As was reported in www.star.bnl.gov/HyperNews-star/get/embedding/497/1/1.html, there was an issue with deep int pseudorapidity distribution for reconstructed electrons.

One of crosschecks which was proposed was to check if this might be drift velocity problem. To do it, I projected both the MC and reconstructed tracks (with number of common points >= 10)  to the radius = 3 cm and plotted the difference in the Z position (fig 1.). As was suggested, I took fully east and fully west tracks - vz*eta>0 The obtained distribution has a Gaussian shape centered at 0, so I guess the drift velocity is not the issue.

Fig. 1 - Delta Z for MC and reconstructed tracks projected to r=3 cm