Some days, I think I'd rather gouge my eyes out than read another email about a new roguelike or roguelite. This confuses me, because many of favourite games are roguelikes or roguelites, including Dead Cells, Balatro, FTL: Faster Than Light, and the recent Morsels, a reeking procedural dumpsite that speaks to the overproduction of Rogue/rogue derivatives at large.
Roguish games are everywhere right now. According to SteamDB, 1602 games tagged "roguelike" were published in 2024 out of 18567 total, versus 312 out of 9655 in 2020. Stir in roguelites and the countless games that advertise themselves as having "roguelike mechanics", and I sincerely worry that you're describing the majority of PC releases from the past couple of years.
Then again, how many of these games are 'genuinely' Roguish? Roguelike and roguelite have become such broad concepts as to be functionally useless, describing everything from carpentry to casino machines. This was the case back in 2011, when Adam Smith (RPS in peace) marvelled over a peculiar new "Roguelike arcade game" called The Binding of Isaac; it was the case in 2016, when Alice O (RPS in peace) observed that the term roguelike is "so very bendy and too confusing to throw around without explanation".
Personally, I define these games as follows: a "roguelike" adheres more closely to the original Rogue from 1980, featuring permadeath, equipment or character progression, procedurally generated environments, and semi-randomised challenges. A "roguelite" is a less harrowing variant that may not have an explicitly defined rogue character or dungeon - in particular, it may feature an over-arching system of unlocks that persist between runs, taking the sting out of restarts. Still, perhaps a more useful way of thinking about roguelikes and roguelites isn't to tick off correspondences or take them back to their roots, but to look at the shapes they form within the greasy currents of platform economics and player habits.
There are a thousand definitions and mine is just one among many, I'm aware. This is not a "right vs. wrong" matter, it's how you cut things out.
For me, a roguelike has four rules:
People aware of other definitions (like the Berlin Interpretation) will notice my #4 is not "grid-based". I think the grid is just a consequence of keeping individual elements simple, in this case movement.
Those rules are not random. They create gameplay where there are limits on how better your character can get; but you, as the player, are consistently getting better. Not by having better reflexes, not by dumb memorisation, but by understanding the game better, and thinking deeper on how its elements interact.
I personally don't consider games missing any of those elements a "roguelike". Like The Binding of Isaac; don't get me wrong, it's a great game (I love it); but since it's missing #3 (combat is real-timed) and #4 (complex movement and attack patterns, not just for you but your enemies), it relies way more on your reflexes and senses than a roguelike would.
Some might be tempted to use the label "roguelite" for games having at least few of those features, but not all of them. Like... well, Isaac—it does feature permadeath and procedural generation, right? Frankly, I think the definition isn't useful, and it's bound to include things completely different from each other. It's like saying carrots and limes are both "orange-like" (carrots due to colour, limes because they're citrus); instead of letting those games shine as their own things, you're dumping them into a "failed to be a roguelike" category.
So would you consider Slay the Spire and Balatro to be roguelikes? I think they meet these four criteria.
Slay the Spire: yes. All four rules are there, specially in spirit. It's also a deck-building game but that's fine, a game can belong to 2+ genres at the same time.
I'm not sure on Balatro. I didn't play it, so... maybe?