Tobold's Blog
Sunday, February 07, 2016
Losing in games

I'm playing the just released XCOM 2 this weekend, and that made me think about the game mechanics of losing. To some degree that is an underdeveloped part of game design; there are about a million different rewards games hand out for winning, but only a single consequence for losing: You are forced to play some part of the game that you already did again. Losing is an essential part of games, as it increases the interest of winning. But there are a lot of things that can go wrong with loss conditions in games, and the XCOM series of games does have some examples of them.

One rather fundamental requirement for a good loss condition is that the game accurately tells you *that* you lost, and preferably gives you some information of *why* you lost. Games with a long campaign of many battles like XCOM or Heroes of Might & Magic even manage to fail the former sometimes: Either a decision you made that leads to inevitable doom will not give you any feedback at all, or worse you win a Pyrrhic victory and the game tells you that you won, while in reality you just sealed your loss.

One eternal problem with losing is how much it depends on randomness. I do think that randomness has its place in these games, and XCOM does it well by telling you exactly what your chances are for a shot. But an 80% chance isn't the same as certainty, and losing a soldier because you missed a shot with 80% chance can feel somewhat frustrating. But I do think that is more a problem of players not being good at risk assessment, and it gets better with practice.

If losing means having to replay part of the game, the obvious question is how much of the game. Console games frequently have "save points" visible or implicit, and a loss sends you back to the start of the level. As long as the level isn't too long, and you don't need to replay it 20 times before getting some jump pixel-perfect, that works quite well as a consequence of loss. Games like XCOM have loss consequences that are somewhat less transparent: If you lose a soldier, you need to recruit a rookie and train him back to that level. As the game doesn't necessarily give you the time to do that, you might actually lose the game while replenishing your losses, and then the penalty becomes a much harsher one of having to start the game over from the very beginning.

In many situations in XCOM, if you are reasonably clever you will know why a soldier of yours just died. He moved forward to quickly or otherwise exposed himself unwisely, most of the time. But there are situations where your soldier died because you come for the first time in contact with a new element or scripted part of the game. Traps you can easily circumvent the second time you meet them, but that are quite deadly the first time you see them, are rather bad loss conditions. They make people look up solutions on the internet instead of having fun trying things out. In a tactical game like XCOM you can learn a lot of general tactics by trying out, for example how close together you should keep your soldiers, or how fast they should advance. If that trial and error results in a minor wound, you learned your lesson. If it is insta-death and the error wasn't even obvious, that only gets frustrating.

The weirdest thing about XCOM and many other PC games is that you can frequently choose the consequence of losing. You aren't forced to accept the loss consequence the game just handed to you, but can opt for a different one instead by simply reloading a saved game from minutes ago. That leads to some perfectionists saving before every action and only accepting perfect outcomes, using the load function every time something didn't work perfectly. In some game the random number generator even gives different outcomes every time you reload, so you can reload until every attack is a perfect hit. Unlimited save options thus can completely negate the designed loss consequences, and make a game rather boring. XCOM fortunately has the ironman option to disable saving and reloading. I once recommended playing XCOM games on easy plus ironman, and I still think that is a good option to learn about consequences in the game. But the option most veterans will be use is an intelligent management of save game files to suit loss consequences to your personal preferences.

There is a clear but non-frustrating loss consequence: losing a "life". The game goes on like nothing has happened (for example dead XCOM soldier or WoW raider jumps back with full health), but a "life" counter goes down.

This would allow all games to be "ironman" by default, offer the same difficulty to all players as the difference between being a pro and a total noob is not gameplay outcome, but the count of lives left at the end. All XCOM players would have the same smooth experience, yet the pro could stand out as his high life count win would get to the official toplist.

The best part is that in multiplayer it would allow players of different skill to play together without frustration. If a noob stands in the fire, he dies and now the raid must kill the boss with one less man. With "lives", he would jump back to continue helping his team. The difference would be between noob and pro is that if loot is given to someone who died on the encounter it automatically loses ilvl.


