As a player I’ll switch off if I get the sense that a game doesn’t like me. That’s not the same as saying a game can’t be hard. But if a game is hard, it has to be hard in a fun way. In a Super Meat Boy way.
As an aspiring IF author I’m very aware that what I have made so far has mostly had only one successful path, because that was all I could write in the time I had. That means my stories might have seemed hard in a very not-fun way.
The remedies are either:
To adhere to a set of conventions, so that the player already has training in your game because it’s the same kind of thing as the last game they played
To surprise everybody but understand that it will take many iterations before a random member of the public is able to engage with what you’ve made and come away from it happy.
Make those actions feel decisive and irreversible. If the powerful magic item only works once, make that clear beforehand, and then make sure the description of using it has appropriate gravitas.
Don’t let the “walking dead” state last very long. If you have a powerful magic item that can only be used once, don’t give it to the player at the 1/4 mark and expect them to save it until the 3/4 mark.
Now, this assumes that you want there to be “walking dead” states at all, while it sounds like OP wanted to avoid them entirely. And I’ve found most of the time in design it’s possible to make the wrong irrevocable decisions lead to a swift game over or equivalent. I’ve seldom actually encountered a situation where an action logically needs to be possible, and logically needs to lead to a walking dead state. (I can think of one in Scroll Thief that I worked around in a rather clumsy way: the player character pauses to consider whether they should burn up the irreplacable item before going through with it, as an unsubtle warning to the player.)
I prefer “game-over with an undo opportunity” by default, but I also think the best choice depends on the exact game involved, what the author intended and what sort of experience the player has been led to expect. It’s probably also a good idea to emphasise that “undo” is an option in the game-over if available, since a some players forget in their reaction to having hit a failure state in the first place.
There is one failure state in the Budacanta demo, which I handled with a game-over text (taking several screens due to format limitations) and a single-click undo (which goes back to the decision that caused it - there are several valid answers in addition to the game-ending one).
Instead of instant death or notifications, there could be alternate solutions.
The Bethesda RPGs for example seem decent at that. If you failed to grab the key to a locked door for example, you can try to find a way to break in. Or if you refused to help an important NPC and now you can’t get help in return and can’t progress the game (the NPC might have died because of your decision), then another NPC contacts you and a “split” in the story happens. It’s usually a fake split, but you can’t really tell unless you restore a previous save to find out what would have happened if you made a different choice.
Obviously this is way more work compared to game over screens The payoff from the player’s perspective is huge though. It feels like there’s actual decisions I can make that matter, even though in reality, and if you play through the game twice, you discover they still amount to the same result. But on the first playthrough, the illusion of free will is there.
I agree with the idea that undo changes things quite a bit. In particular, an “instant death” has effectively zero cost, since it’s easy to back up and try some other thing. This also erases the difference between “tough” and “nasty”, since when you realize you’ve done the “irrevocable” thing you do in fact have a way to revoke it.
That leaves only the other half of the scale: are there any “walking dead” states, and do you know when you’re in one? I don’t really mind running into such a state, as long as the game is very clear I can still undo out of it… a bit weird, but it lets you keep experimenting if you’re in the middle of something. (For example, if I can use my one-use scroll in the “wrong” place to get to some new area, I may still want to keep playing for a bit even if I know I eventually need to restore.) And then there’s all those games where experimenting and replaying is sort of the point, like Hadean Lands or All Things Devours.
Maybe we need to move to a formulation that’s more about “time wasted” – and fold this particular class of annoyance in alongside mazes and the other things that one hates nowadays.
Maybe you could have ‘checkpoints’ at critical moments in the game, and an undo takes you back to a checkpoint. Only having these checkpoints occasionally still means there’s some disincentive for a player to undo.
Not sure if this a reasonable option but it’s something I think I as a player would be okay with rather than not knowing I had made a mistake and wasting many hours of frustration.
This is what I’ve been thinking. In the side scroller Shovel Knight, you have checkpoints that you can return to after a failure, but you can also destroy the checkpoints for bonuses, allowing you to risk a more punishing replay for a higher score.
This makes me wonder if we could merge this system with the “missions” from today’s sandbox games. Light this lantern if you’re taking on this challenge; if you get completely flummoxed, blow out the lantern and everything goes back the way it was, so you can start over. Alternately, smash the lantern and collect the trophy inside, but you’re on your own for this challenge!
I keep returning to the phrase “game over” and wondering if it’s a relic of a bygone time. Does the game need to be over if we’re not trying to scam quarters out of the players’ pockets?
I really like this idea. Would work really well for an option dungeon in a game for instance where the reward is that you might get some very useful things at the risk of some (insert unfortunate circumstance).
While slightly off topic and very subjective, I think for the most part games should be fair. Difficulty and fairness are very different things but often get confused. If the player feels like they died/lost/etc. but they feel like it was because they weren’t good enough or they should have tried something different, this is fair, however if the player is playing by the established ‘rules’ or expectations and they still die/lose, due to what they might perceive to be a bug, I don’t think this is generally enjoyable for the player.
Of course some people love the challenge and relish in playing a game for 10 hours only to find out they did something in the first hour that breaks the game later on. I have done this several time in infocom games and have rarely thought that these were intended, just bugs or things not implemented because they were so obscure. One of the things I find most frustrating is when you have very limited inventory space so you often need to drop items, however so many of the items are obscure and not obviously needed for something and of course you end up needing one of the things you discarded earlier. I’m trying to balance this out in my project by allowing adequate space but limiting consumables to avoid hoarding. Probably my all time favourite inventory management system in a game is resident evil 4. It rewards consideration and restricts hoarding to ensure there is constant tension. This of course changes when the game diverges from survival horror to action horror, but nonetheless it rewards the player.
