Short and Biased Minutes of

W Jamboree 10/04/2000 – PS Auditorium – 10 am to 17:30 pm

  1. TGCs from Single-Gamma (192-202) (A. Jacholkowska)

    2. Agnieska presented the final Single-Gamma analysis of the 1999 data, using Gary's selection. The couplings are measured to be DKg = -0.1 +- 0.5 and Lambda_g = -0.2 +- 0.8, which represent a relative improvement on the statistical error of 30% and 10% w.r.t. the values at 189 GeV. Systematic errors are being evaluated and should not be a problem to be ready for Summer conferences. On the other hand, Agnieska has also started a new analysis using a selection of two photons + missing energy, to put limits on the Quartic Gauge Boson Couplings (WWGG), with the Z's decaying into neutrinos. Work has started in this direction. It was requested on the meeting, if possible, to use Gary's selection (two photons + missing energy) to determine the QGC, or at least, to keep the changes w.r.t. Gary's selection to the minimum.

  2. Simulation systematic effects studies (A. Venturi)

    3. Andrea gave an extensive report on the studies he has been working on charged track acceptance. The basic question to answer is how well the MC reproduces the acceptance nearby the beam pipe hole. Andrea has developed some tools to check the simulation. Several (time-dependent) discrepancies between data and MC have been found, in particular on the costh distribution of the track closest to the beam pipe in a jet (differences between the acceptance in the 4th and 5th pad rows). The best would be to redo the W mass analysis (and any other analysis concerned) requiring 5 TPC hits for the good track definition, and see if this has any effect. Andrea had also studied the relation between fragmentation and beam hole. Any discrepancy between data and MC around the jet axis can create different energy and direction biases in jets close to the beam pipe. The first look at the problem may suggest that the effects are small, but also the idea to redo the W mass analysis using 5 TPC hits seems wise here.

    Conclusion: Redo W mass analysis using 5 TPC hits to define a good charged track.

    Andrea has also look at the possible "double counting" in the way we have estimated the detector systematics. Mixed classification of the systematic errors is a potential source of "double counting". Need a careful study and a "double counting" finder campaign! The posibility to improve the Calorimeters simulation was also adressed. The bottom line is: anyone interested in doing precision measurements should try to help in these areas... Andrea is looking for collaborators.

  3. Colour reconnection (M. Pearson)

    4. Matthew has update the multiplicity measurement with the 1999 data using the same method as Nick. NN selection is softer (cut at -0.9 in NN19) to increase the sensitivity to CR. There is a systematic trend to measure a positive difference between N(4q) - 2 N(2q), which goes in the opposite direction than most of the CR models used, with the notable exception of the HERWIG model. As the bkg is substracted from data, it would be nice to see the shape of the bkg, in particular in the soft momenta region. Also the effect of BE correlations in this measurement could be a potential problem to be quantified. It was sugested to update the draft of the paper with this new measurements, and circulate the 1st draft the sooner the better. New CR MC is needed for last year data.

    Conclusion: Update draft of paper with 1999 data.

  4. Looking for CR evidence in Particle and Energy Flow (M. Chalmers)

    5. Matthew has had a first look to the idea developed in L3 to look for CR effects in the energy and particle flow region between jets. The selection is the same than for the W mass (NN14 > 0.3), and a topological selection criteria is added. This reduces the 4q efficiency to around 30% with a purity of around 85% (78% correctly paired). This is not exactly the same kind of L3 selection (two times more events in Matthew's selection), because L3 restricts their selection to very planar events. After this selection, the rescaled angle phi_i = phi_i/phi_jk in the plane defined by jets j and k is computed, (i refers to an individual track). As events are not planar, the angle phi_i is transformed onto the four planes spanned by each pair of adjacent jets. The effect of CR is clearly seen in the interjet region, and has opposite sign in the region inter W's and in the region intra W's. Thus, an asymmetry is defined and significant differences between predictions for CR and non-CR models are observed for the first time. Data seems to lie somewhere in between the Non-CR model and SKI(with 100% reconnection probability), but compatible with both models. As the analysis has been only applied to 189 GeV data, and there is room for improvement the situation looks quite promising, and it's worth the effort to try to combine the results from the four LEP experiments.

  5. Particle Flow a la L3 (T. Ziegler)

    6. Thomas selection is much closer to L3 selection, therefore events are much more planar in this case. The 4q selection efficiency is around 14% and a similar purity (85%) than Matthew's selection (88% correctly paired). The rest of the analysis is very similar to the previous one. Thomas has compared the phi distribution with different CR ARIADNE models, and JETSET models. The conclusions are very similar than Matthew. A more global selection was tried too (in the line of Matthew's selection) and the results are very consistent.

  6. Particle Flow a la L3 (R. Jones)

    7. Roger has been investigating the effect of the different selections. It seems that the L3 type of selection removes a large fraction of events colour reconnected. He has also investigated new variables in the momentum space that could show more sensitivity. Work is in progress in this direction.

