Read on Twitter
If you're going to automate your tests, and you will be automating some, it's important than you can be confident they run correctly.
So how do you do this?
First, be sure about what you expect from the code, and therefore what you expect from your tests.
I like to run my tests over and over again. Sometimes, I will eventually get results I don't expect. Then I know they are unreliable, and need firming up.
Another method I use, is to change the code I'm testing. If my test now fails, it was probably correct.
If they always pass, they might be missing something. Is it incomplete? It could be missing an assert statement.
UI tests that frequently go red could be using a finicky locator. If there's no other way than the finicky locator, it's a testability problem. You may need to talk to the dev team to get that sorted.
