Dead. Again. Forty-seven minutes into a Hades run, Zagreus hits the floor of Elysium and the screen dissolves to the House of Hades. You pick up the fishing rod. You exhale. And then, almost without deciding to, you press the button to start the next run.
That moment is the whole game design problem in miniature. You failed completely. You lost everything. And you kept playing. Why? Because the failure felt yours The spike that killed you was readable. You made the wrong call in a fight you understood. The game gave you real information, and you used it badly.
Now contrast that with the moment you quit a different game. One where the menu to respec your character was buried four submenus deep, where a tutorial prompt obscured the hitbox you needed to see, where the checkpoint system seemed almost random. Same failure on the surface. Completely different cause. One is a designer’s deliberate decision. The other is sloppiness wearing the costume of difficulty.
That distinction, intentional friction versus accidental friction, is the actual fault line in modern game design. Getting it wrong costs you your players. Getting it right is what separates the games people remember from the ones they refund on day two.
The Original Sin of Permadeath (And Why It Wasn’t What You Think)
Permadeath has a reputation problem. Ask most people what it means and they’ll say punishment. Harsh. Brutal. Not for everyone.
That framing is historically wrong. Roguelike co-creator Glenn Wichman gave an interview to Game Developer in which he stated clearly that permadeath was never supposed to be ‘about pain’. It was a technical workaround for limited storage that accidentally produced better emergent storytelling than any save-scumming system could. Each run had weight because it was unrepeatable. The loss mattered because the run mattered.
The irony is that modern designers who reach for permadeath as a difficulty signifier are often misreading the mechanic’s actual function. It isn’t there to punish. It’s there to make every decision carry stakes, which makes every decision feel real.
Hades, Dead Cells, Spelunky. None of these games feel brutal to seasoned players. They feel legible. The rules are visible. The risk is proportional to your knowledge. You die because you didn’t know something yet, and the dying teaches you. That’s intentional friction doing exactly what it’s supposed to do.
Risky challenge that teaches isn’t just tolerated. It’s sought. Soulslikes built an entire genre identity around it. FromSoftware’s Elden Ring sold 25 million copies by March 2024, a number that dismantles the old industry assumption that accessible-means-easy.
What Intentional Friction Actually Does to a Player
Designers talk about friction like it’s a dial. Turn it up and players work harder, turn it down and they coast. That model is too simple. Friction has a direction, not just a magnitude.
Intentional friction, resource scarcity in Darkest Dungeon, the stamina bar in Dark Souls, the limited inventory in early Resident Evil, teaches through constraint. Every restriction communicates a design value. Inventory limits in RE1 say: this world is dangerous and you can’t carry it all, so choose That choice IS the game. Remove the constraint and you’ve removed the meaning.
The academic research on this is pretty clear-cut. A 2020 study published through the University of California’s eScholarship platform on challenge and retention in games found that players who reported high challenge also reported significantly higher retention. But only when that challenge felt fair and comprehensible. The moment challenge felt arbitrary or opaque, retention dropped sharply. Same difficulty level. Different design communication.
That last clause is the one designers miss. You can have a hard game or an opaque game. Only one of those keeps players.
Supergiant Games understood this with Hades in a specific, almost surgical way. The pact of punishment system lets you dial up difficulty explicitly. You choose which restrictions apply. That transparency removes the arbitrariness. You know why you’re suffering. It feels like training rather than hazing.
Compare that to early No Man’s Sky, where survival mechanics ate resources at rates that felt disconnected from any logic the player could grasp. The friction wasn’t instructive. It was just there. Waiting. Until you stopped playing.
Accidental Friction: Where Good Games Go to Die
Accidental friction is harder to spot precisely because designers usually don’t know they’ve created it. It’s not a deliberate choice. It’s what happens when UX, onboarding, and systems design aren’t pulling in the same direction.
Common forms: onboarding that tells players what buttons do without showing why they matter; progression systems that obscure what the player is actually working toward; technical friction (load times, unskippable cutscenes, mandatory account creation) that inserts waiting where there should be playing.
This last category is interesting because it shows up outside gaming too. And the lessons from adjacent industries are sharper than game designers typically acknowledge.
Platforms that required extensive identity verification before letting users do anything discovered, repeatedly, that users bounced at the verification step in enormous numbers. Not because they were fraudulent, but because the friction felt disproportionate to what they were trying to do. The detailed guide to no kyc casinos published by Pokertube documents exactly this tradeoff in the online gaming space: platforms that stripped mandatory verification out of the onboarding flow saw session-start rates climb significantly, because the first thing a new user encountered was the product itself, not a bureaucratic gate. The mechanic that existed for compliance reasons was functioning, in UX terms, as accidental friction. A barrier that cost engagement without improving the experience for the user who cleared it.
Gambling involves risk. Play responsibly and only wager what you can afford to lose. If gambling becomes a problem, visit BeGambleAware.org.
The parallel to game design is direct. Every mandatory tutorial your players can’t skip, every server-authentication handshake that delays the start screen, every account-linking requirement before the first level loads. These are verification steps your design is forcing on someone who just wants to play. Some of those steps are genuinely necessary (anti-cheat systems have real value). Most aren’t. And players make no distinction between necessary friction and unnecessary friction in the moment. They just leave.
The Progression Trap: When Challenge Becomes Confusion
One of the cleaner tests for whether friction is intentional or accidental is what the player learns from failing.
In Celeste, every death is informative. Madeline bounced off a spike? Now you know where the spike is. You misread a jump arc? The game’s physics are consistent enough that you can model it differently next time. Deaths in Celeste are data. Fast data. The respawn is near-instantaneous precisely because the designers knew that delay is accidental friction. It interrupts the learning loop.
