To Save, Or Not to Save

How much control should the player be afforded over save games? Is this an area of game design in which the player should be fully empowered, or an element of the architecture of games which should be abstracted away to sit seamlessly in the background?

Although Ernest and I work as part of the same company framework, we mostly interact virtually, and only face to face at numerous conferences and speaking engagements throughout the year. We have very compatible views on the games industry, but still there are always some disagreements. In his recent Bill of Player's Rights (inspired in part by Peterb's  piece), Ernest suggests the following right:

The Right to Choose Not to Save the Game. Some games give the player only one save point, and the game saves to it automatically. Occasionally, the game does this beyond a point at which the player will inevitably lose the game—which means the previous save is lost and the current one is worthless. This is evil, bad, and wrong. Besides, a good many players like to establish an extra challenge for themselves of making it through a tough spot without saving. Players should be able to save when they want, and should not have to save until they want.

I disagree with this. Not the spirit of the clause - obviously the player should never be able to get into an untenable situation - but I do not support the player's right to choose not to save (at least, not in games which hope to reach the Casual market).

Let me pause briefly to point out that I am wholeheartedly in agreement with Ernest's "Right to Quit, Pause, Save and Resume the Game". I believe that a bookmark save is an essential element these days - the player should be able to stop playing at any point. But, crucially, I do not believe that the player should be able to resume play repeatedly from that save. Bookmark saves that delete themselves after use strike me as a perfectly legitimate measure.

Why dispute the player's right to not save? Firstly, because a Bill of Rights must be something which all games should adhere to. But games such as Animal Crossing are undermined if the player can avoid saving (this particular game takes amusingly elaborate steps to chastise the player if they try and go against this!)  Therefore, the right not to save could potentially act as a barrier to play invention, which I cannot support.

However, there's a bigger issue in the context of player needs. The game literacy of Casual players is considerably lower than the typical core gamer - save games can be a confusing element of some games, and a potential barrier to Casual play. For instance, when I play games with my wife (who is quite game literate) she will still have no part in decisions about how and when to save - because she wants to be playing the game, not playing the game of when to save.

I suspect the reason that Nintendo has settled upon its system of offering a save slot which is then automatically saved to is that their user testing has shown than anyone can deal with this system. I freely admit, however, that this approach does count out certain styles of play - in particular the logistical quick save style of play.

This, for those not familiar, is how a player goes about completing a game that is badly designed, or designed specifically to be completed in this manner, or at the very least open to being completed in this way.  In essence, you play for a while, then when you detect you are temporarily 'safe' and have made some progress, you quick save. Gradually, you optimise towards a solution for the level. Saving and loading, in this regard, has become a game in itself - albeit a game which can only be enjoyed in a strategic-logistical fashion.I don't believe we can or should stop the above form of play, since it seems there are players who actually do enjoy playing this way. But this style of play is completely anaethmatic to the Casual audience as far as I can tell.

Additionally, there is the problem of creating the implicit obligation to reload. This happens when a game creates a situation from which a reload is by far the best option. For instance, in GTA, you generally lose all your weapons and equipment when you die or are arrested, hence there is often an implicit obligation to reload. Similarly, if a game allows you to place big gambles, there is an implicit obligation to reload for any player who wants to continue to make progress in the event that they lose. Note that in both these instances it is not that the player must reload - it is that there is significant advantage to do so. And we should expect many if not most players to take any course of action that gives them an advantage.

In the case of GTA, the game design team presumably decided that there had to be a cost to failure. Hence, a small fee and loss of weapons. Except that much of the time, what they did was create a very different cost to failure: the implicit obligation to reload. In this instance, I don't believe they needed to add to the cost of failure - since the player already has to travel to the mission initiation point, and so has acquired a cost in terms of travel, and a psychological cost in terms of having failed in the first place. This was sufficient. Adding the additional costs just created an implicit obligation to reload.

In the case of gambling, the aleatory play of Juiced fails to some great degree as many players work out that they can remove their memory card to prevent the autosave.  As a result, an implicit obligation to do so emerges for many players - instead of creating a game which captures the essence of gambling on car races (which probably wasn't a good idea to begin with, to be honest) they create a game which many players are forced to spend some of their time physically interacting with their hardware instead of enjoying the game itself.

