To me it seems that the issues with C2 performance are related to HTML5 implementations. IE seems to actually run my games super smooth, but chrome seems to do worse, even thought they're all pretty simple, and even when Chrome says 60fps, it doesn't 'feel' quite as buttery smooth as running it in IE. C2 is also the easiest to develop with if you're not a programmer (and even if you are for simple things it's quite easy). It has very nice built-in behaviors that will do a lot of the work for you.
So if you want to go with an HTML5 game, probably C2 is your best bet. But as I said before, HTML5 to Android / iOS can have a variety of results. For example, in my case there are documented Chrome bugs that Google doesn't care to fix for years (as gaming is not their top priority, instead they focus most of their efforts on making Chrome secure), making some aspects of my game's polish not possible. The reason the state of Chrome matters is because the main export option to android has been (at least until recently) Intel XDK which is based on Chrome. There are other options like CocoonJS, this was abandoned for a while and did not support native AdMob ads but now ludei seems to have updated his C2 plugin, so I'll probably give that a try soon. It seems like it may be a good option again ,but i can't speak to it just yet, and it is not officially supported by Scirra (although it was in the past and may be again in the future, the technologies C2 depends on and supports seem to be in a state of constant flux, although again, Intel XDK seems to have gotten the most focus as of late). Results seem to vary wildly by each game, on which export option works the best, which to me is just another way to say that mobile export is 'unstable'. It works, sometimes well, sometimes poorly, and you won't know ahead of time which. Another issue to consider is that Intel XDK adds about 40mb of useless browser engine (since it's based on Chrome) unrelated to your game. So your game will have an extra 40mb in size than using a different export option, meaning you may run into the google play / iOS store file size limit more easily.
If you want something that isn't based on HTML5, which is likely to run your game better on mobile platforms in spite of what Scirra / Scirra loyalists may say, there are better options, with my favorite being Unity + some toolkit appropriate for your game-type. There are actually quite a few Unity toolkits that can do a lot of heavy lifting so that you don't have to do much coding, but this will probably end up costing you somewhat more than C2 as each is tailored to a specific game-type, so you may need more than one if you want to work with different genres.
Overall, C2 is a good value for the money it costs, so you should buy it anyway. But I would not try to make a production-quality mobile game with it, nor a production-quality game with it on any platform that involves substantial amount of work. Part of the reason is because support on bugs is lacking since it depends on 3rd party technologies that are still in alpha, beta, or other development stage, meaning that Scirra cannot fix those bugs and will often redirect you to said 3rd parties. However, 3rd parties often don't care about the bugs as high priority, leaving you stuck. This in turn means that making a game that requires a lot of time is very risky since your work can be wasted. C2 is great for smaller/simpler games or web games though, and for putting together a quick prototype of various gameplay mechanics to see how they work together before making them into production on a different tool.