Exploring Loss Conditions

Probably one of the areas in game design that's most lacking is the treatment of loss conditions or failure.

There are very few games that are built around challenge and of what few games do include real challenge, most treat loss conditions with "reload from checkpoint/save game".

One of the only games I've ever seen build out real game play around loss conditions was Mount and Blade. In this game, when you lost a fight - you were taken captive for a bit then released stripped of some of your valuables. Your companions were scattered randomly across the world, or sometimes kept hostage.

After losing a fight, you could search the world for your companions and re-find them. Sometimes they would even be imprisoned and you would have to perform a difficult rescue mission to break them free.

The interesting thing in this example is that they took "losing a fight" and built a really interesting set of mechanics out of them, these mechanics worked flawlessly in the world and narrative, AND they made for exciting interesting game stories.

I've never seen another game come close to such an elegant system. But, there are a lot of games out there... has anyone ever seen another game that handled losing elegantly / interestingly? Do you have any ideas for how losing could be handled in a way that enriches game play?

 

That's really interesting. But that's not "losing" so much as "suffering a setback" — you've lost the fight, but you haven't lost the game (or your virtual life). Is there any way to completely lose the game? Or is it, no matter how stupid you are, you just can't die?

There is Rogue Legacy (whose key feature has also been done elsewhere, I think). You lose the "game" but in another sense, the game continues, as some of your skills/attributes/equipment pass on to the next play-through. So there is sort of a larger metagame that you can't lose, though you can suffer a setback.

In some of the LEGO video games, taking too much damage causes you to drop all the collectables you have collected wherever you are, as you immediately respawn. You then have to run around trying to gather up as many of them as you can before they fade away (which is pretty quick). So again, it's enough of a setback to be avoided, but doesn't really stop your progression through the game.

We're working here on a roguelike educational game. Because you (the player, not the avatar) are gaining skills as you play, it makes sense not to start completely over, as in a standard roguelike. Perhaps I'll steal that captured-and-stripped idea from Mount & Blade. It might make a lot more sense in this case. But it also changes the whole game design from play-many-times to play-once, which is a very fundamental change with a whole lot of implications (including, for example, the need to make a lot more content).

 

Great topic actually, when you're working on a ten+ hour game. You can't have the player start from scratch, that's an easy way to put them off your game.

Then again it has to seem a little "plausable".

 

I have another question about this. When you win the fight, do you tie up your enemies and steal some of their stuff? Or do you just kill 'em?

I would imagine you just kill 'em, but then this makes a very lopsided moral position... you're running around committing mass murder, while your enemies are kind enough not to kill you even if you are soundly defeated. How do they deal with that?

It could be they simply ignore it — the standard "kill anything that moves, steal anything not nailed down" game morality has become so ingrained most people never even think about it. But since they innovated with the mechanic, I'm curious whether they also innovated on this aspect of the fiction.

 

Thats a very interesting topic in itself: How do violent games deal with violence that the player inflicts himself. A couple of games which have a very interesting twist on that topic are the Metal Gear series for example (especially MGS3) where violence negatively influences progression and they also introduce story elements where you get negative feedback for violence. Another game would be Spec Ops: The Line, where the player has the illusion of choice but he really doesn't while the inflicted violence spirals the story down tragically.

 

One of the reasons the "being taken captive" system worked so well in Mount and Blade was because ransoming prisoners was a significant mechanic. This is somewhat accurate historically as well, nobility was rarely killed on the field because they could be captured and sold for bounty (a king's bounty).

But again, the systems in this game are a level above anything I've ever seen in anything else.

 

Just look at the old sonic games. You want a lot of rings, but getting hit makes you lose them, yet as long as you have one you won't die (from a hit). It's both a punishment and a created gameplay opportunity.


The first issue is that the only punishment (well, less than desirable outcome) in most games is death. The only place you can fail is in combat, while the only outcomes are win or die. Then there is the fact that most games are built to be binary progressions, either you continue or you don't. There is no variability to be had.

