CDF/ANAL/TOP/CDFR/2525 October 8, 1993 COMPARISON OF KINEMATICAL ANALYSES FOR B-TAG CANDIDATES (preliminary, incomplete, but, I think, quite informative). K. Sliwa Tufts University For every event you'll find the KS printout (cone 0.7, px,py,px,et,E-jet energy corrected for underlying event+out of cone correction+scale (cor1); all four energy correction factors (cor4 means scale correction but no underlying event and no out-of-cone correction); jet "mass"; phi and eta). Below you'll find Lina's or Brian's printout with corresponding information about jet with cone 0.4 and full corrections (what would correspond to cor1 in KS printout). What follows is the summary of fits; the jet assignments are listed in the following convention : e lb lh W1 W1. Results are listed in the following order (left to right - Lina,Brian,Kuni,Dolly and KS, where: i) (full) means using full method with structure function, angular and Pt(t-tbar) factors turned ON, and with jet correction cor4 ii) cor1 means full method but with cor1, i.e. including out-of-cone and underlying event correction iii) plain means method with structure function, angular and Pt(t-tbar) factors turned OFF, and with jet correction cor4. To the rightmost you'll fine the values of the structure function, angular and Pt(t-tbar) factors. At the bottom you'll find my (KS) comments. I am writing a summary of this comparison now, and will distribute it as soon as it is ready. Here are my most important observations: a) the mass values are basically in very good agreement between ALL methods of fitting, as long as one uses the same jet configuration as input. b) Two events were found to yield very different sets of jets, depending on whether clustering was done with cone 0.4 or 0.7. I think one should reject those events, it might actually be a good procedure to run jet clustering with cone sizes of 0.4 and 0.7 to verify stability of jet clustering. Kinematical analysis rely on how well the energies are measured, if clustering is not stable I would reject those events (46510_139604 and 46818_221912). c) In one event the difference between Lina and KS is related to square root ambiguity when finding the longitudinal component of the neutrino. The two solutions tend to give very similar mass values in most cases, but in 40758_44414 the difference is about 10 GeV ! In KS method the solution is solved geometrically, there is no ambiguity. It seems that KS's solution corresponds to the other neutrino solution, at least in Lina's case. From my past experience (E691, kinematical analysis of D->Kenu, where I had to find longitudinal neutrino component using the square root, we found that choosing the one which gives LOWER MASS was closer to reality !) d) In event 45705_54765 the are actually two solutions of very similar quality (chi2 or L), both with b-tagged jets as b-jets in kinematical configuration. The difference in mass is 25 GeV !!! KS picks the lower one. This is not because of using the structure functions, angular distribution or Pt(ttbar) factors, the plain method finds the same result. KS would argue that the fitting method based on scanning explicitly the fine grid in the entire parameter