Metric to capture cell fragmentation rates based on hashtagging UMIs

Has anyone derived a metric to capture an indication of 10X run quality with respect to cell fragmentation based on ratio of cells with high hashtagging umi counts (maybe >10) to those with low umi counts (<10?)?

See an example here, with two sets of libraries: one set of good quality libraries, manually validated by wet lab scientists, and a second set of bad quality libraries (in comparison to the first set).