I don't believe in implicit obligations to reload. Where possible, I believe that they should be factored out. This is because I don't see the play associated with saving and loading to be as engaging, rewarding or interesting to most players as the actual play of the game - although with certain badly designed games, it can get quite marginal!

Ernest's viewpoint is underpinned by his belief that the game designer shouldn't take steps to force the player into a particular course of action, on the grounds that the player bought the game and should be able to do whatever they want. I sympathise with this viewpoint, but I have always somewhat disagreed. (Ernest is slightly more liberal than me, perhaps!) I have seen first hand that when players find a way to take advantage of a game, they will do so - even if exploiting this flaw or imbalance destroys what is enjoyable to them about the gameplay. The drive to progress or conquer is so strong in many players, that it trumps any notion of 'fair play' - after all, one's opponent isn't another person worthy of honour, but just the computer.

I think that one of the reasons that Ernest and I are in slightly different camps with regards to save games is because Ernest primarily plays PC games, and I primarily play console games. Player control of save games is much more relevant for PC games, where a hard-drive gives you the capacity to store effectively infinite numbers of save games, than on a console, where memory for saves is at a premium. (In this regard, the original Xbox might perhaps be considered a user-friendly PC). That consoles may eventually acquire hard-drives as standard is, I believe, besides the point. Consoles are the home of the mass market of games, and they need to reflect mass market issues in the design of their games.

I believe it is reasonable to take away the player's right not to save, but in exchange the game should offer the player the capacity to opt out of save games entirely. Indeed, for the mass market, I believe that games should be designed to minimise the extent to which saving and loading of games impinges on the game itself. Mimicry, ilinx and alea can all be undermined by player control of save games; only agon (competition) can seem to benefit, and then only tangentially.

I believe that Hardcore gamers have become accustomed to treating save games as part of the play of their games - but that mass market games should not make the same mistake. We should focus on the play of the game itself, not on the play inherent to the state space mechanics of saving.  We should be sensitive to a Casual player's desire not to be burdened with decisions at the meta-level, about when and how and whether to save, but rather to lose themselves in the joy of the game they are playing, and have faith that we the game designer have taken care of their needs invisibly, without them even knowing what we have done.

I'm sure we've all got ourselves into trouble with games that give us control of save games - saving the game in a situation which is completely or partially untenable. It happened to me only last month with Rygar: The Legendary Adventure, when I inadvertently saved at the foot of the final boss with inadequate provisions- thus destroying the save which could have saved me. I already had the right not to save, though - I just made a strategic mistake because I did not know what was about to happen. Isn't this always the case? That because we do not know what is coming ahead, we can never solve the problem by simply giving the player the right not to save.

Games like The Legend of Zelda series, which have save slots, do not ever present this problem - but they do obligate the player to tackle the design issues that result from an automated save mechanism. Although this implies some work on the part of the game designer, this is the sort of challenge every game designer should be willing to shoulder.

We are still free to make games with player-controlled save systems, but these games should perhaps be niche market titles. Perhaps we can even make games which default to automatic save management but which have the option for the player to take control, if that's what they want, although I'm uncertain this is necessary.

Let us aim to have the mechanics of saving become part of the seamless architecture of mass market games, and accept the design obligations this implies.

Comments

[Corvus]

I've started about nine comments and haven't been happy with any of them. I think I'll let this simmer a bit and dish it up on MBB if it seems tasty enough.

Good post, though!

[James O]

" I have seen first hand that when players find a way to take advantage of a game, they will do so - even if exploiting this flaw or imbalance destroys what is enjoyable to them about the gameplay."

On the other hand, by limiting the save structure you neccessarily are imposing limits on the player, discouraging them from experimentation. In save-anywhere games, I will sometimes make a back-up save, and then just play around with game mechanics (can I lure this guard through his own traps? Will he notice the pile of his comrades in the corner? What will he do if I sneak past him and then do weird things with the items in the room?) Games that have more limited save structures say to me that my personal experience is less important than the developer's experience. When my saving is handicapped, I am more likely to simply use the most obvious and rote tactics, rather than try something unusual.

