I think that Apple screwed the pooch a bit by allowing the app development at the time of Watch release. WatchKit 1 just wasn't what it needed to be, and I think a lot of developers must have looked at it, shook their heads, walked away and didn't come back. It was really hard to muster any kind of enthusiasm to develop for it, when it was clear as day what the watch was really capable of, and that Apple themselves had APIs that were lightyears ahead. If they waited those few months and launched developer initiative with the launch of Watch OS2, I think things would have been better.
Even now with WK2.0, everything feels very sandboxed and limited, and you have to jump through hoops to create something outside of that sandbox, if it's even possible. It really shouldn't require some kind of genius level creative workarounds to create something like a Flappy Bird game, but that's precisely what you need even now to create a simple arcade game for example. For a while I was very worried that I won't be able to create a seamless switch between the hot/cold dials in my app and have them both smothly animated and controlled via the digital crown. Something so seemingly simple at first looked like it would be impossible to do the way I wanted, but thankfully it wasn't. All the while that Mickey Mouse watch face that's certainly using spritekit or even vector animations, laughs at you and shows how much more access to APIs Apple still has.
Then there's the fact that indeed a lot of things, including arcade games, just don't make much sense on the watch, when it can be served so much better on the phone. Watch app needs to do something really quick and really useful, or people won't bother.
Lastly, even if you have a good idea for something useful that falls into this sandbox where it's possible to do it, there's a level of unreliability to the way things work even on WK2.0 that I wasn't expecting. Things start becoming unpredictable if your app layout is too complex, or if you rely on stable and timely watch/phone connectivity. Sometimes things that you're 100% sure should be working, simply don't work reliably, and then you have to be creative with what to do if that happens. You can see this so well in the Trivia Crack watch app. I can't even count the number of times I had to press that "Retry" button after the question I answered failed to commit. Even though the developer has released multiple updates trying to tackle that issue, it still sometimes happens. I totally understand their pain - making this app work as reliably as it does, was way more difficult than I ever expected it to be.
