Unwinnable states

What are people’s opinions on unwinnable states? When I first started my game I decided not to have any, then I started thinking it would be a cool way to make puzzles not so obvious: in other words, just because you found a piece of fruit and gave it to a NPC who ate it, doesn’t mean you did the right thing.

Are unwinnable states expected? Do they tick off the players?

How about if they’re included but addressed in a “help” tip? (“Maybe you shouldn’t have been so quick to dispose of that fruit.”)

Well, there’s several different ways your game could go:

  • The game can be put in an unwinnable state easily, and unwittingly.
  • The game can be put in an unwinnable state, but only by going out of your way to do something obviously stupid (like flush a key down a toilet before you’ve used it, e.g.)
  • The game can be put in an unwinnable state, but immediately ends or otherwise notifies you when it happens.
  • The game cannot be put in an unwinnable state.
  • The game cannot be “lost” at all – i.e., there are no losing moves.

My impression is that the more “modern” sensibility is to either avoid unwinnable states altogether, or to have the game end as soon as you put the game into an unwinnable state. However, I don’t think a game with unwinnable states would be shunned, just regarded as more “old-school”. For some people that’s a plus.

If your game can be put into an unwinnable state, I think it’s considerate to say so in the ABOUT text, so that people don’t go in with the wrong expectations and get frustrated. I also like the idea of giving the player the option to be alerted when the game is put into an unwinnable state, but that would take some programming.

Something else that occurred to me. One of the major cornerstones of game and puzzle design is feedback – that is, does the game give you positive feedback when you’re on the right track, and/or negative feedback when you’re barking up the wrong tree. When you try a wrong solution, does the game clearly acknowledge that you tried something? Is it clear why what you tried doesn’t work? Does a more correct solution suggest itself. Is it even obvious that there is a puzzle here that is meant to be solved?

Good feedback is generally considered an essential component of good puzzle design. The problem with leaving your game open to unwinnable states is that those sorts of traps tend to provide little or no feedback. I think you have to be careful to provide clues somewhere, at some point, that the player has done something wrong, even if it’s not at the precise moment that he made the mistake.

For example, if there’s a puzzle in your game that requires fruit, it should be relatively clear, either from the puzzle’s appearance or through feedback, that the player needs fruit to solve it. If the player smacks his forehead and says “Oh, crap! I need the fruit, but I gave it to the NPC and he ate it over 50 turns ago! Now I’m going to have to start all over!” – well, I think that’s fair play.

But if the fruit puzzle has no feedback and it’s not obvious what is needed there, then you run the risk of the player never realizing that he’s put the game into an unwinnable state. He’ll never realize that there is no solution anymore; he’ll just keep wandering around, revisiting dead-ends, until he eventually decides your game is too hard and gives up. You don’t want that.

So include them if you like, but keep an eye on feedback.

IIRC, that’s the exact reason I never finished Zork Zero, substituting (minor spoilers at this point, I suppose, but better safe than sorry)

“flamingo” for “NPC”, “flamingo food” for “fruit”

and “500” or so for “50”.

Actually, that may not have been the exact reason, but I did find out about it in advance from a friend and would have gotten stuck had I made it that far. (Come to think of it, I had the exact same problem with one of the later fully graphical Zorks – Return to Zork, I think? Something you can do in the very beginning that seems innocuous actually hurts you much further down the line.)

That’s not to say that I think unwinnable states are bad. Depending on how they’re implemented, they can range from only somewhat nasty to downright cruel. Certainly as a player I prefer the former.

There are some quotes from various IF reviewers about games with unwinnable states here:

jacl.game-host.org:8080/dfisher/ … 2.html#2.9

To me it’s much more fun knowing that a game either isn’t going to allow me to get into an unwinnable state, or that there will be a very clear warning if it does … then I can relax and explore / experiment with everything without having to worry about whether I’m doing something irrevocable.

But there’s room in the IF world for all kinds of games! So long as there’s a clear warning in the “about” text.


Personally, although I’ve got nothing against old-school games, I definitely prefer the modern way about unwinnable states. For me, this kind of fairness is even one of the biggest advantages of modern IF games!

Cconroy mentioned Zork Zero: I remember that I had a similar problem with another object of that game (I discovered after many turns that the game was unwinnable and I had to start all over, or almost). I still loved Zork Zero, but despite that flaw, certainly not because of it!

Death and losing endings are OK for me, though – as long as you can UNDO. There’s also a TADS 3 extension (I don’t know if something similar exists for Inform, but I guess it’s possible to do, anyway) which was used in All Hope Abandon and implements the RETRY command, which is basically a more powerful UNDO which automatically takes the game back to the last winnable state (when you die but UNDO isn’t useful because the game has been unwinnable for more than a few turns): this is OK, too.

Note that unwinnable states can be much more bearable if the game is very short (for instance, if it’s the sort of game you have to replay several times in order to win, anyway). The longer the game is, the more painful it is to start over!

I wouldn’t like that, either: when I’m stuck, I sometimes try stupid things, and I don’t think I should be “punished” for that, at least not with anything worse than a losing ending (preferably a funny one).

This is ultimately a decision about what kind of audience you’re aiming for. Many players will also think it’s fair play to throw up their hands and stop playing at this point, afraid of what else they may have unknowingly broken.

The trouble is that the number of syntactically valid commands that an IF game can accept is simply enormous. Players can only expect a tiny fraction of them to actually produce a result - and when they get a result, they see that as a positive achievement.

Not to mention that for every otherwise well designed game that punishes you for flushing keys down the toilet, there’s another badly designed game where you’ll lose all your items and wind up trapped in a locked sewer - and you don’t really know which you’re in until you’ve finished it.

Good feedback, everyone.