One experience I have here was Gran Turismo 4. In the prior GT games, most players would save up cash, save their game, and then go through a buy car -> test drive car -> reload sequence of gameplay. In GT there is normally no way to test out a car like you could at a dealer, and I certainly don't want to blow several hundred thousand dollars on an NSX that might not have the acceleration I want without at least testing it - that's just unrealistic. So I would try a car, reset, try another car, reset, and go through my shopping car until I found a winner. Save/load was a fast process, so it was no burden on me. Life was good. Unfortunately, an autosave feature in GT4 prevented this. Autosave always activated on purchase of a new car, and people online who tried to remove their memory cards beforehand often suffered corruption or other errors. Now I would have to go into car buying blindly, which greatly detracted from the experience (especially given how many cars were available.)

Giving players the option to save anywhere should not impact the experience of players who are willing to just use the built-in autosave/checkpoint system. No one is forced to use the extra save functionality if they feel it will hurt the game experience. If the player has freedom to save whenever, it might expose poor game design; but it is not itself responsible for the design being poor.

An example here might be Juiced - a game design based around LOSING as a center point of interest should immediately set off red flags for the developer. If it was a PvP style game I might tolerate that style of play more, because it is another player that bested me. However, if it's the heartless, rubberband cheating AI stealing my car (that it has no use for), I would be rather frustrated. The problem here was bad design, the fact that players could enjoy it more without the save limitation just exposes it for all to see.

"Consoles are the home of the mass market of games, and they need to reflect mass market issues in the design of their games."

As near as I can tell, the mainstream of gamers hates not being in control of their own save files, or having to wander a map for hours searching for a save point. Memory cards are no longer a limitation with the PS2 card at a rather generous 8MB.

"It happened to me only last month with Rygar: The Legendary Adventure..."

At least in this situation, you could blame none but yourself (though it was certainly a tough pill to swallow.) However, imagine if that happened because of an autosave - you lost all your progress because some developer thought it would be clever to autosave without your permission there. Now not only would you be FUBAR'd, but it wouldn't even be your own fault! You would have lost your game due to poor design. The problem wasn't the manual save. The problem was the game allowed you to get to an untenable position with no way to backtrack. The save issue just highlighted it in a painful way.

As a libertarian, it is important to me philosophically to allow the player as much freedom as possible. Handicapping their save files is just low and unneccessary. Any game play enhancements conferred by revoking save control can still be simulated by simply choosing not to save, and to rely on an autosave system instead. The casual players won't engage in meta-play save structures as they are casual and are probably not interested in the try-reload play anyhow. The hardcore, however, will want that option, and by removing it you are only hurting, not helping, the audiences. Saving should ultimately be a choice for the player, who is actually playing, and not the developer, who must simply hope the player is playing the way he anticipated.

[James O]

I was thinking about this a little more, about games that have a lose-continue mechanic where the player can lose something but continue on (e.g. Fire Emblem character death, Juiced car loss). Geting a player to accept certain losses requires deeper engagement with the magic circle. I would posit that embracing, not being repelled by, these losses, would imply a MORE hardcore player (but along a different axis of core-ism.) A casual player would (I think) be less engaged with the magic circle and thus less willing to accept loss and hence more prone to a fail/reload gameplay. So therefore a bookmark-save-only style game would be better for the hardcore whereas a save-anywhere style game would be more appropriate for casual players (and personal observations confirm the latter portion.)

Also, I got your book today, so I shall be reading it soon! :)

[Chris]

Thanks for this detailed look at the other side of the coin on this. I am reasonably convinced by your argument that access to a full save functionality could be of significant importance to a certain proportion of the Hardcore. Unless you're off playing devil's advocate, again - in which case I demand to see your cards. :)

It's fascinating that you characterise it as 'handicapping their save files' - you really *do* want and expect to be able to control the save files, don't you... For me, I'm quite glad to be free of the burden most of the time. There might be some deeper issue at work here that we haven't got to the bottom of yet.

It still strikes me that the save game functionality would do better to begin as a seamless part of the architecture... The Hardcore players are generally game literate enough to find things in the menus and options, so the game could (should?) default to a hermetic auto-save configuration, but allow the player to strip away this safety shell, if they would prefer to do so. (Again, I reiterate this is only really an issue in mass market games.)

