Short Biased Minutes of

W meeting 24/11/1999 in 40/SS-C01 from 9 am to 1.30 pm

1) lvlv mass status (Djamel)

Djamel presented results concerning the lvlv mass measurement using three observables, Emax_lepton, Emin_lepton, Missing energy. The analysis is performed using 189 GeV data. He has tried three fitting methods all based on reweighting. The first employs 1 MC at 80.35 GeV for the reference, similarly to what was done for the measurement at 183 GeV given at Moriond. This method suffers from important fluctuations due to the presence of sizeable weights when far from the 80.35 GeV point. The second fitting method uses several reference masses (actually 5) combined for the intermediate points. The weight of the various MC reference points in the combination is controlled by an exponential regulated by a parameter (see the transparencies for its definition). The fluctuations present in the first method are overcome, the calibration curve shows a nice linearity. The expected error is minimal at large values of the parameter. However in this case the likelihood shows a non-parabolic behaviour originating from the statistical fluctuations of the various reference samples. To overcome this last problem a third method based on a neural network is introduced. The network automatically smooths the likelihood. Again, good calibration curves are obtained. The three fitting methods shows similar expected errors. On the data the presence of a large fitted value for Emiss (82.40) requires the need of a further MC point at 82.85 to complete the interpolation. Systematic errors are under evaluation.

2) Progress report on the tau channel at 189 GeV (Helenka)

The treatment of the tau semileptonic channel has been improved since Siena. The "golden" events, passing a 0.05 Prob(Chi**2) cut are fitted with 2D technique (m_inv,error). The events not passing this cut, together with the ones not passing the similar (0.01) cut in the e/mu channel (mostly taus) are treated separately as "silver" and "bronze" candidates. The difference between these two categories is the value of the invariant mass in the kinfit e/mu hypothesis. This treatment is improving the stat. error by 10%. Calibration curves are reasonable. Work on the final systematics is in progress. Helenka also showed that the Creta Cunningness factor for aleph taus, when properly calculated and when using the latest improvements, is at least as good as the best others. Work on the systematics for the other channels (e,mu) is in progress.

3) The width in the semileptonic channel (Barbara)

The calibration curves are studied in the 2D fit (mass,width) . The mass window has been satisfactorily replaced to the one used for the mass (70-90 instead of 74-94). Contrarily to the mass a P(Chi**2) cut is not yet applied because a small increase in the stat error is observed (it has to be pointed out that the ev by ev error is not used here). The calibration curves are reasonable, even if fluctuations close to the nominal sm point are observed in the muon channel. The study of detector systematics, particularly of the resolution effects, is in progress.

4) Fragmentation in the semileptonics (Jason)

The method, based to MC/data comparison, has been described in previous meetings (see for example Jason and Eric transparencies in Siena). For the semileptonics Jason is using four variables : mass of highest mass jet, mass of smallest mass jet, charged multiplicity and lepton isolation. The first two variables have been updated with respect to Siena by employing standard jet corrections. The mean value from MC reference has been corrected for the difference in mass m_ref-m_fit. It is found that the mass of smallest mass jet is the most sensitive variable. The final fragmentation systematic error is 20+/-11 MeV. In the discussion has been pointed out that it would be desirable to evaluate a weight distribution in bins of the sensitive variable, such that the data and MC are forced to agree in that variable then apply those weights within the fit to determine a mass shift. A note is available on ~wardj/public_aleph/Notes/frag.ps for comments.

5) 4q mass status (Lluisa)

Main changes with respect to 183 GeV :

a) Selection. ISR (see Jeremy talk), B.E. (see Hugo), now NN14 is used.

b) Clustering: Durham-pe

c) Kinematic fit : abcfit

d) Pairing: Matrix Element now used

New n-tples have been produced. Include more variables.The running on DST (necessary to implement the HCAL calibration LEP1/LEP2 correction) gives 4 evts less and 10 MeV shift.

6) 4q W width status (Roger)

Latest news on systematic studies have been reported. One important effect is originating from the difference between measured and SM cross section (80 MeV). It is also observed that differences between abcfit and kinfit disappears when a 0.01 Chi**2 cut is applied. A generalized likelihood has been tried yielding consistent results (both on width and normalization, i.e. cross section). Detector systematics are in progress. One worrying point is the fragmentation systematics : -146+/-87 MeV is found from jetset/herwig comparison. The error on lep energy has to be computed. It has been pointed out that also the lep energy dispersion has to be studied.

7) 4q : BE studies (Hugo)

The cuts on the distributions proposed by Jorgen/Frederic (pion multiplicity and minimum q2) have been studied.There is a major concern related to the fact the pion mult. distribution is different data/mc (0.6+/0.2 in average). This needs further understanding (after the meeting it has been clarified that the same effect is present in Franck/Bolek analysis). The W mass is independend on J/F cuts when no BE corr between W is present. A dependence is observed in the BE between different W case. No dependence is seen in data.

8) 4q: ISR status and detector studies. (Jeremy)

The almost final results with ISR cuts have been presented. 45 events with ISR in the detector have been identified. The BE cut (see above) rejects 181 evts. The new ECRESC and HCRESC routines (see Franco web page on systematics) have been tested with not yet final cards. Corresponding jet corrections still have to be determined (in progress at the time of writing). The conservative upper limit on the syst is 20 MeV.

9) TGC results at 189 GeV (Gonzalo)

Gonzalo presented final results on 189 TGC with second order optimal observables (OO2). A detailed study of the systematic errors has been performed, for 1D as well for 3D fits. The systematics are treated together with a Chi**2 technique in a correlation matrix. The systematics appearing in more than one measurement are assumed to be fully correlated or anticorrelated depending on the relative sign of their delta(OO) shift. Systematics are small (1-3%) for semileptonics, while can be up to 25% for hadronics (dominated by cross section uncertainties).

10) Latest on Zgamma at 189 GeV (Eugeni)

The differences between Eugeni and Benjamin analyses have been scrutinized in detail.Now they get basically same events and same energy , still 3 sigmas from LEP official energy. A check made using 187.6 GeV reference yields the same result. Only possibility for final cross check of the result is an independent estimate of angular bias with order of 1 mrad precision (note of the minute writer). Otherwise result is to be considered final.