I was just stating what the data from the USA offsite is showing, which is based on a pretty large sample size. I don't really have a logic behind it, but appearantly it's true. In any case, if you are training JS on bigs and are willing to train out of position for it, I suggest using 1v1 guards over JS guards since bigs can use the extra DR/HA from the 1v1 training. Only exception is if you really want a little JR instead of the sizable DR/HA, then you could train JS guards for your bigs.