space is more reliable. In any event, the effect found in this event constitutes a new systematic error. e) The three dynamical factors used by KS do not seem to change the mass of the solutions. What is affected (but in most cases slighly, and in NONE of the events in any decisive way) is the value of L, which estimates how likely a particular combination is due to ttbar production and decay. This is expected, and KS would argue that this information should be included since it reflects our knowledge of physics. e*) IN EVENT 45879_123158 the dynamical factors actually make a difference. f) The mass shift due to using/not using the out-of-cone and underlying event corrections in KS method is about +2%, as expected from a study based on a larger sample of events, and mentioned in cdf2256. ============================================================================= -------------KS printout RUN : 42548 /EVENT: 143286 # px py pz et en q ELE: 1 -12.60 63.77 -29.03 65.00 71.19 1.00 ELE: 2 -5.05 -10.28 -3.93 11.46 12.11 1.00 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta * JET: 2 -31.16 -59.39 -41.45 67.06 82.85 1.28 1.20 1.30 1.23 25.46 4.23 -0.50 JET: 3 -38.88 -2.73 -27.74 38.97 48.91 1.34 1.24 1.38 1.28 10.18 3.21 -0.60 JET: 4 25.26 -9.38 75.24 26.95 80.48 1.26 1.14 1.31 1.20 9.51 5.93 1.75 JET: 5 10.64 3.45 -24.54 11.19 27.24 1.62 1.37 1.80 1.53 3.85 0.31 -1.43 JET: 6 -8.27 -3.75 25.49 9.08 27.25 1.56 1.27 1.76 1.46 3.24 3.57 1.74 MUO: 1 -2.31 -1.10 -1.98 2.56 3.24 MUO: 2 -0.90 -2.05 -0.42 2.24 2.28 MET: 628.91 230.77 46.13 6.00 75.32 0.38 1.58 7.38 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 2 3 5 4 mass 150+-6 153+-6 149+-6 5.9 0.34 1.5 chi2/DL/L 0.37 0.14 0.28 comments: KS finds a good solution (L~0.4) with jet 5 which has corrected energy Et>10 GeV. Others may have not tried because Et>8 GeV cut on uncorrected energy. Jet 2 has a soft lepton tag, it is a leptonic b according to this kinematic analysis. There is very little difference in mass values obtained by KS with full/plain method, the value of L changes, but not that much. Including the out-of-cone and underlying event correction shifts the mass by 1.8%. ================================================================================ -------------KS printout RUN : 45705 /EVENT: 54765 # px py pz et en q ELE: 1 11.78 51.23 39.86 52.57 65.97 -1.00 ELE: 2 1.78 -12.58 10.36 12.71 16.40 1.00 ELE: 3 3.71 -12.07 8.75 12.63 15.36 1.00 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -81.23 -53.68 88.37 97.37 133.41 1.30 1.24 1.32 1.25 22.56 3.73 0.89 JET: 3 -45.94 2.37 -10.11 46.00 49.62 1.36 1.26 1.39 1.29 15.63 3.09 -0.09 *JET: 4 11.81 -44.70 33.74 46.24 57.87 1.34 1.25 1.38 1.28 8.52 4.97 0.78 JET: 5 24.76 -6.58 88.68 25.62 92.69 1.29 1.17 1.35 1.23 8.39 6.02 2.03 JET: 6 -3.53 10.04 -4.99 10.65 12.27 1.60 1.34 1.78 1.51 3.50 1.91 -0.32 JET: 7 -7.18 5.63 -18.28 9.13 20.58 1.90 1.55 2.14 1.78 2.40 2.48 -1.32 MUO: 1 -1.62 -0.19 -2.29 1.63 2.81 -------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 5 1 45705 54765 0.000 0.00 jet index in KS list: 1 3 4 5 input pt e b1 b2 j1 j2 55.79 96.77 51.71 47.78 26.41 Mj1j2 60.1 eta e b1 b2 j1 j2 0.70 0.81 -0.23 