Presumably, this ceases to be an issue if the play is being delivered in very short chunks? Less than 10 minutes, say. There must be some threshold whereby the ability to explore the game space via saving becomes irrelevant because the play segments are too small, musn't there? What do you think?

[Chris]

I'm curious... why would it be the case that the Hardcore player (who is characterised primarily by playing more games) would be more prone, rather than less prone to be carried along by the Magic Circle? I would suggest the visitor is more prone to being carried along by the play experience than the experienced player, who has learned they can push against the walls a little. (In some cases, a lot!)

And thanks for buying the book! :)

[James O]

For me, I prefer that the game can intelligently manage saving for me at appropriate intervals - for example, Halflife 2 had very natural feeling checkpoints to me, and only rarely did I feel the need to manually save. Having to exit to a menu to manually save tends to break the experience, but I want the option to be there for the occasions I need it - for example, if one particular part of the game is especially fun for me, I want to be able to make a save file right before it so I can always access that point of a game (this can also include parts right before a neat cutscene that cannot otherwise be replayed.) I also like having a manual save at an important juncture in the game (whether it relates to picking one of two plot paths, or picking upgrade a vs upgrade b). Some developers dislike this, saying it makes the game too easy; although honestly that's a stupid way to add "replayability" to the game (forcing you to replay it completely to choose different options.) If the game is good enough, I WILL replay it fully to choose a new path; however, in the event that would require replaying many tedious parts, I like being able to set up my own personal bookmarks that I can revisit whenever I wish.

For me, the optimum save structure:

1) Has automatic checkpoints throughout the game in particular safe areas (transitions from one level to another, after completion of significant objectives, etc.) These checkpoints are logged seperately in an index file. On a larger PC harddisk each save could be held indiviually, but on a console I could at least have a "Chapter selection" menu where I could revisit each checkpoint with a "standard" preset loadout. Maybe Mission 5 Part 2C was especially fun for me. A DVD style chapter/subchapter menu allows me to revisit those parts quickly and easily.

2) A quicksave/quickload button. Some developers resent the "save point crawl." That's garbage. If I have to save point crawl, then the developer fouled up in his level design. If the developer put me in a situation where I have 10 bullets for 10 enemies, and cannot afford to miss once, then it is his fault, not mine, that I will be forced to endure a tedious and frustrating save point crawl. Quicksaves (set to a special quicksave slot) allow me to make a backup before a dangerous encounter without breaking the game flow. If I'm playing Splinter Cell, and the success of the mission depends on me sneaking through a lengthy room undetected, I will quicksave. I shouldn't be forced to replay the parts leading up to it over and over; that's just frustrating. Quicksaves should be available at any time and place, even if the quicksave is untenable (checkpoints and the next type of save mitigate this.) If a quicksave becomes unplayable, I can fallback to a previous save. One thing that is critical is that the quicksave be on a seperate slot from level change autosaves/checkpoint saves. I don't mind a quicksave being a potentially unreliable type of save, but I should have fallback options in the event I quicksave at an inopportune time.

3.) Finally, a manual save (from a menu) should be available for when I want to make my own checkpoints. The manual saves can be restricted to certain safe areas or have some other features (if neccessary) that ensures they are always safe to play from (e.g. the Zelda style save system, where a save in a dungeon saves all progress but simply starts Link at the entrance of the dungeon.)

I think a fair balance can be achieved in save structures by balancing user freedom with safety - some of these save types limit player freedom, but in return are very safe to replay from. However, the option of a potentially unsafe, yet immediate, quicksave balances the equation.

If gameplay is in small enough chunks, then midway saving isn't neccessary all the time. I don't need to midpoint save while in the middle of a short race or Katamari session. However, in an endurance race (especially the punishing 24 Hours of Le Mans in GT4) absolutely must have a midpoint save (perhaps in the pit section), and the fact that GT4 omitted that was unforgivable. However, checkpoints can still work in short gameplay style games - witness Super Mario World, with it's midlevel checkpoint. It's not a permanent save, but if the player fails, they can retry the level from halfway through.

