You're welcome to make a Steam platformer in C2 with over 20 thousand players too and then come back and tell me I am wrong
The engine has bad bugs, either in its engine loop and platformer plugins, and/or its NWJS wrapper.
Yeah, I hear you. This was exactly why I had to leave construct for the final product. I mean I literally read through just about every c2 runtime file I could. The fact was, besides performance, I needed certain collision details revealed to me from the deep dark magic box that contains c2 collision logic. Because the code wasn't accessible, I had to write my own collision system as events which ran like poop. I then wrote it as a jscript behavior and it ran better, but then the real question was: Why am I fiddling with construct when I'm programing almost everything - I'm working harder for less power. I love constructs ease of use, but everytime people want more power there is a fear that will lead to ease of use diminishing. Unity isn't hard to use. It is simply hard to navigate by comparison. Construct has the gui and flow figured out and is meant for 2d. It just needs more power. But to be fair, it also has a small dev team.
Now to be fair to contruct, I tried for like a year to make it work. But every time I requested something like - "Can I have the collision contact angle exposed by construct" ... It was met with we don't think we see the reason for needing that. Dismissed. etc... This makes making a game via the real construct method to be impossible unless your game functions like most other games. Think about a game like, "the floor is jelly". You couldn't make it with construct very well. Similarly, sonic the hedgehog and mario are also not possible (by events).
Then we get to the performance issue. I have an okay laptop. Its old. But it runs unity, and runs my game at 180fps. Then I run the game prototype on the same laptop and get 12 fps. I'm also not an idiot when it comes to optimization. I like wasting time optimizing for no reason, but I had to make so many theoretical compromises to get the game to work with c2, so I switched.
I like the power. But I liked the ease of the event system in cosntruct. I wasn't meaning that the devs should drop or require unity, rather that they should look at a way to get more power with only the work of creating the interface and SOL magic in unity. I think it would be a smashing success over there. Since I suggested it 2 more event style systems have been released (BOLT I hear is good, but haven't tried it) anyway. Now that I wrote a book.
I have to address the bit about complaining and walls. I don't see this as a complaint but a point of reality. It could change, thus the motivation for sharing the struggle. I want change, thus I take the time to suggest things I would see as helpful. The point is, I like construct and would work with it if I could. The fact is, I cant anymore and that is still a bummer.
Yes, I have made games with construct. They took a month to make, my friend actually even made outside tools that would go through construct save files and do things in bulk that were tedious to change by hand. I worked construct through and through. But the results that construct excelled at making were simple type games. Yes, you can make complex things, but there is an exponential increase in the work it takes to make things scalable and complex in construct. I need real oop for example and Construct still lacks that.