From the developer's perspective that doesn't work for raiding because without full wipes, everyone would kill all raid bosses on the first night. Developers love raiding as content for exactly the same reasons I dislike it: you spend 95% of your time on organization, wipes, and standing around waiting for other players. This is how they only have to make a handful of 10-20 minute encounters to occupy players for 20+ hours a week for months.

From the player's perspective, the main problem is there is almost no way to objectively tell if a player's death was their own fault, or the healer's fault, or maybe even the tank's fault for not picking up adds. When you add a penalty for death, player toxicity goes through the roof, because you know who they will always blame for their own deaths: someone else.
Well, there might be some issues with Gevlon's idea in contexts such as raiding where the objective is just 'kill the boss'. But I could see it working in other contexts where there are repeated raids on the same boss for loot (think dungeons). You could set performance targets that allow for some frequency of player death. Eg. say you defined gold star players (never died), silver star (died once), bronze star (died less than five times). Then a breakpoint for certain loot is that you must have five that are at least gold star, three more at least silver star, one more at least bronze star, and it doesn't matter if there is a 'noob star' tenth player.

As for the loot, that could itself be divided according to stars, or not.

It wouldn't work in all situations, and it's a bit abstract, but I could see something along these lines encouraging cooperative play for players of various skill levels. Even the noobs could aspire to helping out the team by making it as bronze!

"From the developer's perspective that doesn't work for raiding because without full wipes"

Yes it does, assuming the system is built for that approach. In Marvel Heroes (an online ARPG) the raids are for 10 people, and each encounter has 30 lives. There lives deplete as people die, but they can just rejoin as they spawn. When all 30 lives are lost, the encounter fails. In the harder raids (i.e. Genosha) it is very common that you will be wiping again and again while you learn the fights, and it can get even worse when 1 clueless person can just drain those lives to such a degree that he becomes a hindrance. However someone with lesser 'skill' can die 4-5 times and still be carried by the pros that have the encounter down.

Also, in Blade and Soul you can 'buy' up to 2 extra lives before the boss fights by spending specific in-game currency. What happens when you get defeated is that your life instantly refills and you carry on normally. What this means that new players that don't know the encouners (or people with bad reflexes) can spend some resources to ensure survival, while the pros just don't bother and get to save their currency.

Some food for thought.

I don't know if it's just me, but for a post about personal "lose" conditions, I'm seeing a distortion occurring here. Many of the comments are dealing with death(as it pertains to Multiplayer games) as it occurs through a lack of "skill" or experience, and many here are offering ways that developers could implement design elements that would help reduce the effect that "unskilled" or "inexperienced" players would have on another players personal "win" condition.

Back before I made the move to MMO's, I was very active in the FPS(first person shooter) genre. Back then, the community I was involved with was very tight knit and we relied on the developers to make fixes where overall "balance" was concerned. The community policed itself, in a sense, through the formation of teams or "clans", where tryouts were held before someone was allowed to join. Now, if someone wasn't able to join team "A", they could always try out with team "X", "Y" and so on until they found a team where their skill or experience was best suited. When WoW was first released, this is the same mechanism that raiding Guilds used to populate their rosters. Servers were able to police themselves - in that not only were skill and knowledge used to determine a players positive contribution to the guild, but negative elements such as bad behavior, ninja looting and what have you, would land a person on a server wide "black list" and that player was pretty much guaranteed to not be included in most raids or dungeon runs. Now, I realize that most people consider the LFR or LFD tools to have been a success where inclusion was concerned, but did these tools not also come with the price of splitting the community with what are known as watered down "win conditions" where lesser I-Level gear was offered as a reward?

