Difficulty is a place you choose

Difficulty is not a weather system that genres inherit. Hollow Knight: Silksong, Team Cherry, 2025, explores a woven map, and that exploration is not the cruel part. The bosses are. Someone placed the strain. A hard game that cannot say where the strain lives is just uneven.

Strain is what makes a win memorable, and it is what makes an unfair hit unforgivable. The condition is that the player can see why they died. A tell, a pattern, a mistake that was theirs. Without that, difficulty is a tax. With it, difficulty can carry a story about competence, or about a world that does not care, as long as the world is consistent. Give people a way to change the pressure if the strain is not the point of the scene. Do not pretend the pressure was an accident of the genre.

Difficulty is authored, not inherited. Hollow Knight: Silksong is a Metroidvania built around threading an interwoven map for upgrades, hidden routes, and secrets, and that exploration layer is not especially punishing by itself. The strain lives in the bosses: long fights that ask you to read patterns, find safe spots, survive phase changes, and learn exactly what Hornet's moveset can and cannot do.

The first Hollow Knight had hard stretches. Silksong leans on the pedal harder. Bosses in the early and middle of the game can eat dozens of failed attempts before you know enough to close one out. For some players that is the whole appeal. For others it lands as dejecting, demotivating, or plain cruel.

Both reactions are real, which makes the game a good place to ask the underlying question. Why do designers choose to make a game hard? What does that choice buy them, and where does it stop paying off?

The plainest reason to make a game hard is that the size of the win tracks the size of the obstacle.

A boss that dies on the first try leaves nothing behind. A boss that takes an hour, and sharpens your reflexes, your knowledge, your timing, and your nerve on every failed run, gives you a moment you keep.

That trade is the whole pitch of a game like Dark Souls: joy manufactured out of hardship. Silksong is reaching for the same thing. A fight like the Last Judge holds you tense and locked in for a long time, then flips to euphoria the instant the last hit connects.

Easy games almost never produce that particular release, because the release is made of the struggle. You remember the victory because winning required you to get better first.

The second reason to raise difficulty is to keep the playing in agreement with the telling. A game that describes a world of misery and death, then hands you enemies that die in one hit and hazards that cannot hurt you, is arguing with itself.

Celeste shows the alignment cleanly. Madeline's story is stubborn persistence through setbacks on the way up a mountain, and the player's story is persistence through repeated failure. The difficulty is not a challenge bolted onto the narrative. It has the same emotional shape as the narrative.

Silksong runs the same play. Pharloom is withered and haunted, other characters have already failed their pilgrimage, and Hornet only continues by growing stronger and wiser. Practice, exploration, and persistence make you do the growing alongside her.

The third reason is instructional: difficulty steers you into a specific way of playing. It works like a teacher, refusing to reward shallow tactics and paying out for the behavior the game was built to produce.

Furi is about patience and observation. You dodge, you parry, you wait for the rare gap and punish it. Let the player walk in, soak a few hits, and mash until the boss falls over, and the elegant exchange the game exists for stops happening.

That is the general pattern. Easy games tolerate a single repeated strategy. Hard games tend to demand the full surface: every mechanic, every tool, every system, and every part of the level.

Silksong asks for exactly that. Its bosses want pattern knowledge, the right charms and tools, and a deep feel for how Hornet moves. Difficulty compels attention, and it converts systems you might have ignored into things you have to know.

Those three payoffs, satisfaction, narrative agreement, and discouraging dull play, are real. But difficulty remains a tool, and a tool can be swung badly.

Crude difficulty is cheap to make. Line the walls with spikes. Restart the player from the beginning on every death. Tuck enemies where nobody can see them coming. All of that is harder, and most of your players will leave.

The useful version is a balance. You want the player challenged, possibly frustrated, but still able to see the route to the win. They should feel themselves improving. They should be able to name the reason they died. They should get enough chances to adjust and go again.

