It is a good experiment and fun to watch it unfold.  Part of the question about having a target block generation period of less than 10 min is a bad idea was the amount of forking that would take place. With such low difficulties, the effect is probably multiplied and would serve as data for anybody else planning to change the target interval.