I'm sorry but I agree with @Konidias 100%. I've been saying the same things for the past year.
I've been the first to publish a "game" for iOS made in construct and again the first to support the native Game Center (on PhoneGap). I say this not to boast my achievements, but to prove that I'm speaking from experience.
First of all the argument that optimization fixes the issue is complete bullsh*t, I optimized my game in unimaginable ways, rewriting even part of the engine to use the latest hacks, I got to 40fps, but on iPhone4 it lags a lot and I won't risk my company name releasing a half-assed job. Not only that, but to achieve those abysmal fps I had to strip down the graphics to the bare, resulting in a project that look less nice than what I wanted. And let's be frank, my game is very very simple, I did a quick test with another game maker (that exports native apps) and I was getting 60fps constant with a layer of particles that I added on top of it to test speed.
The other argument I see often made is that you can get good fps with mobile Safari, true that, but nobody gives a crap about web games on iOS, plus they are extremely hard to monetize, it's either the AppStore or time wasted (unless is just an hobby).
Meanwhile the WebView (the one that PhoneGap uses) yield horrible framerates (certainly not the fault of Scirra, it's just not in the interest of Apple to make it work).
So we have to rely on external services (like CocoonJS) that are an expense on their own and are clearly not updated regularly (who follows the development of cocoon knows perfectly this). In the end you have: One level of abstraction in the form of Construct, and another level of abstraction which is CocoonJS. If any of these companies fails or decide to stop updating you are done as they are both (partially) closed source. I am personally waiting for the past 3 weeks for ludei to add the promised Game Center support which I'm not seeing coming.
I write this because I believe that construct 2 is the best game maker ever made and a very very good engine, even comparable to some big boys like Corona as far as game development possibilities: you could write something as big as Diablo II with c2, no doubt about it. And on Chrome you'll alse get amazing fps.
Still and I know I'll get bashed for this, HTML5 gaming is a failure, and that's it, there is no point to be made, no redeeming qualities, it's just that: a failure. Let's be honest about it.
HTML5 is the future of Web Applications, and at that is big, beatifull and perfect, but gaming no, and never will be sorry.
Why? Simply, because any game engine there is can be ported to multiple platform, every game engine, every game maker on the market (except C2) has native exporters to every possible platform, and the few that don't are taking steps to do it.
Any argument can simply be countered with this: let's say you develop a game, you get big, you get on Steam, you want to support steam features (achievements, cloud, workshop) good luck doing that with HTML5, it's certainly possible, but whatever your game will be it would be less time consuming to just port it to another engine, even controller support is still sketchy in HTML5!
I finished my game in 2 weeks with construct, then spent 6 month optimizing and waiting for people to finish their libraries and now I'm still sitting on it and I lost any drive to finish the job, it will be certainly the last game I develop with construct 2 (unless a real miracle happens here) because the time I lost optimizing I could' ve wrote it native and be done with it already.
I'll certainly use it for prototyping, but it's such a shame, such a horrible shame, like having a Ferrari in your garage but you can't drive it because you don't have the right fuel.
I believe that if Scirra decides to finally drops HTML5 and start thinking about a cross platform language (like Haxe or lua) they could dominate the market, as things stands right now I see it as a very unlikely eventuality...