Personally, I don't care that failure isn't explored more, because if you really flesh out the outcomes, failure can be an undefinable thing. Any outcome could be desirable in some light, and possibly should be to make it interesting.

 

In the game ZombiU, if you die, you start as a new randomly generated character back at a base camp. The idea is that another wanderer discovered the camp. You lose anything in the previous character's backpack and all experience, but if any of your previous characters left items in the camp's locker, you can grab those items.

If you go back to the area where your previous character died, you can find him or her as a zombie. Kill that zombie and you can grab the backpack and get the items.

It was a neat idea, but I think I would have preferred a traditional checkpoint really.

 

I remember scenarios like that happening in Black Caldron and Kings Quest. where failure meant you usually found yourself locked in a dudgeon somewhere. Oh ya, and Maniac Mansion as part of the plot you had to get locked up in the basement. and Gantlet rewards you for loosing. A lot of survival games have incorporated that you drop all your inventory when you die but you can go back and pick it up.

I'm reminded of street fighter where you can be rewarded a round where you see how much you can beet up a car. I think there was some reward a satiated with that. Maybe that would be a good implementation for when you loose some challenge that you have to play a mini game and perhaps get rewarded. and based on how well you do in the mini game determines how much of your inventory you loose when you "die".

I had a DM that if someone rolled a 1 their weapon got swallowed by the enemy then if we defeat the enemy and get the weapon back it would have some cool new property to it.

 

There is always the XCOM model where you can play one of two ways, either constantly save the game("savescumming") or you can recruit new soldiers who eventually level up. You have a decent size roster of soldiers to choose from so either is viable, although I would say that savescumming is probably the route most take(just a guess).

 

There's a game which came out quite recently, I think, which had an interesting death mechanic. I can't remember what the game was called... It was a 2D side-scrolling platformer, where you played as a knight defeating various enemies and collecting new swords. However, you could only switch weapons once you died, which made you look forward to seeing how your new weapon worked, and softened the blow of having to restart the level.

Drat, I wish I could remember it's name... :/

 

Rougelikes have already been mentioned but I was a huge NetHack fan for many years. I always got a kick out of the concept where everyone's dead characters are occasionally saved as a ghost opponents (with all gear and skills)that you may have to fight later in the lower levels. Still permadeath but it was a fun twist.

 

While it is an immensely interesting topic in its own right I feel for your game the more important question would be "What's within scope and where should I invest my time to generate the most value from it?". Ultimately fleshing out scenarios around players losing a fight might be investing time into content the players either will not or don't really want to experience. I'd focus on things first that benefit every player for as much playtime as possible. FTL has a nice way of handling this I think. The real final failure just ends the game with the death of your crew and usually explosion of your ship. But up to this point the systems in the game are so complex that there is a broad margin between "flawless victory" and "barely made it". I think the "close calls" are a lot more fun than the "failures". If you want high stakes gameplay with emergent narrative around a broad spectrum of how well things are going for the player, I'd focus on implementing mechanics that support your goal in a natural way, instead of creating elaborate (pseudo)failure-states for the game. I'd handle the loss condition where there is no more way for the player to win the game with a game-over screen that presents a bit of written narrative for closure. That could be modular and either randomized or created with influence from the point in the game where the player failed and other data. I think in a game hard enough for many people to ever see the game over screen, it could be time well spent to make the game feel more engaging and the narrative more "believable" by not presenting an ever same cookie cutter game-over story. But I think the real time investment should be with making all the mechanics deep and engaging enough. Jagged Alliance 2 would be one of my favorites for handling that, but I've already talked at length about that in the thread I made about it. I was sure you had participated in that, but it seems I'm misremembering:

http://forum.unity3d.com/threads/stakes-and-tragedy-in-games.365294/

 

Close calls are absolutely the best game play moments, no doubt. The trick is close calls need to fail sometimes in order to be close!

