I want people to play around with [Spelunky 2] and not focus so much on "beating" it. I think it's a difficult concept for people to wrap their head[… - Derek Yu

There's a certain glee [among game developers like Bennett and me] from players feeling frustrated and disappointed. And it's not because we dislike players. It's actually because we really care about players, and these are emotions that you need to feel to have a well-rounded human experience in a video game.

The key to making a challenging game that's not frustrating is to give the player all of the tools that they need to overcome the challenge— and never make it the game's fault that they lost. [...] Roguelikes give you lots of those tools, so even though they're really tough, it's always fun.

The business model of arcades— I don't know how it will be replicated ever again, but it created such an interesting category of games, just based on the unique features of it. And I think it's one of the few places where the business model of the arcades really forced this type of design that was— I call it "lean and explosive". [I say] "lean" because you have to push players along to the interesting parts of the game as quickly as possible. And you just don't see that in modern gaming and modern game development. [...] And I say "explosive" because they don't save anything for the end. The experiences are quite short: to play through an arcade game, it's 30 minutes to an hour, tops, for the longer arcade games. And you don't want to save anything for the end because players are renting the game a quarter at a time. And so, starting with Stage 1, you've got to put it all out on the table, while still — in the later parts of the game — giving people something to look forward to. And I think that has been very influential on Spelunky 1 and 2, and it's just a type of design I really enjoy.