For better or worse, software development is largely a "ship to learn" process these days. You don't have the time or luxury to do years of UX/UI testing to create something that works as well as a hardware device that's been in use since 2006, which was itself an iterative upgrade of a device introduced in 1994. Even in an industry like model trains where the level of user patience is extremely high, the world just doesn't work that way anymore, if it ever did. The mentality is to ship a minimally viable product and then iterate as fast as you can until you have something everyone can adopt. In the case of both MTH and Lionel, they had the additional benefit of knowing that folks like most of us already had a working system to use, and we could decide to move into the new space/hardware/tech at our own pace. New people to the hobby were going to default to the Base3/WTIU, and so most of the testing went into the most basic of features, which is perfectly logical. If all you've had is a LionChief remote, the Cab3 app is a huge upgrade. I'm betting those customers don't see the Base3 the same way that someone like me, with 300+ locomotives who's been using a CAB1/CAB2 since they were released, does.
The real question is: do you trust Lionel to get it right over time? The Legacy system wasn't perfect at the beginning, but over the years it's become pretty solid, and the known issues have been taken care of. Which is a better bet? That the oldest, most established company with the largest install base and largest revenue stream is going to figure out how to make people like us happy with an app given enough time, or that the cost/joy/pain/frustration of moving to a new hardware/app/tech stack (whether that's DCS or Blunami or DCC or whatever) will get you the experience you want. It's like betting against Apple or Google. Those companies aren't always the fastest to market, and don't always introduce new features first, but they very rarely fail to make their customer bases happy. You can always find little companies that are on the bleeding edge, but if history tells us anything it's that the tortoise really does win A LOT.
Personally, I choose to believe in Lionel. I use the CAB2 for everything, but I have the app installed and keep tabs on it to see how it's doing with each upgrade. The 4-digit addressing is a HUGE feature for me just because of the number of engines I have, and I'm curious to see whether we'll ever be able to upgrade the electronics in our older TMCC/Legacy fleet to take advantage of it. I trust that given enough time and constructive feedback, the Cab3 app will be something I have installed on dedicated devices/tablets around the train room just like I have CAB2 remotes today. We aren't there yet, but I've been in this ecosystem a looooong time, I can be patient.