  7. W mass measurement in the tau nu qq channel (D. Boumediene)

    8. Djamel has started to work in the reconstruction of tau nu qq events. He has developed a specific tau jet finder which has a different treatment for one-prong and multi-prong tau decays. The performance of this new algorithm is better than the official analysis used so far. For instance, around 75% of the events have all tracks from the true tau reconstructed and not mixed with other tracks after selection (to be compared with 51% with the official analysis). The resolution of the rescaled hadronic mass after selection has an RMS of around 5.7 GeV (to be compared with 11.4 GeV with the official analysis). Looking in more detail, the performance of the new tau jet finder is particularly good for the one-prong tau->e,mu decays. A first attempt to measure the W mass using the rescaled hadronic mass (not using a full kinematic fit) has been performed. The expected statistical error is 238 MeV (to be compared with 278 MeV using the official analysis in the same conditions, i.e. not using a kinematic fit). With this method, Djamel determines the W mass to be 80.461 +- 0.265 GeV, (to be compared with 80.239 +- 0.308 GeV using the official analysis in the same conditions). Overall the work done by Djamel looks very promising, and some improvements are still possible.

  8. 4q calorimetric selection (F. Ligabue)

    9. Franco has finalized his 4q calorimetric selection and documented it in an ALEPH note (ALEPH 2000-033, PHYSICS 2000-007). The motivation was to have a complementary check to the official and "only charged tracks" selection in the measurement of the 4q cross-section. The basic idea is to drop the charged tracks in the event, re-run Eflow, force a 4-jet topology and build a linear discriminant analysis from a combination of "well-behaved" variables. The measured cross-sections at 189,196,200 and 202 GeV are in good agreement with the official values. The jet corrections have been recomputed for this selection, and found to slightly over-correct the High Energy side of the distribution. Similar problems to reproduce the bkg, as in the official analysis, are found.

  9. 4q fragmentation error (L. Mir)

    10. Lluisa has been working on the comparison JETSET vs HERWIG trying to understand the difference on MW (-29.8 +- 7.4 MeV). The difference between both MC models seems to be the same before and after selection. Lluisa has studied in detail posible sources of differences: Multiplicity of baryons, Y34 distribution, Pt out and b-tagging. Reweighting those distributions to agree between both MCs change the difference in the W mass to: -24.5, -26.1,-28.5 and -29.0 MeV respectively. Thus, the biggest effect is seen in the baryon multiplicity... work in this line is in progress, and is absolutely needed to understand the fragmentation error we are quoting.

  10. Evaluation of the BE systematic with KINAGAIN (H. Ruiz)

    11. Hugo has used the brand new produced BE MC using KINAGAIN to reevaluate the BE systematic error on the W mass. Using only 80k events (from the 100k) generated, he determines an error of +16 +- 15 MeV (to be compared to the value in the paper +27 +- 21 MeV using 100k MC events). Thus, the efforts invested in KINAGAIN to correlate both MC samples start producing interesting results. Moreover, the highest correlation between both MC samples allow to investigate the effect of cutting those events with pions from different W's nearby in momentum space. With the previous MC sample, no conclusion could be drawn due to statistical fluctuations, while with the KINAGAIN sample, a tendency to decrease the difference between both MC samples when removing these events maybe seen.

  11. Zgamma status report (B. Trocme)

    12. Benjamin gave a status report of the progress in the Zg studies. The analysis of the qqg events at all energies up to 202 GeV indicates a lower measurement of the LEP energy compared with the values from the LEP Energy calibration group. The effect is at the level of about 3 sigmas. A similar analysis has been performed using dimuon events and the same tendency is observed, although with less significance due to lower statistics (about 1.5 sigmas lower). There is a clear disagreement between data and MC for very collinear dimuon events in the s'/s distribution (data is much more spread in the region 0.lt.x.lt.0.005, being x = 1-s'/s, than MC). This could point out that the resolution for this very small acollinearities is not well reproduced in the MC. Also, could be interpreted as soft ISR not being reproduced by MC. Benjamin, worked out a toy model to study the second hypothesis, and found that by reweighting soft ISR in the radiative return events, he could shift the measurement to agree with the LEP energy calibration group. It was requested to check dimuon events at the Z peak, because if it is ISR the effect should be energy dependent. This has been done after the meeting, and found the same disagreement between data and MC, making the ISR hypothesis less likely. Benjamin also pointed out small problems with the calibration curve of the method, but it was recomended to correctly propagate the errors from the reference MC to achieve any conclusion. Benjamin also interfaced the fitting program to ZFITTER and found nice agreement. A lot of work is still going on to understand the difference between the Zg measurement and the LEP measurement. The decision on what to show next week at the workshop with the other LEP experiments was posponed to the general W meeting on wednesday.