Some RPGs structure their difficulty around stat gates that aren’t visible to the player. You walk into an area, get one-shot by an enemy, and the game’s only explanation is a damage number. Was that enemy too strong? Were your stats too low? Was there gear you missed? The player doesn’t know. They might grind. They might quit. They have no real information on which to act.
The XCOM series, something Josh Bycer has examined in depth on this site, lives in particularly interesting territory here. Its percentage-based hit system is intentional friction: you know the odds, you make a call, and sometimes the dice disagree with you. That’s the game. But early XCOM implementations had UI issues that made those percentages feel less reliable than they were, which tipped the intentional friction toward accidental friction by muddying the player’s model of what was happening.
The fix wasn’t to make the game easier. It was to make the odds more readable. Same difficulty, better communication. Players stopped feeling cheated.
The Churn Point: Reading the Dropout Data
Game designers who track completion rates know something uncomfortable: most players never finish the games they buy. Steam achievement data from 2025 suggests the median completion rate for narrative games hovers somewhere under 30%. For games with permadeath mechanics, that number drops further. But so does the refund rate, which suggests something counterintuitive. Players who engage with permadeath tend to chooseto engage with it. The ones who bounce do so early, often before they’ve hit the mechanic at all.
What’s losing those early players isn’t usually the challenge. It’s the friction before the challenge.
This matters for roguelikes especially. Binding of Isaac, Spelunky, and Risk of Rain all have brutal first-run experiences. But all three have carefully designed first-ten-minutes flows that make the mechanic clearbefore it becomes hard You understand what you’re signing up for. The contract is legible.
Games that front-load confusion, unclear objectives, opaque resource systems, interfaces that don’t explain themselves, lose players before those players ever reach the difficulty that might have hooked them. The accidental friction ate the intentional friction’s audience.
The emerging design principle around this is something like: earn the right to be difficult. Show the player what your game is, clearly, first. Then make it hard. A developer who gets the order backwards is making a design mistake, not a philosophy statement.
Difficulty Settings Are Not Surrender
One persistent myth in game design circles is that difficulty settings dilute a game’s artistic intent. This argument shows up most loudly around soulslikes. The community-level debates about whether Elden Ring should have an easy mode are almost a genre tradition at this point.
The argument misses the actual design question, which isn’t “should the game be easy” but “what is the friction for”.
If the friction is intentional, if the difficulty IS the design, the way permadeath is the design in a roguelike, then a difficulty setting that removes it changes the game in a meaningful way. That’s a reasonable artistic choice to protect.
If the friction is incidental, if it emerged from technical constraints, time-to-ship pressures, or assumptions about the player that turned out to be wrong, then protecting it is protecting a mistake. Accessibility options that address control schemes, text size, reaction-time windows, and UI clarity don’t compromise a game’s challenge. They strip out accidental friction while leaving the intentional kind intact.
Celeste demonstrates this with its assist mode. Every assist option is labeled clearly. The player knows what they’re changing and what they’re keeping. The game treats the player as someone capable of making an informed choice about their own experience. That’s not surrender. That’s design confidence.
FAQ
What’s the difference between intentional and accidental friction in game design? Intentional friction is designed constraint that teaches the player something or creates meaningful stakes. Think stamina bars, permadeath, or limited inventory. Accidental friction is a barrier that slows players down without communicating anything useful: buried menus, unclear objectives, or mandatory steps that exist for technical rather than experiential reasons.
Why do roguelikes use permadeath if it’s so punishing? Permanent death makes individual runs feel consequential. When every session can end in failure, every decision carries weight. The mechanic doesn’t exist to punish. As Rogue’s co-creator Glenn Wichman noted, it originally emerged from storage limitations and turned out to produce richer emergent storytelling than save-scumming ever could.
Does adding difficulty options hurt a game’s design integrity? Only if the difficulty is the core design. Accessibility options that address control schemes, UI clarity, or reaction timing remove accidental friction, not intentional challenge. A game confident in its design can offer options without compromising its artistic intent. Celeste’s assist mode is the clearest example of this done right.
Why do so many players quit games before reaching the hard parts? Most early dropouts aren’t caused by difficulty. They’re caused by confusion. Unclear onboarding, opaque systems, or slow-to-start flows lose players before they ever engage with the intended challenge. Accidental friction in the first ten minutes is a far bigger retention problem than a steep difficulty curve later in the game.
How can developers tell whether friction is intentional or accidental? Ask one question: what does the player learn from this constraint? If the answer is “something specific about how to play better,” the friction is intentional. If the answer is “nothing, they just wait, or re-enter data, or work through an unclear menu,” it’s accidental. Playtest data almost always confirms the difference, because accidental friction shows up as session abandonment at consistent points in the flow.
—
The Stakes Are Real
Game design has never had more tools for measuring friction. Playtesting platforms, heatmaps, session-abandonment analytics, A/B-tested onboarding flows. A modern studio can pinpoint exactly where players are leaving, and often why. The problem isn’t information. It’s philosophy.
Some teams still conflate difficulty with quality. Hard equals serious. Accessible equals compromised. That framing ignores everything the best roguelikes have shown over the past decade: that clarity and challenge aren’t opposites. That a player who understands your game will tolerate enormous difficulty. That a player who doesn’t understand it will leave regardless of how forgiving you make it.
Permadeath works because it’s honest. You know the stakes before you start. Every run is a contract between the designer and the player, and the terms are clear. When designers extend that same honesty to every layer of their game, its menus, its progression, its onboarding, its communication of risk, they stop losing players to confusion and start losing them only to the difficulty they chose to build. That’s the only churn worth accepting.