0.68 1.94 phi e b1 b2 j1 j2 1.34 3.73 3.06 4.98 6.00 rawet e b1 b2 j1 j2 0.00 73.98 36.57 33.86 17.38 metx mety 79.2 31.5 tagged jet, eta phi rawet 0.68 4.98 33.86 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 4 5 1 3 mass 171+-13 167+-8 168 173+-7 167+-5 170+-5 167+-5 2.9 0.39 1.5 chi2/DL/L 1.3 7.2 7.9 5.5 0.16 0.07 0.05 e 5 4 1 3 mass 145+-11 141+-6 144+-7 140+-6 2.2 0.41 1.1 chi2/DL/L 1.5 0.17 0.09 0.07 154+-6 +Brian mentions this solution with b-tagged 0.5 jet as one of W jets comments: The jets found with cones 0.4 and 0.7 agree quite well. There are two close fits in both Lina's and KS, which have b-tagged jet as one of b jets in the kinematics. In Lina's case the chi2 is marginally larger for the e4513 solution, while in KS e5413 is better. The difference is not due to any of three additional factors in KS method (structure functions, angular distribution and Pt of ttbar system). The difference may come from different way of minimizing, scanning over a fine grid in parameter space, I would argue, is more reliable, it is independent of any assumptions about the minimized function. The decisive factor here is how the chi2/likelihood(L) is defined and how the minimi- zation is done. There is very little difference in mass values obtained by KS with full/plain method, the value of L changes, but not that much. Including the out-of-cone and underlying event correction shifts the mass by 1.8%/+2.1%. ================================================================================ -------------KS printout RUN : 45880 /EVENT: 31838 # px py pz et en q ELE: 1 2.05 4.61 2.71 5.05 5.73 -1.00 ELE: 2 3.30 9.88 2.38 10.42 10.69 1.00 ELE: 3 -6.26 -26.60 4.21 27.33 27.65 -1.00 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta ?JET: 1 101.69 -7.32 -20.61 101.95 104.30 1.21 1.16 1.23 1.17 7.73 6.21 -0.28 JET: 2 -25.33 -53.90 77.80 59.55 98.42 1.46 1.37 1.49 1.40 9.30 4.27 1.02 *JET: 4 11.55 28.18 12.39 30.45 33.91 1.32 1.21 1.37 1.26 8.30 1.18 0.31 JET: 5 -22.02 -14.41 -14.77 26.31 31.30 1.40 1.27 1.46 1.33 8.31 3.72 -0.57 JET: 6 11.35 -12.21 4.62 16.68 18.22 1.45 1.28 1.56 1.38 5.72 5.46 0.18 JET: 7 -7.68 9.61 30.80 12.30 33.56 1.45 1.23 1.59 1.37 5.08 2.24 1.51 JET: 8 1.31 11.59 -9.32 11.67 15.25 1.65 1.39 1.81 1.55 3.07 1.46 -0.77 JET: 9 10.01 6.02 -16.82 11.68 21.13 1.97 1.67 2.18 1.86 5.21 0.54 -1.16 JET: 10 -4.17 9.35 -21.79 10.24 24.21 1.59 1.32 1.77 1.50 2.50 1.99 -1.54 JET: 11 -5.86 -3.24 -26.20 6.69 27.17 1.90 1.46 2.24 1.78 2.60 3.65 -2.07 -------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 6 1 45880 31838 0.000 0.00 jet index in KS list: 1 2 4 5 input pt e b1 b2 j1 j2 25.76 106.92 60.90 30.79 24.34 Mj1j2 60.5 eta e b1 b2 j1 j2 0.16 -0.20 1.09 0.38 -0.59 phi e b1 b2 j1 j2 4.48 6.21 4.28 1.15 3.70 rawet e b1 b2 j1 j2 0.00 84.19 39.65 20.78 15.86 metx mety -53.7 58.2 tagged jet, eta phi rawet 0.38 1.15 20.78 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 2 5 1 4 mass 145+-12 144+7 141+-8 145+-8 139+-7 5.4 0.41 1.5 chi2/DL/L 0.4 0.6 0.06 0.08 0.02 e 5 4 1 2 mass 137+-8 124+-4 126+-9? chi2/DL/L 8.2 7.8 8.5? 