The best fit for Z And was when the only frame taken after the meridian flip was used, and it was calibrated using the mixed Balmer and lamp method. The Balmer lines were found from Pi And (also taken after the flip).

The lamp was the one taken for Z And, after the flip. No heliocentric or corrections for radial velocity were made.

The errors were 

H alpha                 -0.4

H beta                   +0.65

H gamma             +1.2

H delta                  +1.0

 H epsilon             -1.0

When the same frame was calibrated using its lamp lines only and not the Balmer lines from Pi And

Errors are:

H alpha                 -0.4

H beta                   -1.45

H gamma             -1.8

H delta                  -2.5

H epsilon             -3.5

So for me it seems the mixed method gave the best results