k0s event is generated using Lund7.4. The momentum is generated uniformly between 0.5 to 5.5 GeV and the polar angle is between and .
For the fit, candidate vertex point is obtained by using the class JSFVertexing. This class return the closest distance between two helix as the vertex quality. If the quality of JSFVertexing is less than 10cm, they are applied to the geometrical constraint fit by JSFGeoCFit. Among 10000 of ks0 events, 5283 events has two CDC tracks whose vertex quality less than 10cm. The rest of the events are mostly . The macro file used for this study will be found in the file, macro/V0TagStudy.C in the JSF distribution.
The decay length distribution of generated k0s is shown in Fig. 1. Confidence level distribution of the fit is shown in Fig. 2. The confidence level distribution is flat, which confirms the correctness of fitting.
The number of trial of the fitting is shown in Fig. 3 and the distance between generated decay vertex and the fitted vertex is shown in Fig. 4.
Since ks0 has long decay length, some of them does not have enough hit points in CDC. I suspect that such ks0 are the source of entries whose number of try are greater than, say, 10 in Fig. 3 and distances greater than, say, 3 cm in Fig. 4. To get the good vertex resolution by using the CCD vertex detector, ks0 must be decayed before the second layer of CCD whose radius is 3.6cm. These are the entries in the first bin of Fig. 4. The entries in the rest of the bin in Fig. 4 are considered to be obtained without vertex detector information.