1xe-18 e 4 6 2 5 mass 110 100+-6 chi2/DL/L 14.0 very bad MET match e 1 2 4 5 mass 143+-10 143+-6 chi2/DL/L 2.9 0.004 e 4 1 5 7 mass 128+-6 131+-6 130+-6 6.9 0.40 1.0 1.04 0.36 0.13 comments: The jets found with cones 0.4 and 0.7 agree quite well. Brian rejects e2514 thinking jet 1 is tagged. He likes e5412 more, because of b-tag. Kuni finds e4625 as the best combination. Notice that jet 1 is not considered a part of the t-tbar system. Dolly (most likely) finds the same solution as Brian, thinking jet 1 is tagged, and perhaps rejecting e2514. According to Lina, it is jet 4 that is tagged, and she picks e2514. KS allows in his fits jets with Et>10 corrected, he finds e4157, with large L. If one were to reject this combination, say, by requiring Et jets >8 GeV uncorrected, then e2514 will be the combination chosen (except that it has low L<0.1). There is very little difference in mass values obtained by KS with full/plain method, the value of L changes, but not that much. Including the out-of-cone and underlying event correction shifts the mass by 1-2%. In this event the factor due to structure function is bigger, but results of full vs plain KS method do not change. ================================================================================ -------------KS printout RUN : 40758 /EVENT: 44414 # px py pz et en q ELE: 1 10.92 2.51 -1.24 11.21 11.28 ELE: 2 -11.52 21.90 8.32 24.75 26.11 ELE: 3 -96.22 -51.67 49.85 109.21 120.05 ELE: 4 11.66 -7.88 27.65 14.07 31.02 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -118.97 -65.89 59.46 136.00 148.60 1.18 1.13 1.19 1.14 7.21 3.65 0.35 *JET: 2 87.60 27.47 -21.63 91.81 94.86 1.23 1.16 1.24 1.18 10.06 0.30 -0.30 JET: 3 -26.07 72.54 23.45 77.08 82.73 1.23 1.16 1.24 1.18 18.81 1.92 0.22 JET: 4 46.05 48.22 44.52 66.68 80.54 1.26 1.19 1.28 1.21 7.68 0.81 0.56 *JET: 5 22.79 -16.99 58.10 28.42 64.95 1.35 1.23 1.40 1.28 6.00 5.64 1.38 JET: 6 5.92 6.25 -13.11 8.61 15.86 2.09 1.69 2.38 1.96 2.32 0.81 -1.23 -------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 2 3 40758 44414 0.000 0.00 jet index in KS list: 2 5 3 4 input pt e b1 b2 j1 j2 106.85 95.12 31.41 83.44 71.44 Mj1j2 86.2 eta e b1 b2 j1 j2 0.44 -0.24 1.46 0.30 0.62 phi e b1 b2 j1 j2 3.63 0.30 5.66 1.91 0.80 rawet e b1 b2 j1 j2 0.00 74.01 20.19 64.08 51.93 metx mety -37.4 -91.6 tagged jet, eta phi rawet -0.24 0.30 74.00 ca jp csq mt dmt pnul pnut 0 1 10.8 174.74 11.16 18.66 56.60 0 2 10.8 174.74 11.16 18.64 56.61 1 1 0.3 177.34 16.63 1.07 93.51 1 2 1.1 165.34 12.18 87.19 85.45 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 5 2 3 4 mass 177+-11 176+-8 168 177+-9 166+-8 166+-8 chi2/DL/L 0.3 1.8 18.2 2.4 0.070 0.084 2-nd neutr. mass(2) 165+-12 * 169+-8 172+-8 170+-8 chi2/DL/L 1.1 * 0.128 0.097 0.161 1.4 0.36 1.5 * ( sum of e5234+e5324 ) e 5 3 2 4 mass 177+-12 178+-7 173+-8 174+-8 chi2/DL/L 0.1 0.7 0.058 0.077 2-nd neutr. mass(2) 167+-12 chi2/DL/L 1.0 comments: The jets found with cones 0.4 and 0.7 agree quite well. In Lina's fits I found something interesting. In both cases listed above the two solutions obtained with two neutrinos (there is a quadratic ambiguity in Lina's way of fitting, Kuni also solves for a neutrino with square root) are very close to each other in terms of Chi2, while they differ in mass by about 10 GeV. The solution