> the engine limits the kinds of games that are thinkable. Does anyone actually do that? Most people just sit around thinking things up, and only later are the

> the engine limits the kinds of games that are thinkable. Does anyone actually do that? Most people just sit around thinking things up, and only later are they sobered to the fact of how much work it would be to implement a niche feature that isn't a substantial part of the game. > Which is why so many games look and play exactly the same. I'd argue that 90% of that is caused by default shaders. It's why people complain about "unreal engine sheen" it's how most of the materials get set up for you if you don't care to change it. The other 9% is default character controllers for the same reasons. 1% or less is going to be caused by some unalienable quirk of the engine that can never change. > Steam success is somewhat correlated with making your own engine. Novelty and standing out from the crowd are the things that sell your game. Successful games often have novelty and stand out. But there's an aspect of survivor bias. Probably plenty of games that are novel or stand out in all the wrong ways. It's also not clear that making your own engine causes novelty or you to stand out. > It's a lot harder to do new things if what you can think is limited by your infrastructure. Again, not convinced that's the way people go. Maybe they try to implement it and it's clunky, or doesn't scale, or trashes performance. So they scale it back, rework it, or write some custom code. An engine isn't magic, any amount of code you have to write to implement a feature is less than a whole new engine. Would Balatro have been a better game if it had a completely bespoke rendering pipeline? I'd say probably not.