Honestly, I'm not sure what's best for my game specifically and I'm definitely not committing to the mount and blade style hostage rescue. There's just no harm in talking about those ideas, and who knows, maybe this thread will turn up some gold and show me a really cool system that I missed somewhere.

City of Mordheim (my main inspiration) did something that I really liked. Your guys won't usually straight up die after being knocked out in combat, most of the time they come back, maybe half the time they gain some kind of perm injury. Now the power gamer in me would normally just fire a guy who was maimed, but they added a little flourish.

Every injury was visually presented in a kind of awesome way on the character model.

You lose a leg? Your guy gets a peg leg. Lose an eye? Bandage across the head. Etc.

It's kind of weird, but I loved those characters. The fact that Jim had a peg leg made him way more awesome than just a few extra stats would. The scars and battle wounds gave them character, and filled in their backstory.

I donno if I could even do a system like that given my art deficit, but it's definitely worth thinking about.

 

This sounds really cool, I love stuff in games that isn't just about numbers.

Necromunda (a Games Workshop boardgame, just like Mordheim originally was) had similar rules for how your gang members would suffer permanent injuries. I think there was even something like horrific scars that strike fear into your enemies and make morale tests harder for them.

Maybe if you describe what a typical progression through your game is at the moment that might spark some ideas in others reading it here?

 

You want to separate this discussion out into loss conditions and what happens when a player looses.

Some common loss conditions

• The player runs out of health
• A specific event occurs
• A timer runs out
• Another entity wins the game

Some common things that happens when a player looses

• The game must be restarted from the beginning
• The game resets to the most recent save or checkpoint
• The player gets some penalty
• The player does not get a reward
• Nothing at all

Some interesting cases I have played

• Creeper world 3: You can't explicitly loose this game. If you fail you just see the world awash with creeper. You have to actually call the loss yourself and choose to exit the level
• Lego anything: You simply loose points and respawn
• Civ: As long as your civ is still in existence you can keep playing as long as you like after someone else wins
• Skyrim: Being caught by a guard left you with several options. Fight, go to prison and wait it out, or go to prison and attempt a jail brea.
• Supreme Commander: There were separate cut scenes if you lost. They showed the faction you were meant to be protecting overrun, and included some scenes on the specific characters you interacted with in the briefings
• Shadow of Mordor: Being defeated by a orc granted that orc a promotion and made them stronger. The orcs would also remember killing you and brag about it to their friends, or taunt you next time you encountered them. The player also got known as the 'grave walker' after dying enough times, and orcs would actually ask "How often do I have to kill you?"
• Dark Souls: Dying left all of your souls at the place where you died. The player would revert to the last save point and the enemies would all be reset. However claiming a save point would also reset the enemies. This led to an interesting push your luck mechanism, how far did you dare venture between save points?

 

• An arrow flies
• his pants fall down

... At least there's a joke in this spelling mistake.

Shadow of Mordor is a pretty interesting case since dying means creating a relationship.

 

That one catches me out a lot

It goes one step further. The final boss battle involves a nemesis show down with the Orc who killed you the most.The design is slightly off, my nemesis ended up being a low level guy I ran into a lot in the early game when my character was low level and couldn't do much. But the principle is cool.

 

I really like how games workshop games handle wounds and perma-death and stuff.

I think the Games Workshop games feature these kinds of things because it gives the player a chance to mod their models, and make little custom details - like the peg leg. The same with their mutations and stuff. It was an opportunity to customize your models and make them really visually unique.

This is something that is really hard to do with video games, but I think that permanently changing the appearance of characters in response to events in game (not just buying cosmetics or something) really adds something to a game and makes you care about the characters more. It makes it feel like what happens in the game really matters.

I really regret not being able to play Shadow of Mordor. The nemesis system really sounds interesting (I've watched a ton of videos). Did the system actually work well in practice? Did it bring a lot to the game?

 

It was a coolish looking toy, rather then something that worked well. The whole game felt like it was put together with really cool pieces, but none of them quite fit together right.

