When there is a very low ratio of test failure, it reports 0% failure and 100% success. For instance:

Success Failed Skipped Total

1. % # % # % #
   2371 100% 7 0% 0 0% 2378

In that case, I believe the precision should be higher, and it should rather display 0.3% failure and 99.7% success. At some point, it's impossible to be precise enough not to report 100%. It could for instance add a < and > to indicate it's not quite 100% or 0%:<100% >0%.

Another possibility is to color/highlight in red the lines with at least one failed test.