My point with this is that each type of game offers a personal win condition and a "personal" lose condition. As you indicated on your post, simply dying might be a lose condition to an individual player, but it sounds as though this discussion is taking the position that developers should be responsible for creating "fixes" that lessen that persons negative effect on their team if they don't perform well or die. It's almost as if players are accepting that random matchmaking systems are fine, but not wanting to accept the natural outcome of such systems. In my FPS days our win/lose condition revolved around ranking and win/loss record. There was no I-Level or gear induced e-peen issues. We went into a match after a careful selection and roster population process. We had only ourselves to blame if we lost our match. We didn't have a randomization process we could shift the blame to if we performed badly on a given night, and we never, ever blamed the developer or asked them to change the dynamics of the game by adding more lives, health or what have you, to change the odds of a win or loss.

I realize that the magic circle of Gaming has been permanently broken perhaps, and that the hard-core/casual debate rears its ugly head anytime someone mentions skill or experience as qualifiers, but insisting that game developers fix all of these problems because players no longer have the desire or inclination to be involved in the process speaks volumes about the state of gaming today.
Being able to "save the game" has always been a moderating factor in offline games. When you are crushed in an ambush (And you will be, in some way or another.) You generally have a max of 15 minutes to half an hour since the last save. In addition, since you know now know where the ambush is, and conversely, where it is not... you can progress through that time much faster the second time through.

As such, the designer can't really pound you too hard with the "Surprise!" bat, as you'll just respond by saving more often.

When online games first came out, designers had no such moderating factor, and their true colors came out. They LIKE pounding you with the "Surprise!" bat in an effort to make your progress as difficult as possible.

Games initially had merciless death penalties that penalized you hours if not days of playing. What emerged was "safe grinding" to avoid that. No risk taking, just mindlessly grinding the same mobs to get the slow progress.

I hated that and spent countless hours arguing with producers that merciless penalties on death were bad in that they stifled risk taking and promoted safe but really boring play. Many were so hung up on "Actions have consequences" that they couldn't see that characters are no longer made to die in a campaign and then a new character is rolled. One guy believed that quests should be massively difficult, with no in game record so you had to take copious notes to record the critical information, because there WAS going to be a test. That SOUNDS cool? But it means the quest is going to take forever to the point you would be better off grinding boring mobs.
Yeah, that is why people called it "Evercamp". Not many people went on quests, unless they had looked up the solution on Allakhazam and the reward was good. And nobody visited level-appropriate dungeons, because the risk just wasn't worth it. So let's rather go to the West Commons and shout "camp check" to see where we can grind some orcs.
I would say your point here has one major flaw. You are comparing gaming from 10+ years ago to gaming today. And I don't think what worked back then can work now. The gaming industry and community has grown and evolved so much that the sheer amount of people playing a game would make it next to impossible for a community to "police" themselves without moderators or dev intervention.

I remember games exactly like you describe. So con on PS2 was very much like that with trials for clans and players setting up tournaments and all that. However your talked about a time where to even play online meant you had extra hardware for your console, and expensive internet connection. You were basically guarenteed to be playing with only other hard core players and there were maybe 50k people playing the game regularly.

In today's world where we have hundreds of thousands of people (or millions) regularly playing a game I don't think a system like that is feasible if you want your game to be anything more then a niche title.
Socom not so con. Stupid autocorrect.
Playing a level again isn't that bad if the fault actually is with the player and the game gives feedback to convey that.

There are games that resort to "jump scares" that are just unfair and will insta kill you if you don't know them before hand. To counter frustration there has to be either a way to anticipate the insta kill or you must be able to avoid death by being good at twitch movement.

Most real old school first person shooters were made a way that you always had enough ammo (decent hit ratio expected) to live through every jump scare if you were on your toes and could strafe out of danger fast enough. If you got the rocket launcher you knew shit's about to hit the fan.

Newer shooters resort to insta kills which is just bad game design IMO. Instead the game should make it clear in level design, show a monster falling through the trap door, highlight the dangerous area in some way, just give a rather obvious clue.

