I understand this pessimistic point of view.

I’ll ask a different question: Does automatic tests save time?

Let’s measure it… Writing good tests takes time, a lot’s of time. Let’s say 25–50% of task development. Maintaining tests, and keeping them valuable takes time, let’s say 10–25% of development time.

How much time do good tests save?

I cannot answer.

The experience tells me that writing good tests pays itself.

The experience tells me that Mob ends up with better architecture I could ever do by myself; with great tests; the code is easy (really easy) to maintain and easy to add features.

Mob also helps with decisions — not only technical, but also business/priorities.

Anyway…. I don’t have numbers, only experience :-)