"I'm curious... why would it be the case that the Hardcore player (who is characterised primarily by playing more games) would be more prone, rather than less prone to be carried along by the Magic Circle?"

I assume here that the casual player might be more nervous to fully engage the magic circle, compared to a seasoned player. Specifically I had traditional pen/paper RPGs and LARPs in mind - LARPs are certainly not very casual games, and require very deep engagement with the magic circle. I assume here that casual players would be less likely to engage in that part of the magic circle, the roleplaying sort (Type 4?) However, this is why I qualified it as hardcore along a particular axis. I think a certain type of hardcore (say, the hardcore agon of a fighting game player) is less likely to open themselves to the magic circle than those of another type (the LARPers and SCA fellows.)

[Chris]

That seems like a coherent approach to saving. I'm glad you included a solution applicable to console games, otherwise your argument would have fallen slightly flat. :)

The chapter approach you allude to does indeed seem to offer an excellent end run around memory restricted saving. Although the game itself is flawed, Drakengard uses this very system to afford the player excellent access to different points in the game.

"I think a fair balance can be achieved in save structures by balancing user freedom with safety"

Indeed, this is the essential tension, isn't it? Between protecting the player and giving the game-literate player better control of their experience. My argument boils down to the idea that in a mass market product the former role outstrips the latter, I suppose.

Regarding RPGs and LARPs, I can see your point of view here, but I feel there could be an a priori factor here - that it's not how much role-playing you've done, but whether or not you are predisposed to a deeper connection with the role-play. It might be hard to test. :)

[Dan Boutros]

How about a Saves List...

1. Checkpoint Save, automatically done by the developer. It's a level start or Mission Start

2. Chosen Save, you chose a point in the game where you'd like to leave it.

Interface: A list. The last two chosen saves are always available. The last two Autosaves are always available. That way all designers have a fallback plan. Let's just hope the guys behind EA's film licenses don't get lazy and start relying on such a system.

It's not perfect, but it'll work!

[psu]

I have never read a convincing argument in favor of not implementing save-anywhere. And this discussion is no different. The truth of this is so clear that I can't fathom why, as a game *player* anyone would want anything different. I don't really care about the game *design* aspects of this. Let me play how I want to play. Period.

If you want to put autosave on top of manual saves, fine, go ahead. But life is too short to deal with checkpoint-only games (even though I play them, I hate them).

Just the other day I got screwed again by the "No, you can't try out that other level again because I will delete the single checkpoint that I saved for you on this HUGE hard disk" brain damage in FAR CRY.

Developers who do this should be forced to use a development system for their games that only allows a single save slot for all their source code. That way the games would never make it to release.

[Chris]

Physical media restrictions are the chief reason you don't always see it. The more dynamic the game world, the more data must be saved in each save file. For some games, the cost (in time and money) of implementing such a save system can actually be prohibitive!

For instance, we never got the save anytime I wanted in Ghost Master because of the problems involved in saving and reinitiating the large number of dynamic particle systems in the game (or at least, so the programmers claimed!) :)

On PC, where save space is effectively infinite, this is less of a problem, but on consoles it can be a significant issue.

There are also logical game reasons i.e. the risk of the player saving in an unrecoverable situation. This isn't an argument against saving anywhere - but it dictates that a save anywhere scheme is insufficient on its own.

As an Xbox owner, you probably also suffer because a lot of games are designed for multiple platforms and only the Xbox has a hard drive by default. :)

[peterb]

To be frank, all of those reasons sound like justifications rather than explanations.

Storage is not an issue anymore. It is 2005. Storage. Is. Free. In nearly every game we are talking about, the choice is not between "save 8 bytes of data, or 3 gigabytes", but "save the player's position as a single 16 bit int, or write out a couple of kilobytes of data."

Game designers choose the former not because of memory constraints, but because they are _bad designers_ who are making _bad design decisions_. That's all.

[Chris]

Programmer time isn't free. QA time isn't free. To be able to save at any time requres either the equivalent of a core dump, which does exceed media storage in most cases even now, or a design which allows for that eventuality *in all cases*. Do you have any idea what the QA implications for that are for a typical upper market console game?

