That's a whole different argument than soil moisture prediction accuracy. Sure, a 100% local-based system is generally better than one that's dependent upon cloud services. But in this case you just have different points of failure (your Ecowitt sensors) than I do. In both our cases, there are things that can go wrong which turn our respective systems into dumb timers.
Far as being "very" expensive, Rachio cost me $100 over your solution ($219 at Costco)., but it turned out to be one of those rare things that I could just install and forget, so for me that cost is insignificant. If my lawn looked like crap or my water bills were high, I'd certainly have a different opinion. Nothing wrong with your approach though.