found by KS is very much similar to the 2-nd neutrino solution. The difference may come from different way of minimizing, scanning over a fine grid in parameter space, I would argue, should be more reliable, it is independent of any assumptions about the minimized function. This may explain why Kuni and KS agree. The fact the quadratic ambiguity may lead to different masses may, potentially, be a source of a systematic error in least squares methods. (In KS method the solution is reached geometrically, there is no quadratic ambiguity.) The results are still within errors for all methods, including KS. There is very little difference in mass values obtained by KS with full/plain method, the value of L changed. Including the out-of-cone and underlying event correction shifts the mass by +1.7%. ================================================================================ RUN : 41838 /EVENT: 168601 # px py pz et en q ELE: 1 -26.66 37.95 5.90 46.38 46.75 -1.00 ELE: 2 8.36 -9.10 9.71 12.35 15.72 -1.00 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -29.03 -71.38 117.33 77.05 141.24 1.44 1.36 1.46 1.38 15.68 4.33 1.04 JET: 2 -34.27 46.61 25.95 57.86 63.73 1.26 1.19 1.29 1.21 6.39 2.20 0.20 JET: 3 -9.76 3.92 -41.28 10.51 42.77 1.75 1.46 1.95 1.65 3.82 2.76 -2.25 JET: 4 6.89 -2.49 -3.80 7.33 8.91 1.89 1.48 2.20 1.78 3.35 5.94 -0.67 JET: 5 -1.60 -5.00 -35.88 5.25 36.33 1.81 1.30 2.22 1.69 2.25 4.40 -2.63 JET: 6 4.62 -0.79 -126.91 4.69 127.00 2.13 1.47 2.67 1.98 0.94 6.11 -4.05 ================================================================================ -------------KS printout RUN : 43096 /EVENT: 47223 # px py pz et en q ELE: 1 23.71 23.06 -29.68 33.07 44.44 -1.00 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta *JET: 1 78.08 101.81 3.11 128.30 128.87 1.26 1.20 1.27 1.21 11.72 0.92 0.08 JET: 2 -67.93 28.67 133.78 73.73 153.27 1.24 1.17 1.26 1.19 12.63 2.74 1.39 JET: 3 12.31 -72.42 86.84 73.46 114.97 1.43 1.35 1.46 1.38 16.79 4.88 1.02 JET: 5 -32.39 -13.49 129.58 35.08 134.41 1.27 1.17 1.31 1.21 6.67 3.54 2.05 JET: 6 -2.63 5.09 4.87 5.73 7.87 2.19 1.61 2.64 2.05 2.33 2.05 0.77 JET: 7 -4.96 -1.85 -42.97 5.29 43.34 1.77 1.27 2.17 1.65 1.90 3.50 -2.71 -------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 1 1 43096 47223 0.000 0.00 jet index in KS list: 1 2 3 5 input pt e b1 b2 j1 j2 30.35 137.03 79.40 72.95 38.53 Mj1j2 86.8 eta e b1 b2 j1 j2 -0.81 0.02 1.35 1.02 2.02 phi e b1 b2 j1 j2 0.77 0.91 2.74 4.88 3.53 rawet e b1 b2 j1 j2 0.00 101.42 57.22 47.43 26.50 metx mety -1.8 -77.6 tagged jet, eta phi rawet 0.02 0.91 101.40 ca jp csq mt dmt pnul pnut 0 1 1.2 167.07 15.51 1.46 58.85 0 2 1.2 167.11 15.51 1.39 58.91 1 1 0.6 285.12 15.15 -60.38 66.62 1 2 0.6 285.20 15.07 -60.92 66.53 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 1 2 3 5 mass 167+-16 171+-9 176 168+-8 158+-5 160+-5 158+-5 chi2/DL/L 1.2 1.9 4.4 0.003 0.0003 0.009 e 2 1 3 5 285+-15 ? chi2/DL/L 0.6 comments: The jets found with cones 0.4 and 0.7 agree quite well. Lina says that the best solution (e2 135) was rejected by everyone as unphysical (mt>280 GeV). The results of the second best are very close for all methods, including KS. There is no difference