I’d love to see some examples of cases where he says the tool gets the wrong answer.
I had a play with the site a while ago, after one of the stories here, and it basically seemed to do the right thing for “obvious” cases.
I think the “testing” issue is a red herring. I doubt that something as simple as this would not be doing what it was supposed to do. It’s just a website that asks questions and does no more than about 10 lines of logic to work out the answer. It seems more likely to me that this Chris Chaplin has entered details of somewhat borderline cases where he disagrees with HMRC about what the outcome should be, and predictably the tool has told him what HMRC thinks, not what he thinks.