fix: DIS: less strict on fits, use mean/stddev stats

This replaces the hard criteria on fits by mean and stddev values on what should be a pretty normal distribution anyway.