0.82 +0.46-0.31 Mar00 3.7 6 6.61 0.91 +0.51-0.34 May00 2.5 4 4.5 0.89 +0.63-0.38 Jun00 3.75 6 6.6 0.91 +0.51-0.34 - - - - - ------------------------------------------------------------ All 27 40.2 0.67+-0.15 This value R is 2.2sigma away from 1.0 (No oscillations). New data are consistent with the numbers shown in Nu2000 Conf. At Nu2000, R (FCin)= 17/29.2(+3.4-3.3)= 0.58+-0.19. 1) FCout events - - - - - --------------- FCout Month CT*10^18 Data Expected R=Data/(No Osc) /Year (No oscillations) R errors Jun99 3.1 2 2.3 Nov99 3.5 1 2.6 Jan00 2.2 1 1.7 Feb00 4.0 3 3.0 Mar00 3.7 2 2.8 May00 2.5 2 1.9 Jun00 3.75 6 2.8 - - - - - ------------------------------------------------------------ All 17 17.1 R = 0.99+-0.31 At Nu2000, R (FCout) = 9/12.4+-2.5 =0.73+-0.28. At Nu2000 (Jun99-Mar00), R (FCin)= 17/29.2(+3.4-3.3)= 0.58+-0.19 2.6sigma from 1.0 R (FCout) = 9/12.4+-2.5 =0.73+-0.28. 1.0sigma from 1.0 - - - - - -------------------------------------------------------------- If we sum two data together, we have R(sum)=0.62+-0.14, 2.7sigma from 1.0. The significance of R(sum) is the same as that of R(FCin). With new data (Jun99-Jun00), R (FCin) = 27/40.2 =0.67+-0.15 2.2sigma from 1.0 R (FCout) = 17/17.1 =0.99+-0.31 0sigma from 1.0 - - - - - -------------------------------------------------------------- If we sum two data together, we have R(sum)=0.73+-0.14, 1.9sigma from 1.0. The significance of R(sum) is the same as that of R(FCin). Even if we keep the same policy at ICHEP, we can maintain the same conclusion consistently, namely "Our data (FCin) disfavors no oscillations at 2sigma level." If people ask, "Why don't you add FCin and FCout together to obtain a better limit?", then, we answer, "Since FCout events have lager systematic error, adding together does not change the significance at all. We use FCout events as a consistency check. We will not use FCout events for spectral analysis since the outer region is not well calibrated." On the other hand, If we change the policy at ICHEP and retract all numbers except FCin, we may be asked more questions: "Are all numbers, shown at Nu2000 a month ago for FCout events, wrong? Why do you change the policy within a month? Is anything wrong with the experiment?" Summary: ------- Regarding options for FCout events, I add one option: (3) Keeping the same policy as Nu2000. As I showed above, we can maintain the conclusion, shown at Nu2000, for ICHEP consistently. Certainly, there are other options (1) and (2). I sent the following mail to conveners the day before yesterday. > After talking with Nishikawa_san, we decide that we should >discuss over the issue at the convener meeting. > If the systematic error in FCout events, +-20% at present, >is not well studied, > >(1) we do not give a systematic error to FCout >events and we use FCout events as a reference and >consistency check (KS plots, CT plot, T-diff plot, vertex distribution), > or >(2) we retract all numbers about FCout events >(No FCout numbers, No distributions). >There may be a choice, (1) or (2) or else. Regards, M.Sakuda ------------------
This archive was generated by hypermail 2b29 : Tue Jul 11 2000 - 09:33:27 JST