Hold that balance and a game keeps players inside the challenge zone for a long stretch. A brutal boss stays gripping as long as attempts run longer each time, mistakes stay legible, and the finish keeps looking reachable.

Unfairness swaps out the emotion. The player stops feeling tested and starts feeling annoyed, cheated, or worn down.

Silksong hits that line at points. Long repetitive walks back to a boss door. Unpredictable troll moves, heavily random attack patterns, and secondary-weapon ammo that does not come back on death, so you go farm enemies before you can seriously try again.

Friction like that shoves players out of the challenge zone and toward the quit button. It also cancels out the first payoff. Quit and you never get the catharsis at all. Win and you may feel nothing sharper than relief that it is finally done.

So it is a mistake to treat difficulty as one undivided quantity. A demanding boss pattern, a long runback, a harsh resource penalty, randomized behavior, unclear telegraphs, and strict checkpoints are six different changes to a game. Each one reshapes the player's relationship with failure in its own way.

One of the strongest tools is letting the player move the difficulty themselves. The obvious form is a difficulty mode or an assist option that alters the game directly, and Silksong, like Hollow Knight before it, ships without that menu.

The omission matters because difficulty is subjective. A fight that reads as demanding but plausible to one player is impossible for a second and trivial for a third. Options exist so those three people can each locate their own challenge zone.

And a game like Silksong is far more than its bosses. It is mapping an interconnected world, digging out secrets, piecing together lore, and being somewhere unfamiliar. Losing all of that to a single combat bottleneck is a bad trade.

Tunic makes the comparison well. Its secrets, riddles, and manual-decoding mystery can grip players who cannot get past its Souls-like bosses, and accessibility settings that soften combat are what let those players see the rest of it.

Silksong does offer one adjustment, even without a menu: you can leave.

When a boss turns sour, you can usually walk away, take another path, pick up new powers, swap charms, raise your maximum health, or find a tool that makes the fight easier when you come back. That structure gives difficulty room to breathe, because you are rarely pinned against one wall until it breaks.

The world carries plenty of optional bosses and relatively few hard bottlenecks. Even losing your body on death is gentler than some neighboring systems. Dropping rosary beads stings, but Hornet does not forfeit half her magic until she recovers the corpse, and there are ways to bank currency before a risky run.

The ammo rule may quietly serve the same purpose. Because secondary-weapon ammo does not refill on death, the game nudges you out of grinding the same boss forever. You go farm, or explore elsewhere, or cool off, or turn up something that tilts the fight.

More on this: Roguelike vs roguelite: permadeath names two rules

The healthiest kind of difficulty says more than try again until it works. It indicates a better way to play or a better plan to bring.

The indicated move might be studying the pattern properly. It might be switching tools. It might be exploring elsewhere and returning with more health, bringing a friend into the fight, hunting down a specific item, or stepping away for the night.

The difference is worth holding onto. A game can be demanding without being thoughtless. It can ask for persistence and still leave routes around the frustration. It can test skill and still accept that players do not all break at the same point.

Difficulty is a tool for building one specific experience. Handled well, it manufactures deep satisfaction, keeps the tone of a story intact, and pushes players into real engagement with the mechanics.

But it is really many tools wearing one name: boss health, damage numbers, healing windows, runbacks, randomness, checkpoints, resource recovery, optional routes, upgrades, assist settings, and more. Every one of them lands on the player differently.

So the aim is never to make the game hard. The aim is to work out what kind of challenge this game needs, and what feeling that challenge is supposed to produce. Then work out which release valves and adjustment paths keep players in the zone where frustration can still turn into triumph. You settle that by playing the build and judging it yourself, not by reading the numbers.

In the Flockbay app, pick the one scene that is allowed to be hard and make the deaths readable. Play until you can name the mistake. If you cannot, the scene is noise. Lower it, or add the tell. Do not raise the rest of the game to match a noisy fight.

An AI soulslike maker and a soulslike template are the strain. One fight.