What Makes a Good Roguelike?
Updated August 3, 2026
A good roguelike is built on three things: runs where your own decisions, not the dice, decide the outcome; variety that changes how you play from attempt to attempt; and deaths clear enough to learn from. Procedural generation and permadeath are only the frame. The substance is a large space of interesting choices, synergies you discover as you go, and feedback legible enough that losing teaches you something.
How is a roguelike different from a hard game?
Difficulty and the roguelike structure are separate ideas that often travel together. A hard game asks you to execute a fixed challenge better. A roguelike rebuilds the challenge every run, so you are asked to adapt to a new arrangement of rooms, items, and enemies each time.
Two mechanics define the form. Permadeath means a lost run ends and you start over from the top. Procedural generation means the layout, drops, and encounters are reshuffled so no two runs are identical. Put together, they turn a single long campaign into dozens of short, distinct problems.
The distinction matters because it tells you where a roguelike earns its replay value. A hard platformer keeps you coming back until you master one gauntlet. A roguelike keeps you coming back because the gauntlet is different tomorrow, and your strategy has to shift with the scenery rather than just the scenery shifting around a fixed plan.
What is the difference between a roguelike and a roguelite?
A roguelike loses everything when you die: the run ends and the next one starts from scratch. A roguelite lets some progress carry over, so each attempt leaves you a little stronger. The strict definition comes from the Berlin Interpretation, the roguelike community’s own attempt to pin the word down.
| Term | What carries over between runs | Example |
|---|---|---|
| Roguelike, strict sense | Nothing. A lost run ends and you start over from the top. | Rogue, 1980 |
| Roguelite | Meta-progression: unlocks and upgrades that make the next run start stronger. | Hades, Dead Cells |
| Berlin Interpretation | Not a game. It is the community checklist the strict definition comes from. | RogueBasin |
This post uses “roguelike” in the loose, everyday sense that covers both, which is how most players and most store pages use it.
What makes a run feel meaningful?
A meaningful run is one where your choices carry more weight than the shuffle. The randomness sets the table; you decide what to do with it. When a player wins, they should be able to point to the reads and tradeoffs that got them there. When they lose, they should be able to name the fork where it went wrong.
Hades, Supergiant Games’ 2020 action roguelike, is a clean example. Each escape attempt hands you a semi-random set of Boons from the Olympian gods, and the interesting play is in how you combine them. The seed gives you options; the build is yours. That authored feel is a big part of why it sits at 98% positive across 307,085 Steam reviews as of August 2026.
The failure mode is the opposite: a run that plays itself, where good and bad outcomes trace back to the roll rather than the player. A great roguelike keeps a firm grip on that dial, giving you enough control that the seed is a starting position instead of a verdict.
Why does variety matter more than a big number?
Variety in a roguelike lives in the size of the decision space, measured in meaningfully different choices, which is a separate thing from the raw item count. A game can list two hundred pickups and still feel same-y if most of them are minor stat bumps. Another can ship a smaller pool where each item rewrites how a run plays.
Slay the Spire, Mega Crit’s 2019 roguelike deckbuilder, is the standard reference here. Its cards and relics are built to combine, so a single relic can push you toward a strategy you would never have planned at the start. The pool is not enormous, yet the interactions between pieces make each ascent feel distinct. If you want a deeper tour of that subgenre, we wrote a whole guide to the best roguelike deckbuilders worth wishlisting.
The design lesson is to count decisions, since raw object totals can mislead. Items that open new lines of play add variety. Items that add a number and nothing else pad a store page and dull the pool.
What makes a death feel fair?
A fair death is one the player can read backward and understand. Readable death is a design job, not a difficulty setting. If a run ends and the player has no idea what killed them, the loss teaches nothing and the restart feels like a punishment. If they can reconstruct the mistake, the same loss becomes a lesson and fuels the next attempt.
Dead Cells, Motion Twin’s 2018 roguelike-Metroidvania, leans on telegraphed enemy attacks and clear damage feedback so a death usually reads as a specific misplay: a dodge held too long, a room entered without a plan. That legibility is what turns “one more run” from a compulsion into an invitation. We dug into why that loop is so sticky in our piece on why deckbuilders and roguelikes get so addictive.
Legibility comes from a few concrete habits.
- Telegraph threats before they land.
- Make damage sources obvious.
- End each run with a summary the player can actually parse.
The goal is for every death to answer the question “what should I do differently?”
Where does Hellmate fit?
Hellmate is our roguelike deckbuilder inspired by chess, and it is coming soon rather than out, so treat everything here as design intent rather than a review. Hellmate is unreleased, and Grandma Left Me a Building, our other project, is a different genre entirely, so this is where our roguelike thinking lives.
The premise: you play a chess piece trying to cleanse its sins and escape hell by playing “hell chess.” You build a deck of Spell, Piece, Consumable, and Joker cards, manage mana, and try to eliminate the opposing king in as few turns as possible, climbing seven randomly generated floors themed after the seven deadly sins. If that theme interests you, we also wrote about how games turn the seven deadly sins into mechanics.
We are trying to hold to the three ideas above. Chess gives us a decision space where position matters as much as the cards in hand, the seven-sin floors and the four card types are meant to keep runs distinct, and the chess board is legible by design, so a lost king should be a readable mistake. Whether we land it is for players to judge once it ships. If the pitch lands, you can wishlist Hellmate on Steam.
Frequently asked questions
Is Hades a roguelike or a roguelite?
Strictly, a roguelite. It keeps meta-progression between escape attempts, so a lost run still leaves you a little stronger for the next one. Most players and most store pages still shelve it under roguelike, and this post uses that looser sense too.
What was the first roguelike?
Rogue, from 1980. The genre carries its name, and its rule that death erases everything is the strict standard later games get measured against. Games that soften the rule with permanent upgrades are usually called roguelites instead.
What makes a roguelike replayable?
The count of meaningfully different decisions it offers, not the count of items in it. Two hundred small stat bumps still play the same way, while a smaller pool of pieces that combine with each other hands you a genuinely new problem every attempt.
Are roguelikes supposed to be hard?
No. Getting rebuilt every run is the requirement; being punishing is a choice laid on top of it. A gentler roguelike still works, as long as the losses stay readable and the player can see which fork went wrong.
Key takeaways
- A roguelike is defined by permadeath plus procedural generation, which turns one long game into many short, distinct problems.
- A run feels meaningful when the player’s choices outweigh the shuffle, as with Hades’ Boon combinations.
- Variety is about the size of the decision space, so count meaningful choices rather than raw item totals, the way Slay the Spire’s relics reshape a run.
- A death is fair when it is readable, which comes from telegraphed threats, clear damage sources, and an honest run summary.
- Hellmate, our coming-soon roguelike deckbuilder, is our attempt to apply all three ideas.
