Dear MG5 team, I'm just playing with LO MLM matched MC samples and I noticed that there are some events with negative weight. Process: generate p p > t t~ add process p p > t t~ j add process p p > t t~ j j Additiona command: set run_card pdlabel lhapdf set run_card lhaid 261000 set run_card xqcut 80.0 set run_card sys_pdf 261000 I found two events having negative weights among 10000 events. Here is one of lhe output: 6 3 -6.4592046e+02 3.01567200e+02 7.54677100e-03 9.99998300e-02 21 -1 0 0 502 503 +0.0000000000e+00 +0.0000000000e+00 +7.5275943115e+02 7.5275943115e+02 0.0000000000e+00 0.0000e+00 1.0000e+00 -1 -1 0 0 0 501 -0.0000000000e+00 -0.0000000000e+00 -3.2013906455e+03 3.2013906455e+03 0.0000000000e+00 0.0000e+00 -1.0000e+00 6 1 1 2 502 0 +1.1403921787e+02 -1.3429936222e+02 +2.5588744069e+02 3.5559618138e+02 1.7300000000e+02 0.0000e+00 1.0000e+00 -6 1 1 2 0 501 -3.1363459834e+02 +2.9217926105e+02 +2.9444846113e+02 5.4805499553e+02 1.7300000000e+02 0.0000e+00 1.0000e+00 21 1 1 2 504 503 +8.3675008823e+01 -1.2766593668e+02 -2.1348620803e+02 2.6244325007e+02 0.0000000000e+00 0.0000e+00 -1.0000e+00 -1 1 1 2 0 504 +1.1592037164e+02 -3.0213962140e+01 -2.7854809082e+03 2.7880556497e+03 0.0000000000e+00 0.0000e+00 -1.0000e+00 2 0.30156721E+03 2 0.11979322E+03 0.20877878E+03 1 21 0.11580914E+00 0.30156721E+03 1 -1 0.49252165E+00 0.11979322E+03 -0.40072237E-05 2 6 1 -1 2 5 1 -1 2 4 1 -1 1 3 2 -1 -1.1139163e+03 -5.8886622e+02 -5.7381283e+02 -7.8855949e+02 -3.5521572e+02 -9.0882155e+02 -4.8044391e+02 -4.6816215e+02 -6.4336956e+02 -2.8981324e+02 -7.5616239e+02 -3.9974142e+02 -3.8952269e+02 -5.3529966e+02 -2.4113190e+02 -9.5241617e+02 -5.1341283e+02 -5.0062564e+02 -6.8225836e+02 -3.1366583e+02 -7.7705688e+02 -4.1888303e+02 -4.0845022e+02 -5.5664064e+02 -2.5591354e+02 -6.4653087e+02 -3.4852122e+02 -3.3984086e+02 -4.6313901e+02 -2.1292650e+02 -8.2203756e+02 -4.5103622e+02 -4.4006512e+02 -5.9523944e+02 -2.7873550e+02 -6.7068364e+02 -3.6799123e+02 -3.5904013e+02 -4.8564369e+02 -2.2741459e+02 -5.5802565e+02 -3.0617796e+02 -2.9873042e+02 -4.0406776e+02 -1.8921466e+02 -9.2593138e+02 -5.0183402e+02 -4.8954850e+02 -6.6321731e+02 -3.0889167e+02 -7.5544848e+02 -4.0943612e+02 -3.9941261e+02 -5.4110544e+02 -2.5201840e+02 -6.2855213e+02 -3.4066115e+02 -3.3232134e+02 -4.5021334e+02 -2.0968564e+02 -7.9168605e+02 -4.3753236e+02 -4.2710884e+02 -5.7381283e+02 -2.7276035e+02 -3.5697370e+02 -3.4846937e+02 -4.6816215e+02 -2.2253959e+02 -5.3742207e+02 -2.9701110e+02 -2.8993528e+02 -3.8952269e+02 -1.8515854e+02 -6.8331019e+02 -3.8437477e+02 -3.7544162e+02 -5.0062564e+02 -2.4238532e+02 -5.5749881e+02 -3.1360351e+02 -3.0631514e+02 -4.0845022e+02 -1.9775722e+02 -4.6385303e+02 -2.6092601e+02 -2.5486190e+02 -3.3984086e+02 -1.6453898e+02 -7.8227600e+02 -4.3293762e+02 -4.2273455e+02 -5.6582394e+02 -2.7115133e+02 -6.3824299e+02 -3.5322495e+02 -3.4490047e+02 -4.6164418e+02 -2.2122682e+02 -5.3103422e+02 -2.9389204e+02 -2.8696587e+02 -3.8409957e+02 -1.8406628e+02 -6.6885842e+02 -3.7746388e+02 -3.6881670e+02 -4.8954850e+02 -2.3943453e+02 -5.4570790e+02 -3.0796506e+02 -3.0091000e+02 -3.9941261e+02 -1.9534973e+02 -4.5404270e+02 -2.5623468e+02 -2.5036469e+02 -3.3232134e+02 -1.6253589e+02 -5.7729673e+02 -3.3160426e+02 -3.2420106e+02 -4.2710884e+02 -2.1277071e+02 -4.7100460e+02 -2.7054913e+02 -2.6450902e+02 -3.4846937e+02 -1.7359527e+02 -3.9188767e+02 -2.2510368e+02 -2.2007816e+02 -2.8993528e+02 -1.4443563e+02 -6.4592046e+02 +4.8737257e+02 -5.0323189e+02 -1.6386290e+03 -1.5989304e+03 -9.2676826e+02 -2.8853283e+02 -1.0062295e+03 -1.6506252e+03 +1.7255055e+02 -9.0984388e+01 +4.0993576e+02 -6.0562481e+02 -8.4948641e+02 +3.1883291e+02 -1.1049902e+03 -1.1959257e+03 -4.3249898e+01 -3.7486750e+02 -2.9306541e+03 -1.2718238e+03 -2.2608081e+02 +5.0891409e+01 -5.1265770e+02 +2.8941238e+02 -6.0211538e+02 +3.1028151e+02 -6.8215556e+02 +1.4104334e+02 -4.7916982e+02 -1.0559768e+03 -1.7099236e+03 -1.0442194e+03 -1.9237848e+03 -4.4079193e+02 -7.6346762e+02 -1.4291308e+03 -6.8483737e+02 -4.1995218e+02 -2.2661533e+02 +2.8852919e+02 -2.6579120e+02 -6.3252097e+02 -1.0238816e+03 -1.5416082e+03 -4.2291052e+02 -2.6101943e+02 -5.5902035e+01 -3.6772813e+02 +4.6481857e+02 -6.6669643e+02 -8.0543815e+02 -4.9155148e+02 +8.9879205e+02 +3.6226317e+02 -1.2256650e+03 -2.3195069e+03 +2.3768053e+02 +3.9529091e+02 -1.8996454e+03 -1.8186350e+03 -3.8857898e+02 -2.1777700e+03 -4.5997124e+02 -1.3749764e+03 -5.2982787e+02 -3.3017525e+02 -4.8711169e+02 -5.0578035e+02 -7.2565567e+02 +9.2447450e+01 -5.5588809e+02 -3.7479926e+01 -2.0569868e+03 -1.1435902e+03 -1.8443050e+03 -6.6486324e+02 -2.1077821e+03 -1.3276187e+03 -6.7330735e+02 -3.2154356e+02 +2.3119580e+01 -3.0857749e+02 -1.3801350e+02 +6.6368143e+02 -8.9433767e+01 -1.4253144e+02 -7.3266384e+02 -1.2742273e+03 -1.1878992e+03 -8.7450714e+02 -7.5166062e+02 +1.5398289e+01 -3.4708738e+02 -9.0113730e+02 -6.3348298e+02 -8.0245297e+02 -6.6128191e+02 -9.4782979e+02 -1.8999447e+02 -4.6293795e+02 I roughly have idea that unweighted LO sample have positive definite weights. The only possiblity that I can imagine is that such events are located at matching boundary and it can be vetoed or not vetoed depending on PDF replica and scale choice. === Results Summary for run: run_01 tag: tag_1 === Cross-section : 645.4 +- 1.327 pb Nb of events : 10000 Pythia8 merged cross-sections are: > Merging scale = 105 : 445.17 +/- 0.67 [pb] > Merging scale = 157.5 : 454.28 +/- 0.68 [pb] > Merging scale = 210 : 449.82 +/- 0.67 [pb] Notice that because Systematics computation is turned on, the merging did not veto events but modified their weights instead. The resulting hepmc/stdhep file should therefore be use with those weights. However, if some of events having weight zero then it looks okay but still sign inversion somewhat hard to understand at a glance. May I know what's happening behind scene? Have a nice day! Best wishes, Sung Hak Lim

When	Who	Action	New State
2017-06-19	Olivier Mattelaer	Answer	Answered
2017-06-19	Sung Hak Lim	Comment	Answered
2017-06-19	Olivier Mattelaer	Answer	Answered
2017-06-19	Sung Hak Lim	Confirm	Solved
2017-06-19	Sung Hak Lim	Confirm	Solved

MadGraph5_aMC@NLO

History of question #644186