That does make sense. As a player, I just don’t want to feel jerked around. Like most feelings, that can vary based on a lot of factors. Of all the Infocom imps, Steve Meretzky’s games were the most likely to make me feel that way, even though I think he was the best puzzle designer out of the lot. Death is just one way to make a player feel jerked around. How about that maze in Leather Goddesses of Phobos?
I think that in terms of narrative and gameplay, the more important the plot, the less flexible an author can be about alternate paths, life after failure, etc. Cinematic action games like The Last of Us may permit some flexibility in exploration, but the PCs don’t make momentous, non-tactical choices. Even more “open” games like those from the Grand Theft Auto series, which offer seemingly limitless possibility in terms of exploration, seldom permit the player to make meaningful decisions. You can’t, for instance, decide to skip the torture mini-game (as much as I’m sure many people would like to) in GTA V.
No matter your alignment, completed subquests, character class, or party makeup, the only ways out of Baldur’s Gate 2 is the death of either your character or Jon Irenicus.
In all of these games and countless others, death obliges the player to replay a portion of the mission and/or game.
I say this all to say that perhaps it is less important to allow players to push through failstates than it is to a) offer an interesting gameplay loop and b) reward the player for thoroughly exploring the world.
The problem there is that even IF games with interesting mechanics rarely reward repeated playthroughs. Unless there is some sort of procedural generation, the player is just re-entering the same commands. The bright spot is that replaying portions of an IF game go quickly, unless some sort of set piece boggler like the cubes in Spellbreaker is there to slow you down. But if you can get through it anyway, why not just undo?
That’s the situation you definitely don’t want: my “ballpoint pen” situation where there is no indication the game is unwinnable and the player can flounder wasting time for hours without realizing. And UNDO is only usually good for a limited number of turns I recall?
Hitchhikers Guide to the Galaxy I think is an example of a game that does this and people excused it back in the day and still do cause it’s good.
I think Infocom’s Bureaucracy is a game pointedly built around what came to be known as Zarfian cruelty - at least in my memory. I never got all the way through - perhaps it was always winnable but that seems like the kind of game that would pull an “Oh you didn’t keep the paperclip from three hours ago…”
“With this character’s death, the thread of prophecy is severed. Restore a saved game to restore the weave of fate, or persist in the doomed world you have created.”
Disclaimer, I have never played the game and only know this from the memes.
Anyway, this was a fascinating thread! The discussion of agency (or the illusion of agency) is really interesting. In my choice-based games, I use the trick of having the protagonist not actually carry out the choice a lot, if it’s a choice I don’t feel like implementing. I feel like this makes more sense in games that have a fixed, well-defined protagonist, where you could just say that this particular character would never take such an action, even if they might contemplate it.
Yes, and it does vary by interpreter… though it’s safe to count on at least one turn these days at least.
FWIW, I mostly agree with the point you made upthread -
…with the one caveat that I feel very differently about instadeath and endings that just sort of peter out… because if I can immediately undo and continue it still feels like an unbroken playthrough, where restarting feels much more heavyweight.
Counterfeit Monkey has an interesting take on this… if you do something deadly, it writes out the ending text, and then says something along the lines of:
At least, that’s what would have happened if you did something so foolish. Should we assume you haven’t? [y/n]
Which mechanically is exactly the same as using UNDO after an instadeath, but feels more like you’re continuing things instead of rewinding them maybe? (And IMO is more fun than having the game tell you you’re not allowed to even try the thing.) On the other hand, Counterfeit Monkey does have multiple “real” endings… and while some are more positive than others, they’re all well-written and narratively satisfying.
It’s important that the player feels like choices are consequential. It’s less important that they really are consequential.
Inconsequentiality doesn’t erode trust if the player doesn’t catch you at it.
At the Bay Area IF meetup a few months ago, we played a choice-based Speed IF game. It was good, and we were satisfied – it had offered lots of juicy, consequential choices.
Except it hadn’t. The author 'fessed up afterwards that the story was single-track. The different choices led the same places. But he had pulled off writing each scene so that it felt like the result of our previous choice. It would’ve stunk for repeat play, but it didn’t matter at all for our one time through.
Yes! That’s similar to the modern Prince of Persia games that employ a rewind mechanic and upon death the player/narrator voices over something to the effect of “Right, but of course that’s not how it ended…”
The nice thing about that in Counterfeit Monkey is that it’s totally explained by the story: you are two minds (Alex & Andra) in one body. The player is Andra & the parser/narrator is Alex, that’s why Alex can say “That’s what would have happened if I didn’t stop you” or something like this.
I voted for No decisions are game-ruining rather than Game over with an opportunity to undo because the latter often seems like the former with a little pointless busywork added. If the outcome is game-ruining rather than a satisfying story ending, just automatically reset after describing the bad result.
I think responses of the type “Picturing how the Machine of One Thousand Knives would instantly mince your body into tiny cubes, you decide not to turn it on while standing inside it.” bother me less than they do some people because I just see them as the most efficient way of telling me I asked to do something game-ruining - complete with the description of why and how.
For cases where the actor couldn’t reasonably know what would happen before taking the action (or where it would be out of character for them to consider the results of their actions at all), I think the undo works well enough but an automatic “time rewinds”, “in a slightly different reality” or something similar would be even nicer.