The nemesis in general was one of those cool concepts that didn't quite work. It was meant to feel like you were working within the orc society, and that your actions had specific consequences on that society. Specific functions that were good:

• Any orc that killed you got promoted to captain. That allowed for some very personal vendetta's, often I would leave the main story line for a significant period of time to hunt down one specific captain that had killed me.
• Captains had special abilities. You could only find out about these abilities by interrogating other orcs or captains. Often these abilities included immunities, so they were important to discover
• Warlords (the top captains) could only be killed by hunting down their body guards first. This lead to some emergent goals, I want to kill this warlord, therefore I must identify his body guard. To do that I had to find a captain and interrogate him. Now who is the easiest captain for me to kill...
• Captains that killed you got more experience and abilities, making them harder to kill next time. This lead to knowing many of the more powerful captains by name and sight. And some cool moments running away when a particular captain showed up

Some downsides:

• It didn't feel like killing captains had any real in game effect. Sure killing a captain meant that there were less captains around, but the orcs just went about as per normal, even if you had killed all of the captains.
• The nemesis system only processed a tick when your character died. That often meant that captains trying to do a specific task would simply wait until the player appeared on the scene. It broke the whole premise of a living world around the player

Other general comments on the game

• I never could quite get over playing Assassins' Creed in the LOTR universe. Most of game play was straight out of Assassins' Creed. Right down to climbing towers to reveal the map. And the theme just didn't quite fit.
• The difficulty curve was messed up. It took me a dozen tries to kill my first captain. The game felt difficult. By the end of the game I was killing a dozen captains a minute. The final battle itself was quite anti climatic.
• QTEs. Enough said.

 

Once you have a system that can run it, or whatever is holding you back, you should give it a try. It's a cool game that I had a lot of fun with.

Interesting, I didn't notice that while playing, or I forgot about it. It's another neat idea to make it more personal.

I think considering what is a "sane scope" for a game, the system worked well enough and served its purpose of creating personal goals and personal relations in the game world. It was a mechanic that actually added something in the fun and engagement department for me, which is a thing that I can't say about many of the common AAA tropes like collecting random relics with lore pieces attached to them, picking flowers or climbing towers. The way it was implemented I think it delivers adequate value to the player in relation to the time invested in the implementation. I doubt that a ratio of equally good "value for dev time" could be reached with creating a level of world simulation that trumps GTA V and Witcher 3. It might even hurt gameplay in unforseeable ways.

I have to agree on that one.

Actually I think the QTEs were really well done for QTEs. I'm generally not a fan of those either, but I think they surve a purpose here. When you performed badly in a fight the "avoid getting executed QTE" was like a slap on your fingers telling you "that was bad, don't do that again", but the first time this happens in an engagement the QTE is easy enough to execute that you'll pretty much always win it and get a second chance of fighting. I'd argue it is excellent feedback on your performance with an instant opportunity to try again and perform better next time. It is a downbeat during combat with a few heartbeats for you to reflect on how to change your tactic and still turn this battle around. The next time the QTE would be harder to execute and I'd have some fails on the second try, so I really wanted to avoid getting that far, third try would be a fail for me most of the times if not always. I think this is a brilliant system to give feedback and second chances without making the game too easy. Why do you think it's bad game design or rather how do you think it would have been better?
I hate QTEs for things like "jump here" or "run down this straigh corridor while a scripted sequence executes", or the god-awful bossfights in Farcry 3. Generally I prefer games without them. But in this special context of near-death feedback in Shadow of Mordor I'll make an exception and say they are great.

 

I think the loss conditions need to be proportional to the gains in the situation. Like it's no good having you lose 10 hours of gameplay and all your gear if you fall off an edge. It needs to be a balanced risk reward where the result at the end is proportional to what you are giving up. This is done well in Souls/Bloodborne where the risk is the amount of souls/echoes you are carrying and the reward is the same thing, so there is a good balance.