WoW raids do this by having trash that uses some of the abilities of the next boss.
"One rather fundamental requirement for a good loss condition is that the game accurately tells you *that* you lost, and preferably gives you some information of *why* you lost. Games with a long campaign of many battles like XCOM or Heroes of Might & Magic even manage to fail the former sometimes: Either a decision you made that leads to inevitable doom will not give you any feedback at all, or worse you win a Pyrrhic victory and the game tells you that you won, while in reality you just sealed your loss."

It's not a failure, it's an intended feature of the difficulty system in strategy games.

When you play at the easiest setting, you can win the war by winning every battle and excelling at everything. For example, if you play Civilization on Settler, you can be the first to get every technology, build every world wonder, create a huge empire, construct all the buildings in every city, become the global leader in culture, assemble the mightiest army and beat the computer to every win condition at the same time. If you play HoMM as a Novice, you can steamroll your way across the map, grabbing every single resource node, winning every monster fight and building up an unstoppable force that never loses. And in XCOM, you can win without losing a single soldier, mission or country - provided you play on Casual.

However, one of the best pieces of advice I've ever received about 4X strategy games was "when you go up a difficulty level, pick one of the areas where you used to excel and learn to live without it". On higher difficulty levels, tactical losses are inevitable and learning how to manage them while still achieving strategic victory is a crucial player skill. By the time you get to King/Monarch in Civilization (never mind Deity), you have to correctly determine which key technologies, resources and wonders you have to prioritize and which you'll have to leave to AI. In HoMM, playing on Hard or Impossible turns you from a steamrolling juggernaut into a guerilla force that launches lightning strikes at exposed points in your enemy's defenses while your sacrificial groups lure *their* doomstacks across the map. And in I/I XCOM, it's no longer just about preventing countries from abandoning the projects - it's about picking the right ones to let go.
After reading some player reviews of XCOM 2, I see many complaints about the turn limits in some missions, while the 'professional' reviews I've read hardly mention them. After playing a few days, what do you think about the effects of turn limits in some missions in XCOM 2?
Turn limits are basically a crutch fixing a fatal flaw in the XCOM combat system: Aliens don't do anything before you "find" them, at which point they "wake up" and get a bonus round of activity. Because of that the best possible tactic in XCOM is to creep forwards very slowly, as rushing forward can get you surrounded by a group of suddenly very active aliens. The slower you move, the safer the combat becomes.

As that is neither intended nor much fun, the developers introduced the turn limits. I don't know how severe these become at the higher difficulty levels, but for me up to now they have been generous enough to allow me to reach the objective as long as I advance at least the distance that costs 1 action point to move. So turn limits haven't bothered me yet. But of course I would rather have a better alien AI than the "waking up" aliens.
AI is a *hard* problem. At least the turn limit can - I presume, not having played - create a tactical challenge without solving it.

Turn limits are basically a crutch fixing a fatal flaw in the XCOM combat system: Aliens don't do anything before you "find" them, at which point they "wake up" and get a bonus round of activity.

Ok, thanks for killing the game for me :)
One of the strong points of the original XCOM was that aliens were active BEFORE you would see them. Do you remember grenades arriving out of nowhere?
Turn limits did not exist, but there were those "Alien Terrorizing London!" events, where the aliens would go around slaughtering civilians if you were camping, forcing you to run around fast if you wanted to save anyone.
I'm surprised that with all the additional computing power available now it's not possible to have aliens which have a predefined actions (guard patrol, attack city, etc.). I can understand X-Mercs to use this trick to save on computing resources on an iPad, but on a gaming PC?!?

The "save the civilians" missions are still there. As far as I can tell some, but not all, aliens in those missions are awake from the start of the mission.
I am just playing XCOM 2 myself and I consider the loss conditions a very good feature of the game. Rather than just saying "you made an error, therefore, you die, game over", XCOM 2 presents a complex web of decision after decision after decision and in the end, you lose because of a combination of lost battles, wrong strategic choices and bad luck.

