On the verge of completing a game with it I originally planned for a long dead game-jam that was on this very website. I am pretty sure I can get it done within 100 events, which is a good free edition limit in terms of preventing needless feature bloat. (this limit is important to know if you have big plans for future projects with it, my project has already broken the 70 event mark and this is just a one-layout game)
 
I will probably post it in it's own topic when I finally polish it off over the next few days, but I will throw a link to it in this topic. I am using the puzzle game I am working on as a intro to working in-depth with the software's event-based scripting, while I spend the year bouncing around multiple middle-ware engines. (I will most likely post the source file on scirra's website, since there are basically NO working examples out there of dropping block type puzzles, and it seems like exactly the thing a lot of new users would want to look for immediate reference on)
 
I am really enjoying working with this engine, even with all the challenge of doing reasonably complex stuff with little to no prior knowledge of how the features work, so I might end up buying the license. I do like how you can get the single-user license fairly cheap, and be ready to publish to a wide variety of platforms. (the way I see it, it is already much more desirable alternative to the newest edition of Game Maker in that respect alone)

I don't like that it's html5-based and I have no experience with the 3rd party exe wrapper they've integrated into it so I can't give an opinion. I did use Construct 1 and it was awesome.