One of the most valuable things players put up for risk is their time, which is what most games use as the condition for losing, is that you lose progress. I'd like it if more games explored other things, like some people have mentioned attachment to characters. Or a change in the world like Shadow of Mordor, where when you die you create a more powerful enemy.

 

Found it (thanks, Playstation Access []). It's Rogue Legacy. And it came out three years ago, so it's not as recent as I thought. [] It looks like a great game, I really want to play it at some point.

 

You could have also found it in the second post in this thread.

So I'm thinking very hard about these things now with regard to the educational roguelike game we're working on. To have any educational value, players need to put a lot of time into it, and they need to be able to continue doing so for, well, pretty much forever. Sure, they may put it down for a few months, but they should be able to pick it up again whenever they feel a need for a refresher.

So this has led me to believe a multi-play design is best, where games are expected to be no more than a couple hours long at most, but you play again and again trying to beat your previous score (or actually succeed at the main goal, which is purposely hard to do). And that's the standard roguelike design, so hooray for that.

But now this discussion has me thinking about the cost of permadeath. As you point out, it's a harsh penaltiy; it feels to the player like they've wasted their time. I know in my Nethack days, there were many times I would have rage-quit if the game hadn't already abruptly ended already! I always came back to try again... but I think many other players (especially these days) do/would not.

I'd love to do something more creative, like the other ideas discussed here. But it seems to me that most of them only work in play-once designs. Any penalty that doesn't actually lose you the game means you can keep going, and you will achieve that ultimate goal, which means your only reason for playing again would be to try and improve your score... not nearly as good an incentive as trying to survive to the next lowest level, and see what's at the end of the game.

Of course I could do something like Rogue Legacy where you start over, but get some benefit from your previous play. I think that softens the blow of permadeath substantially. But what do y'all think? What other loss effects make sense in a multi-play design?

 

Maybe the player could "gamble" some sort of in game currency and if they bet more they get harder educational questions or topics? Then when they lose they get some of that currency back or all of it, depending on how far they got. So if they only made it past room one they get 10% of their money back, room two 20% and so on.
That way there is a risk reward system that punishes loss but gives an incentive to keep playing. Plus with no random chance it isn't really gambling, which may be an issue with the audience of an educational game.

 

I like the idea of putting it into the game as a straight up gambling mechanic. Make a bet on how hard a test of your vocabulary you can master, and win or lose ingame currency depending on how good you perform. It could be straight up gambling at a tavern or a "rite of passage" at a local elder that grants you level-ups or access to harder areas. This kind of mechanic could be used for gating areas of higher difficulty and allow more experienced players to skip as far ahead as they need to, to keep it interesting for them.

 

Some games handle gambling in a really elegant way. Darkest Dungeon is a good example.

Every time you leave to take on a new dungeon, you buy a bunch of (fairly expensive) supplies. These supplies only exist for that single dungeon run, win or lose they're gone after.That means that the player is rewarded for spending the least on supplies possible, without bringing too few to succeed. The more supplies you bring, the better your chance of success - but the lower the net reward.

This system is super simple, easy to implement and provides for really good game play. It's a kind of player generated difficulty modulation and risk/reward calculation at the same time... and the player immediately understands how it works.

 

*Epic facepalm* []

 

You considered going full blown arcade style? Game lengths of a few minutes at most. Multiple play throughs in a single session. Perhaps even no actual win conditions?

 

I don't see that for this game, but another idea the boys and I have kicked around is a tongue-in-cheek, Munchkin-style Roguelike that would be something like that. You level up pretty much every time you beat a monster, and you win at level 10, which generally takes 30 minutes or less.

...OK, I guess it's not much like what you're suggesting after all. It's an interesting idea. Even Gauntlet isn't like that; standard play is to continue the game until you run out of quarters (or, if playing via MAME, until you run out of time or interest). Such frequent restarts would eliminate the character/gear development that is central to the genre, though.

 

Wouldn't something with longterm meta-progression fit the "learning game" purpose better?