It's not one thing that you did wrong, because there is seldom is, in a war. It is your job to figure out what you did wrong and how you can improve. I think that makes the game much more interesting than "one error means game over reaload".

On the flip side, I think the discovery mechanism of XCOM 2 is lousy. Advance too fast at the wrong time, chosing one square too far ahead, and you not only lose this one soldier, but your whole team, because the enemy group gains a whole turn. That incentizes slow advancement, which is countered by turn limits. That's bad game design.
On an only tangentially related note, pointless story time!

I finally got a chance to play Xcom 2 months back and was enjoying it. I often lean more towards the perfectionist side with lots of saves and looking things up ahead of time, so for xcom especially I was trying hard to just let my decisions lay with what happened (although I wasn't brave enough to do ironman on my first play through).

I was doing fine, I had lost I think 3 team members but each and every time it was something "fair", a situation where I immediately knew that I had fucked up, gotten too cocky, and was paying for the decision I made and so losing them felt fine. Then shit went down.

I had just done the mission where we infiltrate a battleship over china, and no one had died on my team but quite a few were badly injured. I didn't think about it too hard since I only had 3 or so more days til a few other people finally recovered (I believe I had 8 total soldiers at that point, and now 6 were injured with 2 recovering shortly). The very next day though, the first terror mission popped up. I was in goddamn trouble. I tried everything I could think of from my last save, but the problem was that everything I could think of to do still took time and money, neither of which I had.

I realized that I didn't really have a choice, I would have to ignore it and take the loss which was AWFUL for someone like me. But you know what, I guess one tanked mission isn't the end of the world...I tried just ignoring it. The very next day after THAT an alien abduction popped. SON OF A BITCH. Same exact problems, except on top of it, losing THAT one on top would be catastrophic to my terror levels, with I think two nations poised to leave and others in multiple regions that were poised to go with another ignored abduction mission.

That was just too much for me...I couldn't bring myself to lose that much. I ended up reloading before the terror missions and doing what I had been trying to avoid for so long in the game: Save scumming. I took my sniper and medic into that terror mission and juggled 2 saves back and forth every other turn. And of course I beat it, it's hard to imagine losing when you are willing to rewind every single turn when anything bad happens to make a better decision, even with the seeded RNG the game appears to use.

It felt really fucking dirty, I tell you what. On the bright side, it still wasn't quite faceroll. There were a series of turns in particular that took me FOREVER to get through where I was facing a Chryssalid that had taken some civilians in an enclosed area. Trying to get the right positioning was a nightmare because of how far those guys can move and still attack. It felt a bit like a puzzle.

Still...I don't know if it was the right thing. I was so invested in my first game and didn't want to lose my progress, but ultimately it was only 2 months in and I'm sure a lot of people wind up losing big in their first playthrough due to not knowing what to do. I wonder if I should've just given up, restarted, and learned.
A positive reinforcement is nice. If you're a programmer and you get a stack trace in your program regard it as your program telling you: "hey, you might wanna take a look at this issue here" instead of "dumbass, you programmed me wrong." Same with losing in a game. However it seems some people don't learn that way, so a repercussion may aid there.

There are also some interesting games where you always lose but its a question of how far into the game. It forces you to play harder, and face the (unfortunate) consequences such as in a game like This War Of Mine. 'Newschool' gamers don't get taught anymore its normal to lose, like in Tetris.
I think what throws a spanner in the works is people expect they will be able to get to the end of the game. Like it's an entitlement.

How can you really have losing when people basically expect they can't lose (because doing so to a sufficient degree means they will never finish the game)

Of course some games seem to sense this and have an ending, but then some additional content that is freakin' hard. Generally this is where the real game is, sadly, with the other part being a long and tricky tutorial.
Post a Comment

<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool