Designing savegame systems

Since we've previously established that the design of the save game system impacts gameplay, even if it requirres specific behaviours by the player to do so, I think it's in order to discuss different behaviors in relation to different save game systems. Before we get that far though, we need to establish just how difficult the subject is to work with, how it can possibly fit within my pattern model, and generally how we should procede from there.

I hope to narrow the subject towards a somewhat monolithic conclusion, but I fear it will be far more divergent than any of the other patterns I've discussed. I don't think it'll turn into a nice uniform set of rules, but rather several sets of conditional rules, that can be followed depending upon what exactly you're trying to achieve. The key to the difficulty, in my view, is the heterogeniety of the player behaviour. Perhaps I should embark on explaining why exactly save game systems are different from other gameplay mechanics, and perhaps more importantly, on how this difference is reflected in the way players tend to percieve save game systems.

Save game systems are technically different from everything else because they can be invoked at any one time during the game, and because they completely change the state of the game. None of the other metafeatures have this effect. Changing your options and configurations do not, as a general rule, effect the play experience; but there are a few exceptions to this rule, if for example you play a timed game against a chess computer you can underclock the cpu that's driving it to make the experience easier. Or, if you get a faster mouse and monitor for an FPS game, or a game with quick time events, it should improve your reaction time by as much as a 10th of a second. And of course any difficulty settings.
And then, there's the save/load feature, the focus of this blog, which alters the games state completely, at any time. Unless there isn't one, in which it simply impacts the players life, if he needs to leave the game, instead. Either way, once invoked, either by shutting the game down because there isn't one, or by saving it because there is, once the player returns, the game will be altered. Either because it allows the player to start from that point regardless of what happens over and over, or because it forces him to start from the beginning again.

Obiously, it's very difficult to anticipate the invocation of the feature, since it's invocation may not relate directly to anything that is within the game as a product. In fact, it's very normal not to respond to the invocation at all, for that very reason, nomatter where or how it occours.
This relies on a, by now completely ingrained, very specific suspension of disbelief amongst players.
Anecdotally, when disguised as anything but a save-game system, this feature would be looked upon as being completely different; a common system could be percieved as an undo-button which you can press if you anticipated that you might need to press it ahead of time. Omnipressence of this feature, this ever-pressent choice, within the game is at the heart of why it is very difficult to work with. But the bulk of the trouble in dealing with it in any innovative way is the vague notion that the feature is to be treated as special and somehow "outside" the game, which has by now festered and taken root in the public consciousness amongst players, even though it can clearly alter the experience very considerably.

Choosing wether or not to cooperate with the notion is not a difficult choice to make on the surface, as the advantages are many and ample. Many players will appreciate incorporating a traditional save game feature because it makes matching their lives with their gaming. This will largely be casual gamers, and if anybody uses it to improve their abilities within the game, it doesn't really relate to them because those who have the largest benefit from a save-at-any-time system are likely so casual that they wouldn't bother. The most hardcore gamers, of course, will likely choose to limit themselves from abusing the system because they want to play the way it was meant to be played, and somehow they feel that not using the save-game system to their advantage is more "real".

But the downside is that, somewhere in between, there will be a demographic that will likely use it to alter their game state in various ways, and it's incorporation will alter their play experience significantly. For this reason, what to do is not obvious, and because the demographic is difficult to pinpoint with any accuracy, it's difficult to say wether or not it is safe to ignore it. If it composes 75% of players, their reasoning behind using the feature to min-max their chance of doing well should be of no concern. They're your target audience, and you should try to maximize the fun they have; that's why they pay you. Not to fuel your artistic expression, but because they want to have fun. Unless you want to deny them that, you must try to cater to them.

But then...is it really worth maximizing their fun by doing something that will probably also inconvinience 30% of your players greatly? The choice is bipolar. But to me, the conclusion of this discussion is clear: If you want to design good, solid games, you'll put care into make the right decision about what save game system you choose for your game, and you'll take both the fun of your audience, the current perception of the save game feature, and the non-linearity of your game into account.

Now that I feel I've argued how to best approach the problem, I want to go over some different save game systems; I'll probably only blog about one today.

The first system is the no-save.

This is often used with arcade shmup games; it's impossible to save the state of the game, often because games in arcades are simple, easy to play, and revolve around learning level patterns in order to optimize and anticipate events that always occour in the same manner. Also, player death is directly what earns the arcades money, so if the player could skip over difficult sections after completing them once, it would be a less efficient manner of getting the players cash.
Another advantage is that playing the same levels over and over again is how you get good at many shmup games - there is no arbitrary skillset, only how good you are at the specific levels. This no-save system, then, lines up perfectly with what the game is trying to do - make the player feel as if he's constantly improving, getting better. But it comes with many drawbacks too, one of them being that when not alligning

It seems that games with more than 20 hours worth of playtime would be impossible to complete for any single player, as that appears to be the point where most get sick without rest. There have been reports of heart attack and similar ails setting in after 36-50 hours of playtime, for instance.

So there is a theoretical limit to the no-save system, as well as a, presumably, much lower practical limit. An estimate for the practical limit could be as low as perhaps 2-3 hours.

This is where the players behavior comes into play however. Some players will only be able to accept a sessionlength of 10 minutes or less for a no-save system game, whereas some players will be willing to accept much longer sessions without the ability to save their games.

This is perhaps one of the largest problems I've encountered so far in my work on game design patterns. Which demographic groups are able to accept which playsession lengths, and which invonviniences? And if so, does a particular type of content, which attracts a particular demographic group, need to be paired up with a particular kind of save game system in order to optimize the design? And how exactly should your demographic distribution be before you should cater to the convinience of one part, or the fun of another?

Naturally, my greatest interest is within one particular type of game - those with adaptive narratives and non-linear storylines. Sadly, those do not even fall within any one of the commonly accepted genres - and they make up only a small minority of those genres they do fall within. So there's no real chance of finding empirical evidence one way or another.

Instead, I have to approach the problem using only anecdotal evidence collected from various demographic groups. And in order to really capture the possibilities, I'll of course have to go in depth with examples from different genres and what kind of concequences game systems can have.

Oy vey. This pattern will take a while.