There’s been a lot of discussion recently about choices in games, and the effect that game save mechanisms have on the ethical impact of those choices. I won’t even attempt to link to the vast majority of the conversation, but two contributions (both involving Nels Anderson) particularly struck me today: slides for a talk by Randy Smith called “How to Help Your Players Stop Saving All The Time” that Nels mentioned on twitter, and an Experience Points Podcast episode on “The Decision Dilemma”.

I’m an obsessive saver when I play games (though, fortunately, these days less obsessive a reloader than I used to be), but listening to Jorge, Scott, and Nels talk on the latter made me realize that many people save games for completely different reasons than I do. The typical scenario that they discussed is a player who saves a game right before a big choice in a game (typically a moral one) and then plays through the different branches, reloading as necessary, before deciding which route to commit to.

The thought of doing that almost never crosses my mind. (Especially if the choice is a moral one.) And when it does, I reject it out of hand. For example, when playing Mass Effect, I wasn’t really thinking too hard when going through the dialogue tree that leads to a choice of which party member dies. I ended up inadvertently choosing to save the party member I liked less; once I realized that, I could have reloaded and not lost much time, but instead I felt a pang of regret and continued playing. (Though, to be sure, I’m not sure I made the “wrong” choice even there—it struck me as the sort of choice that, in the real world, I would want to not make based on personal likes and dislikes, and it’s not clear to me that other factors wouldn’t have swayed me to make the choice I actually made in game.)

Instead, the reasons why I save are quite different: I save because I don’t want to spend time doing stuff that I don’t enjoy. I do not want to have to fight through a stretch of the game, to die, and to have to replay that section. (Unless, of course, it’s a game whose mechanics I’m particularly fond of.) Perhaps worse, I do not want to survive the next section of the game but end up in a weakened state, making battles half an hour later much more difficult. (And probably requiring extra reloading when I reach them!) And, of course, the absolute worst is when I survive by avoiding encounters that would otherwise have given me experience points, forcing me to repeat battles (or grind in order to level up) for the entire rest of the game.

The above mostly plays out in tactical situations. If the outcome of a battle went well enough and it’s easy enough and fast enough to save, then I will typically save; if it went badly enough, I will typically reload; and if it’s in the middle, I’ll play along for a while longer. As I said above, the more strategic choices are much less likely to make me think seriously about reloading; and, even when I’m nervous about a choice, it’s not usually an ethical choice, it’s much more likely to be a choice about which branch of a skill tree to improve my character in.

So the podcast episode was, to me, more a glimpse into other people’s minds than anything that spoke to me directly. Randy Smith’s slides, however, were a different matter—indeed, right near the beginning he talks about frequent saves being driven by a need for safety, which is a good match for my feelings. (Though he branches out into other motivations later in the slide deck.)

I wish I’d heard him actually give the talk; I’m having a hard time grasping the nuances just from the slides. He ends the first part of the talk with a claim that “reducing compulsion is good, regardless of save/load design”, which I tend to agree with—in particular, I don’t claim that my obsessive saving is a good thing, in fact I’m willing to accept that it’s a bad thing. (E.g. because, as he says, it takes my attention outside the game.) I also agree with him that cheap save/load sets up a feedback loop encouraging people to do so more often. (I played Doom rather differently from Marathon, for example.)

It’s not clear to me, however, that I prefer for games to solve this problem by limiting the contexts in which you could save: I’d much prefer to solve the problem by limiting the lack of safety that drives me to save in the first place. (Though I’m also willing to believe that this is a false choice, and that a deeper analysis would lead to a more satisfying resolution. In fact, I’m willing to believe that, if I’d been at Randy’s talk, I would understand him as doing exactly that sort of deeper analysis!)

Consider the basic choices that I outlined above: if I don’t save and then die, or if I do save but don’t reload after a stretch in which I played badly, then I will get punished for my actions, with that punishment lasting in some cases for the entire rest of the game. In other words: if I play badly (or even less than perfectly), the game will reduce my enjoyment of the game for hours to come.

This is a lousy way to treat players. If the game really is about challenging the player’s skills, then of course you want bad play to have consequences; such a game, however, then it should hedge its bets in two ways, both by putting skill-driven play front and center and by having individual bouts be bounded with no lasting in-game consequences from one bout to the next. (E.g. puzzle games, rhythm games, fighting games, multiplayer FPS games.) But if you want your game to have a long-term flow, then don’t treat your players this way.

So: at the very least, bound the negative consequences. There are lots of tactics for doing this; checkpoints are a tried-and-true one, but I also rather like the Zelda technique of both having the game be kind enough that death is relatively rare and having the consequences of death be limited to needing to refill your hearts / bombs / arrows (all of which are available from clumps of grass) and perhaps needing to traverse a part of a dungeon. (Usually a small part: in particular, Zelda dungeons are generally good about giving you a shortcut from the entrance to the boss fight once you’ve gone through the rest of the dungeon once.) (Incidentally, I think part of people’s dissatisfaction with Majora’s Mask is in the ways in which this principle doesn’t hold, or at least doesn’t manifest itself in the same fashion as it does in other games in the series.)

I also really enjoyed the way Lego Star Wars handled this issue: in that game, your character has almost no state at all, which means that the game can simply respawn you when you die. A skill-based player can still take pride in rarely dying when progressing through a level; other people can have no end of fun by simply mashing buttons.

Another issue around saving and loading is the lack of information: most of the time (pre-boss save points being an exception), I save not because I know I’m likely to die in the upcoming area or because I’m likely to play in a sub-optimal manner, but rather because I want to limit my losses in the face of an uncertain probability of death. And, perhaps more interestingly, the reason why I reload isn’t that I know that I played sub-optimally in a fashion that will hurt me down the road, it’s because I know that I played sub-optimally and I don’t know what the consequences of that will be.

If you treat this simply as an information problem, it can be significantly improved without harming gameplay. Start with the reload problem: that shows up most starkly if the game doesn’t put any limits on your capabilities (e.g. your health, your ammo supply, the level of your character). In that situation, if you do anything suboptimal (e.g. miss a single shot!), you may fear that it will hurt you going forward.

If, however, you have caps on these attributes, this problem goes away. For example, if there’s a maximum amount of ammo that you can hold, then if it takes you three shots to kill an enemy whom you could have killed with two shots, and if you subsequently pick up enough ammo that you’re at the ammo limit even after wasting that shot, then you know that the missed shot didn’t hurt you. Concretely, my worry level in Deus Ex declined notably once I started hitting limits of this sort.

These sorts of attribute caps are most effective in directly attacking the problem of when to reload, but they also help with the problem of needing to save in the first place. Your character’s status with respect to various attribute caps give you a concrete way of measuring how vulnerable your character is; assuming that the game has earned your trust that it won’t throw major challenges at you without some advance warning, this will frequently allow you to avoid saving without seriously worrying that doing so will hurt you.

I’d love to see more games with significant moral choices. And I’d be delighted to have not saving be a part of that, as long as that doesn’t destroy my enjoyment of the game in a more mundane aspect.

Post Revisions:

This post has not been revised since publication.