Test Automation

In the book you mentioned some organizations may choose to look at a problem that maybe say failing one in ten times and say if it fails I'll rerun the test. Why is that not a good strategy? We want this failure to occur more often and not less often in order to be able to investigate them but unless we really investigated the failure we cannot assume anything about its root cause. So strategies to repeat, the problem making fails often as possible and then reduce the scope and make things more and more specific until hopefully you can solve the problem. If today we do one retry after it fails and it may be enough maybe in the future it won't be enough.