A manual cannot stop a reload

The manual for The Elder Scrolls II: Daggerfall, Bethesda, 1996, asked players to steer clear of replaying the save. It is a charming request, and it did not work. Text cannot compete with a load option that sits one button away and erases a mistake for free. Players are not disobedient. They are using the tool you shipped.

If you want a mistake to stay made, the rules have to prefer that outcome. Ironman, a cost that survives the reload, a world that has already changed while you were deciding. Or you accept the reload and design the interesting failure to happen before it, so the lesson lands in the thirty seconds you cannot bear to play again. Begging in a manual is not a third option. It is a wish.

You cannot fix save scumming with a paragraph of text. The manual for The Elder Scrolls II: Daggerfall asked players to steer clear of the "replay the save game" strategy, and the habit survived anyway.

The habit is familiar to anyone who has played a computer RPG. You save constantly. A quest turns sour, a pickpocket attempt fails, a random roll goes against you, and you load again until the record shows nothing but lucky breaks and correct calls.

The counter-argument is a good one. Role-playing is not a contest to produce the flawless run; it is about growing a character and letting a story accumulate. Much of the most interesting material shows up only when ordinary mishaps are allowed to land.

The problem is where the argument was printed. A designer with a strong view about how a game should be experienced cannot leave that view in the documentation. It has to live in the systems.

Loading a save after every stumble deletes the messy, unrepeatable stories games produce better than any other medium.

You drop to a sliver of health, turn the fight around, and win anyway. You put down a panicking guard a second before his hand reaches the alarm. A silent infiltration falls apart into a scrambling exit. A party keeps marching, one companion short.

None of those moments can exist without continuity. Each one requires the player to remain inside the same unbroken chain of events after something went badly.

So the design question is not whether players should reload. It is how to make staying more appealing than the quick load at the first sign of trouble.

The first answer is to make the setback survivable. If a single error is ruinous, reloading stops being cowardice and starts being the rational play.

The useful frame here is the failure spectrum: the full range of outcomes that sit between a flawless result and a total wipe.

XCOM has several. A mission can close with the squad untouched, with the objective met and half the team in the infirmary, with the objective lost and a scrappy retreat, or with nobody coming home. Each is a point along the same spectrum.

Some games leave almost no middle ground: you are either perfect or you are dead. Others fill that middle with states, which is what gives failure the space to turn into drama rather than a full stop.

Metal Gear Solid V is generous with that middle ground, and it builds the generosity out of layers.

A guard who notices something does not kill you on the spot. He goes to look. If he spots you, reflex mode hands you a narrow window to put him down. Miss it, and he still has to radio in, which opens one more chance to cut the alarm off.

Once the alarm does go up, Snake has options: run, go to ground, fight it out, or call in extraction and let go of the clean stealth run altogether.

Death sits at the end of all those layers, not at the start. A single slip therefore does not demand the load screen. It nudges you further along the spectrum, into a different kind of play.

A spectrum gets far more interesting when the player can work their way back up it toward success.

Far Cry 2 is built on that motion. It gives you a long health bar, healing syringes, and a buddy who can pull you out of death once. It also keeps hurling problems at your plan: guns jam, malaria hits at the worst possible second, vehicles die halfway through a getaway.

Those problems are not punishments in any simple sense. They are prompts to improvise.

You begin with a plan. Scout the outpost, read the patrol routes, pick your angle, work out how you leave. Then you commit. When something breaks, you are thrown back into planning while the mission is still live around you.

That swing between planning and executing is thrilling precisely because it stays inside one unbroken attempt. The mistake rewrites the immediate objective without ending the story.

Far Cry 2's setbacks earn their place by being minor, hard to anticipate, and possible to come back from.

A jam can ruin an approach you spent five minutes setting up and still not be worth restarting for. A malaria attack reorders your priorities without ending you. A dead engine turns a tidy escape into a frantic one.

The loss is too small, and too random, for save scumming to feel like a good use of your time. Adapting is simply the cheaper option.

Adapting is often the best part of the session. Taking a hit swaps the goal from winning the fight to reaching cover, patching yourself up, or building a medkit. Getting seen swaps it from staying invisible to slipping away, going quiet, or committing to a firefight.

Every one of those goal changes depends on the player still being there after the plan came apart.

All of this collapses if the game leans hard on rewarding perfection or punishing error.

A cosmetic rank or an achievement for going unseen is generally harmless. It hands expert players something to chase without declaring the imperfect run invalid.

The trouble starts when a mistake makes every hour that follows measurably harder. At that point reloading is the sensible response, and players will reload.

XCOM is the cautionary case. Veterans who die are replaced by fragile rookies, the next mission is harder because of it, and a harder mission produces more casualties. Failure feeds itself.

Under those conditions, save scumming is not weakness of will. It reads as self-defense against a campaign that could unravel entirely.

A better answer than tolerating failure is making failure genuinely worth having.

The Shadow of Mordor games manage it with the Nemesis system. Kill the player as an orc captain and that captain remembers it, and may bring it up when you meet again. Dying stops being an interruption and becomes part of a relationship.

That flips the incentive. When imperfect play generates consequences you want to see, and a flawless run is not the only tactically sound route, players are far more willing to let the moment stand.

The principle underneath is that failure should not just take something away. It should change the shape of the situation, add texture, or make the world look like it noticed.

The bluntest fix is to remove the option: no old save file to go back to.

Darkest Dungeon writes over your file continuously, so rewinding a poor decision or a cruel roll is close to impossible. It suits the game. Ugly outcomes arrive and you live with them.

The side effect is that risk starts to mean something. If an old save is always waiting, the question "do I push deeper for better loot?" has no weight behind it. Permanence is what makes the choice heavy.

The Long Dark works the same nerve from another angle, writing a save at the moment things go wrong, such as an injury or an animal attack. You come back to the crisis, not to the calm minute before it.

A game can respect the fact that players have lives without handing them a rewind machine.

Dark Souls separates permanent progress from suspension. You can stop and pick up where you left off, but that temporary save is consumed when you use it. You can walk away without every decision becoming a reversible experiment.

That is a distinction worth copying. A save system can accommodate real life and still protect consequence.

The aim is never to make saving annoying for its own sake. The aim is to stop the save system from becoming the strongest available strategy the moment the game gets difficult.

There are real reasons to keep players from jumping back the second a plan fails.

Mistakes force the goal to shift mid-play. They generate stories a perfect run could never reach. They give risky decisions weight, because the decision cannot quietly be taken back.

What you cannot do is ask players to enforce that ideal on themselves. Given a button that removes the rough edges, most of us press it, and the rough edges were where the memorable moments lived.

So the intention has to be built: close off cheap save scumming, keep setbacks survivable, resist over-punishing imperfection, and give failure as much interest as success. You find out whether you managed it by playing the build and watching where you reach for the load screen.

Get that right and players will let the moment run, because they actually want to see what comes next.

In the Flockbay app, build a short scene with a reload and a version where the reload keeps one consequence. Play the mistake in both. If you mash reload in the first and hesitate in the second, the consequence is real. Keep that one.

An AI RPG maker and a story RPG template will include a save by default. Say what the save is not allowed to undo.