in mass values obtained by KS with full/plain method, only the value of L changed. Including the out-of-cone and underlying event correction shifts the mass by +1.3%. ================================================================================ -------------KS printout RUN : 45879 /EVENT: 123158 # px py pz et en q MUO: 1 46.45 4.20 -10.16 46.64 47.73 MUO: 2 6.73 -11.66 -1.49 13.46 13.54 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -75.50 27.92 -197.60 80.50 213.82 1.18 1.12 1.20 1.13 13.91 2.79 -1.66 *JET: 2 37.05 -66.53 -7.89 76.15 77.28 1.26 1.19 1.27 1.21 10.58 5.22 -0.15 JET: 3 35.86 25.14 4.73 43.79 45.75 1.40 1.30 1.43 1.34 12.36 0.61 0.05 JET: 4 -33.94 -3.89 -16.21 34.16 38.03 1.35 1.24 1.39 1.29 4.09 3.26 -0.50 JET: 5 2.87 25.33 -41.79 25.50 50.03 1.63 1.48 1.70 1.55 10.31 1.46 -1.25 ELE: 1 6.16 4.63 0.28 7.71 7.71 ELE: 2 13.64 9.73 2.22 16.76 16.90 ELE: 3 -24.53 -1.11 -7.55 24.56 25.69 -------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 6 1 45879 123158 0.000 0.00 jet index in KS list: 1 2 3 4 input pt e b1 b2 j1 j2 52.75 83.68 83.16 43.26 37.55 Mj1j2 82.1 eta e b1 b2 j1 j2 -0.21 -1.63 -0.11 0.13 -0.45 phi e b1 b2 j1 j2 0.09 2.79 -1.07 0.66 -3.02 rawet e b1 b2 j1 j2 0.00 69.44 62.06 28.80 25.90 metx mety -51.4 -10.4 ca jp csq mt dmt pnul pnut 0 1 8.9 165.60 9.90 -44.60 33.28 0 2 9.0 165.51 9.84 -48.53 31.95 1 1 11.4 178.35 12.69 -17.50 42.82 1 2 11.4 178.35 12.69 -17.51 42.82 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 2 1 3 4 mass 160+-12 165+-5 187+-8 189+-8 1.8 0.39 1.4 chi2/DL/L 6.0 5.0 0.01 0.03 e 1 2 3 4 mass 165+-10 190 166+-8 165+-8 2.0 0.035 2.2 chi2/DL/L 8.9 64.1 .4e-6 .6e-5 e 5 2 3 4 mass 133+-6 134+-6 17.0 0.43 0.7 chi2/DL/L 0.02 0.006 e 4 2 3 5 mass 133+-6 133+-6 17.0 0.18 0.7 chi2/DL/L 0.03 0.02 comments: The jets found with cones 0.4 and 0.7 agree quite well. This event does not give good fits for any of the method. There is no difference in mass values obtained by KS with full/plain method, only the value of L changed. The difference between KS and Lina may stem from the fact that KS considers all 5 jets (they all have UNCORRECTED Et>10-15), while Lina only look at the first four. KS finds solutions with 5-th jet. Those solution are assigned higher likelihood than the one found by Lina only when the three additional "physics" factors in the definition if likelihood are used. KS would argue that this is PHYSICS, ans that it should be included. ================================================================================ RUN : 42517 /EVENT: 44047 # px py pz et en q MUO: 1 38.89 -24.31 44.80 45.86 64.11 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -52.26 -16.01 20.08 54.65 64.70 1.25 1.18 1.28 1.20 28.22 3.44 0.37 JET: 2 -11.15 -26.19 -22.62 28.46 37.29 1.38 1.27 1.44 1.32 8.30 4.31 -0.66 JET: 3 -10.66 2.13 -59.77 10.87 60.83 1.93 1.63 2.13 1.82 3.18 2.94 -2.33 JET: 4 7.11 -0.27 -22.03 7.12 23.31 1.73 1.37 2.01 1.63 2.70 6.24 -1.74 JET: 6 -5.55 2.04 -6.87 5.91 9.22 2.26 1.71 2.69 2.12 1.73 2.79 -0.92 ELE: 1 -5.35 0.76 2.50 5.41 5.96 ELE: 2 -2.20 -7.42 -5.41 7.74 9.44 ELE: 3 -2.87 -4.95 -4.27 5.72 7.14 Everyone agrees: two jet event. ================================================================================ -------------KS printout RUN : 43351 /EVENT: 266423 # px py pz et en q MUO: 1 -0.61 2.05 1.06 2.14 2.39 MUO: 2 23.23 -1.08 -1.61 23.26 23.32 MUO: 3 -6.66 1.25 -0.25 6.78 6.78 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 116.08 -76.09 212.45 138.79 255.64 1.37 1.32 1.39 1.33 30.88 5.70 1.19 *JET: 2 -84.81 5.76 -14.24 85.01 87.03 1.24 1.18 1.26 1.19 12.02 3.07 -0.17 JET: 3 25.52 -19.31 6.97 32.00 35.05 1.32 1.22 1.37 1.26 12.50 5.64 0.20 ?*JET: 4 -3.47 24.95 15.60 25.19 31.25 1.37 1.24 1.43 1.30 9.94 1.71 0.54 JET: 5 -7.61 12.08 -6.88 14.28 16.72 1.50 1.31 1.62 1.42 5.34 2.13 -0.44 JET: 6 1.39 6.88 36.41 7.02 37.14 2.16 1.70 2.50 2.03 2.14 1.37 2.30 JET: 7 -3.96 -5.35 32.61 6.65 33.31 2.19 1.70 2.55 2.05 1.24 4.08 2.27 JET: 8 -0.13 -4.77 -0.01 4.78 5.07 2.51 1.79 3.10 2.34 1.70 4.69 -0.01 ELE: 1 -1.73 4.79 -3.47 5.09 6.16 ELE: 2 6.86 -7.22 1.36 9.96 10.05 -------------Lina's printout npass = 0 , ev igo ru ev max_d pznu 2 1 43351 266423 0.000 0.00 jet index in KS list: 1 2 3 4 input pt e b1 b2 j1 j2 24.55 145.28 90.73 31.49 19.40 Mj1j2 47.5 eta e b1 b2 j1 j2 -0.07 1.18 -0.18 0.23 0.38 phi e b1 b2 j1 j2 -0.05 -0.60 3.07 -0.73 1.67 rawet e b1 b2 j1 j2 0.00 99.79 68.82 21.97 11.92 metx mety -86.3 66.1 combination item Lina(LBL) Brian Kuni Dolly (full) (cor1) plain F(x1x2) plep Pt(ttb) e 2 4 1 3 mass 148+-9 145+-5 151+-8 152+-8 1.9 0.35 1.4 chi2/DL/L 5.3 5.3 0.06 0.08 139+-7 Dolly says that this is for a combination 2.7 with one tag, I don't know which on it is. comments: The jets found with cones 0.4 and 0.7 agree quite well. Lina, Brian, KS masses agree for e2413. This combination has both jets tagged. Dolly does not accept e2413. It is not a very good solution for KS, either, still this is the best fit in this event. There is no difference in mass values obtained by KS with full/plain method, only the value of L changed (decreased a little). ================================================================================ -------------KS printout RUN : 46818 /EVENT: 221912 # px py pz et en q MUO: 1 -35.20 35.06 60.00 49.68 77.90 MUO: 2 -10.27 2.16 0.66 10.50 10.52 MUO: 3 -6.78 0.97 0.03 6.84 6.85 MUO: 4 -3.93 0.66 0.26 3.98 3.99 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 46.66 -56.29 36.85 73.11 88.64 1.23 1.16 1.25 1.18 33.97 5.40 0.37 JET: 2 -51.66 12.65 -0.06 53.19 55.06 1.35 1.26 1.38 1.29 14.23 2.90 -0.08 JET: 3 7.25 4.51 -3.87 8.54 9.98 1.72 1.40 1.94 1.62 3.42 0.56 -0.49 JET: 4 5.38 7.32 13.82 9.09 16.81 1.96 1.62 2.20 1.85 3.00 0.94 1.10 ELE: 1 5.28 -12.80 3.72 13.85 14.34 ELE: 2 13.19 -8.82 15.04 15.87 21.86 Brian's printout Fit dump: RUN 46818 EVENT 221912 K,mtop,chi2: 1 85.843 252.74 Input 4-vectors: Muon : -34.911 34.771 59.519 77.267 Jet 1 : -53.673 10.133 -1.2997 55.244 Jet 2 : 14.317 -37.292 10.739 42.175 Jet 3 : 27.873 -19.355 30.849 46.176 Jet 4 : 17.498 -13.960 4.4557 23.480 Jet 5 : 9.2752 10.186 .00000E+00 13.777 comments: Results of jet clustering algorithm differ significantly between cone of 0.7 and 0.4. 