💾 Archived View for yujiri.xyz › game-design › saving.gmi captured on 2023-09-08 at 16:29:23. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2023-03-20)

➡️ Next capture (2024-05-26)

🚧 View Differences

-=-=-=-=-=-=-

yujiri.xyz

Game Design

How to save the player's progress correctly

Let's talk about saving philosophy: some games let you save whenever you want (even during a fight), others only at designated points, others only have autosaves and no backtracking. Games also include different things in what gets saved such as consumable items, etc.

Proper saving policy revolves around the concept of a *section*. A section is a self-contained challenge - you start it with full status (health, stamina, mana, or whatever analogous stuff the game has) and your status is restored after it. Every section should have a save point before and after it. You should not save during sections. If a player needs this to get through the section, then either let them lower the difficulty instead, or your section is too long.

Difficulty settings

Cheap difficulty

And please, make boss fights their own sections. There's a massive psychological separation around them, and that must be respected. It's like if the the last two chapters in a novel didn't have a divide between them. If your boss isn't hard enough to be a self-contained challenge (yes I've heard Dark Souls fans raise this argument), then make *the boss* harder, idiot.

That also means don't include long treks to the arena or whatnot as part of the section. Making the player repeat a 2 minute run every time they die isn't "hardcore" or challenging (and yes, I'm talking about Dark Souls again). It's wasting the player's time. Cut that shit out.

Final Fantasy 13 is a good example of the encounter-section progress-saving model. Your party is automatically healed between fights, and when you die, you go right back to the field just before you entered the fight, preserving your progress in the area and letting you reconfigure your team, equipment, or paradigm deck before trying again, or even give up on the fight and go somewhere else. But the enemies are restored too, so it's not abusable.

I can't actually think of a spotless implementation that I've played of the level-section progress-saving model. Halo saves during missions without restoring your status or letting you revert more than one save back, so you can get stuck in a near-impossible situation because you spent all your ammo for any decent weapons and then the game saved on you and now you can't fix your mistake. Call of Duty: Finest Hour takes any unused health packs you were carrying from before the checkpoint when you die (so you can't recover them). Any game that saves but doesn't restore your status when you get to the save point, only when you die and get sent back to it, is unfit for praise here even though it's not as bad as the other problems. Etc.

You might ask, given my other claim that level grinding should not even be offered to the player, whether I think it's impossible to give the player consumable items that can only be used in a single section. No. These things don't count as status. (They still have to be restored on dying after using them, though.)

Why you shouldn't have persistent consequences for failure

The obvious problem is that the player might use all of them before getting to a boss where they need them and then not have any or be able to get any. But there's an (almost) equally obvious solution, which for some reason I've never seen used. That is to let the player replay old sections (with their old stats), and if they re-complete it without using the items, then they go back to the present as if they never used the items there. This also removes the problem a lot of games suffer from of the massive disincentive to ever use these items - if they're irrecoverable after use and the player doesn't know what's around the corner, they won't use rare items when they otherwise would and when it would give them a more consistent difficulty curve.