The fruit question wasn’t hypothetical. I want the player to give it to someone he has reason to believe will have a special interest in it, and as soon as he does so he will get the reward. But I think it might be fair to allow him to give it to any and everyone, who will gladly take it and eat it, so that he has to do a little thinking about who it should be given to instead of just ticking off a list until the result is good. As a player, I think if I gave fruit to someone who ate it or flushed a key down a toilet (without an immediate reward), I’d consider it my own fault when I needed the fruit or key later.

But the points about length of game and informing the player in the About file are good ones. And there are other puzzles where I think an unwinnable state could be even more unfair (than one where an item dissappears without reward). In fact, I am suddenly recalling some old school game where I think you had to give beef jerkey to a clam, which I was pretty upset about – because there was another character it made more sense to give it to, and because giving beef jerkey to a clam isn’t something you’d do in the real world nor was it suggested by the game clues.

You can make it with or without unwinnable states, but either way you should indicate the politeness/cruelty of the game so you know what to expect.

I don’t think there’s a problem with unwinnable states, I just think there’s a problem with unendable states. Getting stuck in a situation where there’s no way for the game to actually end is really really horrible. It doesn’t have to end immediately when the important state change occurs, but I would try to make sure that the game ends any time that no more interesting content can be explored by the player.

Argh, I wrote a long post and then it got eaten by the ether. Here’s the short version:

About Files: Can someone point me to a good example of one? Not sure what to include and how to organize it.

Politeness/Cruelty: Is this an objective index set forth somewhere? I don’t know how I’d categorize my game in its current state.

Thanks for everyone’s comments and suggestions. My game is finished and it contains unwinnable states, but I think they’re fair, the mistake will be obvious later in play, I’ll announce them in the About and Help text, the player can accomplish quite a bit despite making a mistake, and the game isn’t terribly long so it shouldn’t be too much of a pain to start over.

It’s my first game and it isn’t perfect, but it was fun and I’ll definitely do it again, and hopefully improve.

I don’t think you could really make such an index truly ‘objective’, but Andrew Plotkin’s Cruelty Scale is widely referenced. Personally, I think it’s unfair how it lumps death in with other unwinnable states, even though it’s impossible to continue further in a game when dead (at least in the sense it’s meant here), but that seems to just be me. :stuck_out_tongue:

Thanks Pacian.

Does anyone have an example of an About file they think is a particularly good one?

You’re game dosn’t have to be so well designed that you need to have examples on good about files. Just write what you think makes sense, many pepole won’t (how did that o get in there… I’ll ask my english teacher) even read it anyway.

And now I have a question. I was thinking about unwinnable states, and I wondered if it might be a good idea to put in a hint that the player has put the game into such a state. I was thinking something like this:

The laughter, as explained in the hints probably, is the author laughing at you for being a moron, and a hint that you better undo that last action. It might also be usefull for a puzzle, where you think you put the game into an unwinnable state, but in actuallity there’s a way to escape, and you know this because you didn’t hear the laughter.

Yeah, I like that evil laughter touch.

I’m going to rework the game in question (since my testers found out that it SUCKS - no that’s a bit strong, but it needs help) so that will give me a chance to implement some kind of warning like that.

For me this is a question of player freedom.

In the example of keys being flushed down the toilet- I know plenty of people who get upset when simple stupid commands aren’t implemented. “Why can’t i burn my spellbook with the lighter? It’s paper, right?” This obvious self-destructive act shouldn’t be assumed valid because it produces a result.

That being said, some actions seem correct, but screw you over. Flushing the keys down the toilet is bad, but flushing poop down the toilet seems like the right thing to do- until you find the magic spell that turns poop into a horse which you need to ride after the bad guys. That kind of thing is much more cruel. Likewise, having inventory limits when you have areas you can’t return to can be a pain in the butt, because you may need specific items but not have enough room to bring them there, however, going out of your way to lock a key behind the door it unlocks should leave it unreachable.

It comes down to: how much free will should the player have to screw himself over, and how hard to you want the game to be? Not everyone has Zork-style patience anymore.

I think the friendly response is for the PC, a wizard with a pressing need to cast spells, to say “I can’t burn the spellbook! I need it to defeat the dark lord!” You’ve acknowledged that the action is logically possible, but disallowed it as out of character.

You have a point there- but this can become a free will issue too. “Don’t tell me what my character would do” scenario could arise. Also you might not know that you will need something- but should suspect it. Maybe you aren’t the wizard but the wizard’s assistant or something. There are plenty of scenarios where a player who chooses to be destructive can ruin it for himself.

It’s a matter of opinion at that point, however. Killing random NPCs is usually blocked, but should it be? Throwing items off cliffs.

I agree there are plenty of outright mean scenarios (Spellbreaker was horrible about your spellbook getting wet without you knowing you were about to go somewhere that would make it wet), and these should be generally avoided, though I don’t know that I have a problem punishing the player for being rediculous either.

If you don’t block the action outright, or end the game immediately, you are forced to account for its consequences on all subsequent turns. So it becomes a content generation issue, which is fine if the author cares about that content, but otherwise just adds complexity and mediocrity.

Also, unless the character starts as a blank slate, it is perfectly reasonable for the narrator to tell you what the character is willing to do. Or else how would you know?

One thing to think about is why the player is being ridiculous. Often it’s because they have no idea what to do and are trying all sorts of brute force “Try this on that” approaches. My thought is that this is usually a problem with a game rather than the player, so instead of punishing the player you should give them some help.

But it depends on the kind of game you’re making – I’ve been thinking lately about a pretty great point-and-click game called Kafkamesto, which has an NPC that will take any object that you give him, remarking that it’s not what he needs. And then the object, which you might need for some puzzle, is gone gone. Obviously this game is trying to mess with your head in a certain way, which you can probably guess from the title.