People working on roguelikes in the modern era need to start investigating how to massage RNG into something more consistent. https://twitter.com/MckKirk/status/1387739277564518404

If you want your game to be appreciated you need to limit the vectors for frustration and repetition. Most people aren't a fan of that kind of thing.

Here are the questions I ask myself when thinking about designing a roguelite/likelike, and how I evaluate the ones I play:

1. After I die, are the things I do in my next run meaningfully different?

Good: Outer Wilds. I could do anything and the time it took to get to what I was trying previously was almost always trivial.

Bad: Don't Starve. Every time I died I had to do the entire early grind to get back to where I was. There was like an hour of busywork to progress.

2. Does my progress in a play session improve other playsessions?

Good: Rogue Legacy. You unlock stuff which then made next runs easier, which sped up early content.

Bad: Below. You play for hours and die, and there's only a handful of things that can improve your next run, and if you don't find them, it sucks.

3. How punishing is an unexpected or unavoidable death?

Good: Dead Cells. The upper limit on a full playthrough is less than an hour, so your total investment is ultimately hard capped.

Bad: FTL. Since the game has hard gates, an unexpected RNG can torpedo a good run and prevent global progress.

4. If I can take literally nothing from one run to another, how entertaining is the game on its own?

Good: Powder. Who doesn't love insane gods who fuck with you?

Bad: Almost every other roguelike tbh. The amount I laugh when an unknown potion causes me to blow up doesn't offset the fact that it can happen at the worst time in the worst way.

5. Does randomizing content actually improve my game experience?

Good: Most deck-builders. Because it forces you to explore and understand interrelated mechanics in a way you may not otherwise.

Bad: Chasm. The randomization only happens once you beat a game, so many people won't even see it after putting many hours into your first run.

Anyway, I hope these rules of thumb can help you; if you can't answer any of them in a positive way you should rethink your design.

Also it's worth nothing that games that do one of these badly aren't necessarily bad games. But they have pain points that can chase people off.

This reminds me of another question:

How predictable are outcomes? https://twitter.com/wirelessmouse/status/1387771559310462980

Good: Shiren the Wanderer. You can judge how well your run is going and know if you are likely to succeed or not, and the game gives you tools to plan around that.

Bad: Spelunky. You can be having an ace run and then make one mistake and boom, it's over.

This is an excellent thread and observation from Craig which boils RNG down to its simplest explanation: https://twitter.com/craigperko/status/1387772965455122439