0.7 has 2 jets, while 0.4 has four. I would reject this event, since jet clustering so unstable. This is perhaps the reason why results disagree. ================================================================================ RUN : 45047 /EVENT: 104393 # px py pz et en q MUO: 1 -20.31 40.76 -16.64 45.54 48.48 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -33.65 63.35 -105.37 71.73 128.44 1.58 1.49 1.60 1.51 15.81 2.06 -1.13 JET: 2 -9.85 -36.62 16.26 37.92 41.40 1.30 1.21 1.34 1.24 3.43 4.45 0.45 JET: 3 -17.99 32.04 40.63 36.74 55.76 1.46 1.35 1.51 1.40 10.42 2.08 0.95 JET: 4 -2.03 -11.20 26.90 11.38 29.37 1.43 1.22 1.57 1.35 3.14 4.53 1.59 JET: 5 -9.95 -0.97 -9.60 10.00 14.08 1.71 1.43 1.90 1.61 2.46 3.24 -0.81 ELE: 1 -7.30 -23.19 10.63 24.31 26.54 ================================================================================ RUN : 45178 /EVENT: 382599 # px py pz et en q MUO: 1 30.32 16.57 -5.62 34.56 35.01 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -72.38 -16.60 120.98 74.26 143.13 1.39 1.31 1.41 1.33 18.32 3.37 1.25 JET: 2 51.98 -5.22 31.92 52.25 62.92 1.28 1.20 1.31 1.23 14.52 6.18 0.57 JET: 3 21.03 -23.08 -34.97 31.23 47.36 1.53 1.40 1.58 1.45 6.71 5.45 -0.94 JET: 4 -10.91 18.20 28.47 21.22 36.53 1.63 1.46 1.71 1.54 8.58 2.11 1.05 JET: 5 -5.86 9.87 -16.17 11.48 20.13 1.97 1.68 2.17 1.86 3.47 2.11 -1.10 JET: 6 2.54 4.07 3.99 4.80 6.45 2.35 1.68 2.90 2.19 1.64 1.01 0.73 ================================================================================ -------------KS printout RUN : 45610 /EVENT: 139604 # px py pz et en q MUO: 1 35.95 7.62 -6.85 36.75 37.38 # px py pz et en cor1 cor2 cor3 cor4 mass phi eta JET: 1 -13.08 -108.74 -169.11 109.53 207.18 1.43 1.36 1.44 1.38 48.26 4.59 -1.22 JET: 2 9.57 75.37 -59.15 75.98 97.96 1.29 1.22 1.31 1.24 18.00 1.44 -0.77 JET: 3 -24.41 -19.69 0.41 31.36 34.07 1.40 1.29 1.45 1.34 13.30 3.82 -0.08 JET: 4 -8.93 14.46 51.27 17.00 54.53 1.32 1.17 1.41 1.25 7.48 2.12 1.66 JET: 5 -6.05 5.31 -1.61 8.05 8.87 1.72 1.39 1.96 1.62 3.36 2.42 -0.28 JET: 6 -12.60 1.05 -14.66 12.64 19.96 1.85 1.59 2.01 1.75 4.87 3.06 -1.02 JET: 7 -2.32 -8.13 10.60 8.45 13.93 1.87 1.53 2.12 1.77 3.19 4.43 0.92 JET: 8 4.28 -4.53 49.35 6.23 49.79 1.61 1.23 1.89 1.51 2.18 5.47 2.67 JET: 10 4.91 -3.02 -34.17 5.76 34.73 2.36 1.77 2.81 2.21 2.34 5.73 -2.47 JET: 11 3.56 2.44 -22.91 4.32 23.34 3.29 2.26 4.13 3.06 1.00 0.60 -2.41 ELE: 1 -5.82 -6.02 -0.17 8.37 8.37 Brian's printout Fit dump: RUN 45610 EVENT 139604 K,mtop,chi2: 1 126.75 123.28 Input 4-vectors: Muon : 52.757 11.376 -9.7915 54.851 Jet 1 : 11.720 79.159 -60.639 101.53 Jet 2 : -14.255 -73.253 -76.257 107.46 Jet 3 : 3.3430 -36.010 -78.011 86.241 Jet 4 : -17.880 -8.4600 1.4366 19.937 Jet 5 : -8.6712 -12.157 -1.4890 15.134 Jet 6 : -8.1918 11.043 48.131 50.159 Jet 7 : -17.921 -4.9838 .00000E+00 18.601 ------------Lina's printout npass ge 1, ev igo ru ev max_d pznu 5 1 45610 139604 0.000 0.00 input pt e b1 b2 j1 j2 53.48 80.02 74.63 36.16 19.78 Mj1j2 56.6 eta e b1 b2 j1 j2 -0.19 -0.70 -0.90 -1.51 0.07 phi e b1 b2 j1 j2 0.21 1.42 -1.76 -1.48 -2.70 rawet e b1 b2 j1 j2 0.00 58.89 50.91 27.03 10.84 metx mety -31.4 26.7 comments: Results of jet clustering algorithm differ significantly between cone of 0.7 and 0.4. I would reject this event, since jet clustering so unstable. This is perhaps the reason why results disagree.