I have no idea why you are using pre-70 parses with a really low sample count to boot as a comparison of which job parses higher. Pre max level parses hold no weight and adding to the fact that you're making comparisons while not having access to the entire jobs kit. It literally does not matter until you're max level. You aren't going to parse higher than a SAM unless you outgear him or the SAM is terrible, end of story.
