Monday, August 31, 2009

Need More Bookmarks

In this article, programmer Nels Anderson argues that games should use a "bookmark" system instead of outdated save systems that waste the player's time.

Single-play sessions are difficult to design for, because exactly how long they will last is completely unknown. I've played Advance Wars: Days of Ruin for the half-minute or so it takes the elevator in my apartment building to get down to the laundry room floor. We ought to allow single-play sessions of any length, but the inertia of existing save system designs makes this difficult. I believe separating the two similar but distinct purposes of save systems will make this goal much easier to achieve.

Why are most pop songs approximately three to four minutes in length? Classical compositions are rarely that short. The answer is that the original vinyl 78s and 45s could hold about three to four minutes of music per side without sacrificing too much quality. What was originally a technological constraint became integrated into the way music is created. This technological legacy can still be seen today.

There are actually many facets of games that may result from obsolete technical constraints, but I'm going to address just one here- save systems. Specifically, this was inspired by something David Carlton wrote (with Randy Smith's MIGS presentation about saving before that).

The original purpose of save systems was to allow players to finish games that were too long to complete in a single sitting. The first console save systems didn't even use storage on the cartridge, they simple provided codes (or Castlevania's weird weapon grid) that loaded the game in a certain state. Since, they've become more complicated things with numerous implementations. They may or may not be encouraging negative compulsive behaviours (I tend toward yes on this one).

Unfortunately, the original purpose of save systems and the dynamics that emerge from the newer implementations have become confounded. Their original purpose ought to be broken out, leaving the rest of the save system to be addressed independently.

Simply, players should be able to stop playing a game at any point without fear of losing significant progress. To do anything else is to be disrespectful of your audience's time. It's absurd to require the player to wall off a section of their day to play your game. No other in-home media does this and there's no reason why games should get a pass. (As an aside, all cutscenes should also be pausable. Period.)

I like the Metroid series a lot, but I still haven't played Metroid Prime 3: Corruption yet almost exclusively because of its save system. Maybe MP3 isn't as extreme, but previous iterations required about a solid hour of play to make serious progress. Otherwise it would seem like a waste of time, with much of it spent trekking out and back to the save points.

If having that save anywhere/anytime is problematic to the game's design, providing a single "bookmark" save slot that is deleted after it's loaded is sufficient. Many DS games provide this functionality and as someone who has recently implemented a save system, it's not hard to do.

It doesn't even need to load the game exactly as it was saved, but something reasonably close ought to be ubiquitous. David discusses that save/load systems that force repetition of content as a punishment for failure are the reason why he'll save compulsively. And I could not agree more. If you're forcing players to repeat swaths of your game as a consequence for failure, something has gone off the rails.

Playing Little King's Story recently, I can't help but feel that their save system is unnecessarily punishing. You can only save in one place, there are times when you simply cannot save at all and if you fail in combat, you're immediately booted back to your last save. There was a point where I wanted to stop playing LKS but had to continue for about another twenty minutes or be forced to abandon content that could not be restored. And this was just after a boss fight! I can already tell that the LKS save system is going to force me into the compulsions Randy describes because the penalty for not saving is so extreme.

Japanese developers seem worse about this than NA/European studios, but this problem appears everywhere. We've all seen it. It's the save point just before massive, unskippable cutscenes that rolls immediately into a very difficult boss fight. It's the failure that forces you to perform the exact same series of actions again and again. These things don't make the game more challenging, they don't make it more interesting, they simply make the game more frustrating.

Having a save mechanism that's respectful of your player's time ought to come easily if you're empathizing with them. Provide a solution for the original problem the save system was meant to address. Beyond that, we can experiment more with save systems, looking for ways to move away from compulsive save/load behaviour. But unless players believe they can be confident progress won't be lost to punishing save systems, we're never going to move past save/load OCD.

Nels Anderson is a programmer at Hothead Games. He's probably the only game developer in Vancouver (and maybe all of Canada) that was born and raised in Wyoming. He writes about games and game design on his blog, Above 49. A version of this article appeared there.


Adrian Lopez said...

Game designers should learn to uncouple what I call a game's failure response from a game's save/reload mechanism.

Failure response is simply how the game responds to player failure, and loss of progress upon failure is a common and legitimate mechanic. It is, however, usually a mistake, to restore the game to the latest save state in response player failure.

Without justification to the contrary, failure response should be explicitly designed into a game rather than be left to the arbitrary nature of a game's save mechanism.

Post a Comment