Any game with in-game sequences of animation of sufficient complexity cannot be saved in an arbitrary situation in a couple of kilobytes. Hermetic saves are cheaper to implement and more reliable. Game designers who choose this approach will offend some players, but they are not making an objectively bad design decision.

[peterb]

If ease of implementation becomes the yardstick by which we judge games -- or, indeed, any software product -- then we should prepare ourselves to uncritically praise a unending stream of mediocre products.

Fortunately, we don't have to do that. If a developer can't budget adequate QA or implementation time to allow the user to _save their document_, that's fine -- just don't be offended when people rightly criticize the product for it.

[Chris]

You've completely lost me, I'm afraid. I don't think I did a very good job of explaining the points I was trying to make about the financial realities of console development... I'm not feeling at my best today.

Feel free to criticise products in whatever way you wish! You won't offend me. :)

Conversely, if I have said anything that caused offence to you or psu you have my unreserved apology. Such is never my intent.

[peterb]

Oh hey, I don't think you offended either of us! We're just arguing vigorously. I can't speak for psu, but I took the time to comment because I respect your opinions and the way you express them. I apologize if I came across as angry or harsh.

-peterb

[psu]

I just don't understand this statement:

>Any game with in-game sequences of animation of
>sufficient complexity cannot be saved in an
>arbitrary situation in a couple of kilobytes.

When I say "save anywhere" I don't really mean that when I bring the save file back, the complete state of the game will be instantaneously identical to when I saved. I just mean that I'll be in the same place and that the larger state of the world will be the same (i have the same stuff, the same creatures are dead, the crates are in the same place) and so on.

This amounts to checkpointing some high level abstract representation of the game world and the player state, which doesn't seem to me to be *that* onerous. For example, many games come close to this ideal, but strangely leave out some information that would appear on the face of it to be cheap to save, like where the player is in the level, and which creatures the player has killed.

I don't believe you have to dump the entire state of the machine to make this happen. If I'm about to get a bullet in the head just as I hit quicksave, it's not clear that saving the position of the bullet is the right thing right then anyway, and I'm not aware of any game that saves state at that level.

Not making your model persistent at a high level seems to me to be lazy. But maybe I don't have the right insight into what game models look like in the real world.

[Chris]

I see what you're saying here... the save in Dynasty Warriors is a good example of what you are talking about: it saves and restores the general state, but actual actions and so forth are lost so it's not a perfect save.

The point I was trying to raise is that a lot of upper market games now have integrated animation sequences into gameplay (Resident Evil 4 and God of War are good examples) - because there is no generalised model of gameplay to cover all these instances, you need a lot of bespoke save systems to cover all your unique cases - which takes time and therefore money to implement, and also time and money to test.

Although I would love a save anywhere system in Resident Evil 4 ( I want to stop when I want to stop!), I suspect it could have added up to 10-20% onto the project budget - or we'd have had to lose an equivalent amount of features to stay inside budget.

Of course, you could choose to design all games with generic gameplay models to sidestep the problem - but in the upper market, such games are becoming rarer and rarer. Since both Resident Evil 4 and God of War recieved rave reviews, it's hard to make a commercial argument for them swallowing the cost of saving anywhere, if you see what I mean.

Of course, none of this excuses Far Cry from screwing you over with bad save game management. :)

[Dan Boutros]

I was playing Mercenaries just yesterday and it dawned on me how 'saving achievments' would make me not hate the save system in that game.

It *really* pisses me off how playing through a n assassination mission, I have to wait till the very end to save it, and my achievements on route to said assassination are not saved or counted for.

Achievement Saves! Now that's the way forward.

[Chris]

Dan - this is what we call 'ratcheted progress'. I hope and pray it will become standard in many games. But there's a cost: for players for whom fiero (triumph over adversity) is a key payoff, being frustrated by failure can enhance the fiero payoff. I personally don't consider this to be sufficient to screw over the larger number of players for whom fail-repeat gameplay is exceptionally tedious, but I guess it depends upon the audience a game is courting.

[psu]

It's not clear to me why God of War or RE4 can't save anywhere *except* in a cut scene.

I guess the tricky issue in RE4 is that you can't save during a user triggered "cut scene". But there weren't too many of those. I'd still take a manual checkpoint save over what they gave me.

