mean effective gain Cst 2013...some results over zero...

I am calculating mean effective gain of mimo antennas in cst 2013,from the menu

Post Processing->Template Based Post Processing->farfield and antennas properties,farfield result->advanced->,i browse results...then i choose the farfield i want,then MIMO->RESULT TYPE =mean effective gain,power distribution function=gausian for elevation,uniform for azimuth,xpr=0,and i change vertical component and horizontal for deferent degrees...and at some degrees the meg take positive results..What i am doing wrong?

Also does someone know how cst calculate meg?

thank you...