Yeah. Another way I would put it is to also look at inputs that strayed off the happy path. There may be something to learn from them as well, maybe something that adjusts the happy path.