[James O]

"It's not clear to me why God of War or RE4 can't save anywhere *except* in a cut scene."

One reason being that is would be easy to get into an unworkable situation when saving in a dangerous area - for example, saving when Ashley has been grabbed, and then finding you saved at the second before she went through a door in a level, ending the game. Or saving when stuck in a dead end room with no life left and an Iron Maiden/Regenerator blocking your way out.

Of course, you could counter by saying "the game wouldn't allow it in those situations," but that means you would need to come up with every possible situation that could lead to an unworkable save, and disallow that. Can you save during a boss fight? What if you're out of healing and ammo in the middle of a boss fight? etc etc. A true save-anywhere system is dangerous for the player, but having a checkpoint system in the background can add a safety net for reckless saves.

Eternal Darkness: Sanity's Requiem for the Gamecube had somewhat of a save anywhere system - you could save anywhere, and it would return you to the entrance of that particular room (and the rooms weren't all that big, so it's not like you would have to walk far.) The only time you could not save was when "The Darkness is Coming...", apparently the game's terminology for "there are zombies in your room." However, sometimes you would be in a totally cleared room, and the game would mysteriously disallow saves. A little obtuse, but an example of how to work a save-anywhere system.

[psu]

That's why god invented multiple save slots. If you screw yourself, backtrack. Checkpoint saves force you to backtrack no matter what, which is worse.

[James O]

The idea here is that you have checkpoints in conjunction with manual saves - that way, if your manual save file should fail for whaetever reason, the game provides you a plan b.

Keep in mind here that we aren't just talking about a save system that is only accessible for the core gamers. While core gamers like ourselves may metagame the save system, making multiple backup saves at different points of the game, the casual player may not. Manual-save-only is fine for toyplay style games, like The Sims and other god games/building sims, but for a progression-style game, it is easiest (and thus likely most accessible for the casual player) to have a checkpoint system handling the saves for the player automatically. Thus, removing the checkpoint saves is somewhat disenfranchising to the casual crowd. However, it is certainly possible to run checkpoint saves in conjunction with manual saves like I mentioned in an earlier comment - this ensures the maximum number of players find their needs met.

[psu]

Ok. I think we understand each other. All I'm saying is that I'd have been happier with both RE4 and God Of War if they had had manual save-anywhere in addition to whatever else, and that this seems practical to implement. Cheers.

[Dan Boutros]

Chris: Dan - this is what we call 'ratcheted progress'. I hope and pray it will become standard in many games. But there's a cost: for players for whom fiero (triumph over adversity) is a key payoff, being frustrated by failure can enhance the fiero payoff. I personally don't consider this to be sufficient to screw over the larger number of players for whom fail-repeat gameplay is exceptionally tedious, but I guess it depends upon the audience a game is courting.

Dan (me): I believe this category covers the majority of my gaming tastes (though this seems to be changing as I'm getting tired of a lot of the painfully common negative elements in this play type).

Even though I may be growing out of this play type in the grander scheme of things, my love of this game type was often strongest when a single goal, with few sub-goals inbetween (check point to check point) was not stretched out over a half hour or more time period.

Ghouls n Ghosts / Ghost n Goblins are a great examples of this (for most of the game anyway).

I.E., if I played a hard stretch of challenge which only lasted 15-20 minutes at a maximum and died in a situation that felt of my own ill-skill, I would be tense, but excited to go through that again in search of the relief buzz payoff that beating such a challenge brings.

To me (masochistic player who likes very hard games like DMC3, VF4:Evo, Ikaruga and so on), losing after a half hour or more, without any reward or saving of ANY accomplishments is not fun, nor motivating. It's deflating and frsutrating.

I have only tried the challenge I spoke of in Mercenaries three times in the last week and a half thus far. The goal is about a fifteen minute drive from my starting point and there is much inbetween that would only be unnatural to ignore.

In the more old skool examples of games conatining this play type - when this play type was at its most popular I believe - checkpoints (in games that had them) were less than 20 minutes of gameplay apart, if that. 20 minutes may even be too long an estimate. Boss fights are usually what took up most of that time and some games had checkpoints directly before and directly after. Unless it was one of those uber painful 'fight ten bosses in a row' challenges. Then it was taking the piss, but
it was oddly seen as acceptable because such boss fights were expected to be the cream of all the challenge the game contained.

