How do you all define failure of retest? If just the same issue repeats under just the same conditions as described in initial report? Or do you mean "the use-case still has an issue (probably a different one)"? Because for the first one I 17% is big, for the second - small.
?:the art of a constructive conflict perceived as a destructive diagnose.
I am usually not very popular at how I look at retests.
When I find an initial defect - it is included in the failure rate. It goes back and code is changed. If the change fails it is a second defect. I consider it this as I am testing different code and the new set of code does not work.
I have not failed. I've just found 10,000 ways that won't work" --Thomas Edison