Prying Action in Cap Plate Connections

Does anyone know how RAM Connections calculates the allowable tension capacity including the effects of prying action? I ran a couple tests in which I varied the bolt spacing (not the gauge) and the allowable capacity due to prying action remained the same. Perhaps to be more specific, how does RAM calculate the tributary length "p"? AISC 15th ed. shows it can be conservatively taken as 3.5b<=s. If that were the case, a lesser allowable capacity should result when the tributary area is limited to the bolt spacing "s" relative to the greater 3.5b. Comparing ram results to hand calcs it seems like 3.5b is the default tributary length.