Treasure games seem to homage this well. In Megaman, 70 games on, it's starting to get annoying.

Anyways, my point in short:

New skool masochist/Fiero games fail when their checkpoints are too far apart, becuase there is a lot more to worry about in a 3D game then there is in a 2D representation of this genre. Thus, the sense of achievment and desire to try again and again is lot, as it's too much to ask of a player in any category to try and climb a mountain again, if it's taken him half an hour to climb, you've poured oil all over the rocks and just as he hits the top, you crunch his fingers with your boot and kick him in the head, right after you take the gold he picked up on the way up. And all because the screen wouldn't scroll fast enough, or he was too busy juggling duties to rotate the camera with his imaginary third hand.

I do not believe that someone within the audience these games satisfy (me) would feel motivated to try again for the challenge of such a gameplay stretch, moreso the desire to see more of the game. I will play through that Mercenaries area because I'm loving the game as a whole. I'll just have to break my natural desires to take on fights along the way, which although I don't like, I'll have to do to keep enjoying the game.

P.S. It's early morning, so if I seem incoherent, I apologise.

[Chris]

There's definitely a sense in which the distance between checkpoints relates to the key balance between frustration and fiero. Intuitively, I feel five minutes is about the ceiling of what it is reasonable for a player to repeat once or twice; when it gets to repeating dozens of times, even five minutes can become excessive.

Personally, I'd prefer not to repeat at all - but then, the older I get, the more I slip from the Conqueror/Hard Fun template and slide towards a more laid back Wanderer/Easy Fun template.

If I had the resources, I'd run a study on this - but alas... :)

[Dan Boutros]

Is that a hint? ;-)

I'll give you the short version of a study we ran on checkpointing and rewards, which may shed a hint of light on this.

We tested 2D platformers (some of which are hack and slash hybrid games) alongside 3D platformers, again including hybrids.

The main thing we found is that only the most important item finds such as power ups and quota tokens - Power ups in SMB3, Jet Type in SM Sunshine, Chaos Emeralds in Sonic, Precursor Eggs in Jak - were recorded. In the latter, they'd be autosaved imediately. Extra lives, money and enemy kills were not recorded or remembered in both 2D and 3D games.

Also, distance between checkpoints and major checkpoints (end level, or save point) was very much a 2D / 3D split. This meaning that 2D games could have levels beaten from anything between seconds an a minute when going at speed. When playing casually or for the first time, 5 minutes was the average and 12 minutes was the max, time limit constraints permiting, excluding forced scroll levels.

In 3D games, it would be 3 mins average at speed and 15 minutes when going at a casual pace/first time play. When taking time, the levels would vary wildly but the most common examples (which weren't all that regular) weighed in around 18 minutes. Not that much more than hitting the game at a casual pace.

The underlying lesson I felt this showed (as we tested some other stuff alongside this too), was that gameplay and the amount of gameplay types in a small space was far more dense and rich in a 2D game, than it was (and still is it would seem) than a 3D game.

Granted, this is down to technical constraints of hardware mainly and of course the camera view (and sometimes the developer's imagination), but it is a significant find which I think separates the two game eras (2D and 3D) and also gives a hint to why checkpoints in the games we checked out were so far apart - mainly that 3D games had to spread their gameplay over a longer distance and thus their checkpoints would be stretched out too as a result.

Totally agree with you on the 5 minute checkpoint spacing by the way. And I think you may be onto something about ages and game demographics. I love my conqueror games, but the frustration side just doesn't appeal to me anymore and just seems to get worse (above research probably enlightens us to 'why'). I've been playing more obscure freeware games lately alongside Advance Wars DS.

Cave Story (get English patch), Zen Puzzle Garden and Grid Runner have been stealing my time as of late :)

[Chris]

Thanks Dan! I appreciate the info!

[Dan Boutros]

No worries.

I quite like this place. Only place on the net I've found with decent discussion unpolluted by the net's sociopathic stereotypes.