The rules governing when a player may save change how a role-playing game feels far more than most players expect. The save system is effectively a control on tension.
Saving anywhere removes the cost of failure
A player who can save at any moment loses at most a few seconds by dying, so risky choices become cheap and experimentation is essentially free.
That is comfortable and it flattens tension, because the possibility of losing progress is what makes a difficult encounter feel consequential.
Players also tend toward saving constantly once it is possible, which interrupts play and turns a convenience into a habit that they often dislike.
Checkpoints set the unit of risk
A checkpoint system defines how much progress is at stake in any given stretch, and the spacing of checkpoints is therefore a direct pacing decision.
Widely spaced checkpoints raise tension and raise frustration together, since a mistake costs material time and the replayed section is one the player has already solved.
Designers usually place checkpoints immediately before difficult encounters rather than after easy ones, so that repeated content is the interesting part.
Single-slot saving makes decisions permanent
Games that maintain one save and overwrite it prevent a player from returning to reverse a choice, which is how narrative consequence is enforced mechanically.
This makes decisions weigh more, and it is why the technique appears in games built around irreversible outcomes rather than around optimisation.
It also removes a common form of play in which a player explores several branches, so the design gains commitment and loses breadth of discovery.
Save restrictions interact badly with real life
A player interrupted by circumstances outside the game loses progress through no failure of their own, and this is a different experience from losing to a difficult encounter.
Suspend features answer this by storing an exact state that is deleted on resumption, separating the ability to stop from the ability to retry.
The distinction is now standard on consoles and handhelds, where sessions are frequently interrupted, and it has made strict save rules more tolerable than they once were.
Save scumming is a design response, not a defect
Where reloading can improve a random outcome, players will reload, and treating this as misbehaviour misreads what the system invited.
Games that dislike the practice remove the incentive instead, typically by fixing random results at the point they are generated so that reloading reproduces them.
That approach preserves the player's freedom to save while removing the reason to abuse it, which is why it has become the common solution.