News | Forum | People | FAQ | Links | Search | Register | Log in
Difficulty Tuning In Single-Player Games
This is sort of a big catch-all thread. Questions for discussion:

1. As a level designer specifically for Quake (or any game that allows for three or so player-selectable and mapper-tunable skill settings,) how do you go about the task of tuning the three settings? Should the settings emulate the challenge of the original game levels? Do you use your own skill level as a guide? Do you implement the "hard" skill first, then scale back? If scaling back, how do you decide what enemies/items to remove?

2. As a level designer in a game where there are NOT multiple "skill levels" to choose from, what techniques are available to the designer to make the level fun for a range of player skill levels?

3. As a game designer, how do you design game systems that enable the game to be fun for people of varying skill levels? Is it possible for a game to be fun even though it's too easy? Is it possible to for a game to be fun even though it's too hard? Are game systems that adapt to the player's current performace too deceitful to be used? Or, if the player knows about them, can it still be fun?

(my responses inside)
First | Previous | Next | Last
Duh 
think i agree with pretty much anything said so far.

The megaman idea is VERY interesting, but you'd need a whole game/mappack to try it out. oh pity :/

the 'hidden' healthpacks in hl2 are shitty for explorers. I think im quite a good fighter AND i love to explore, so im the guy who runs into every healthpack. of course with its linearity hl2 fucked with all explorers anyway, so i soon got frustrated with exploring and stopped it :/.

Adaptive enemies: are a major turnoff, i remember this being fcking annoying in one of the NFS underground games:
there was easy and hard skill settings, and easy the enemies were no problem at all, but still were 2:30min behind all the time regardless of the map/my driving/number of laps (which totally destroys the fun in dominating enemies) and on hard they cheated to be always just around me, so (in a racing game) you ended up crashing into something in the last lap and loosing 10mins of WORK (thats what it felt like then). (and of course you couldnt get yourself some nice seconds between you and your enemies in the first laps). This creates 'Instant kill on error' gameplay then. IIRC this was in both games.

Leveling:
Getting better stuff is FUN! hack and slay ala diablo is fun in diablo, but not in other games. i dont want hack and slay mmorpgs, ffs!
i always liked the 'get-better' aspect daikatana had - e.g. if you were good you putall your stats into jump higher and run faster and could get more secrets. makes for replayvalue, too. Or system shock2. Yes, rpg in fps is good. v good. 
(part1) 
let me start by saying that I don't think there's an ideal way to deal with difficulty yet. I would like to know :)

The classic fixed difficulty level system works reasonably well, and it is enjoyable because you have a decent idea of your skill level. Feeling that you mastered something is important. Its issues are of course its inflexibility, especially that if you start on normal and in level 6 it is getting to be too frustrating, having to start over in easy mode is a pain. I think any game should allow changing of difficulty between levels (warcraft 3 had this).

The difficulty curve being wrong can also happen inside a level, when you are out of health before a big fight. This is what makes the whole quicksave/load thing necessary. You could say level designers should always have some health especially before big fights, but I am not sure if this can fix things entirely.

Adaptive difficulty: I hate this with a passion, because I always notice it. It seems to get more popular nowadays, I guess with the average gamer being more casual they don't really notice, and more and more gamers are looking to be entertained rather than to be tested. They want a joyride that is only really mildy interactive, yet gives them a great sense of accomplishent and skill. I always assumed that in game design you'd want to maximize the players ability to use skill, but hey, if you can make them think they have skill even though they don't have it... why not?

I just tried Need For Speed Most Wanted, it topping the charts everywhere I had to try what the fuss was all about. And yes, it does feel pretty smooth, until you notice how blatantly the AI just hovers around your position, no matter (well, almost) what you do, how much faster your car is than theirs, or vice versa. They start braking for opposite traffic, cornering slower, or conversely, drive perfectly and go faster than their cars would allow. I deinstalled the game as soon as I saw this. At least when I play gran turismo, I can turn "catch up" off, and then I can judge my progress in learning a track beautifully from how far behind I am off the 1st car. It is only about skill then.

The only AI cheats that you could get away with, are very insignificant ones, so I'd say its not a solution at all, ever. Unless you're in the interactive movie business, like EA.

Which leaves the last solution: make resources in the game infinite, so the clumsier you are, the more time it will cost you. This is still my favourite as it is so flexible along every axis, every part of the game basically has perfect difficulty automatically.

The big problem of course is that it is the wrong way around: crappy players (often newbs or casual players) have to invest 50hrs to beat an rpg, and good players (often the fans of the genre) get only 30hrs out of it. The good players are the ones that would enjoy finding every item in every corner, but are the ones that don't need it. The newbs don't even have the genre experience to know where to look for the extras, yet they need it the most.

One minor equalizer in this is the non-quickload permanent death, in games like diablo (any other games?). Here when you die, you lose x% of your gold or whatever, and you respawn in the village. You can do this as often as you like at no further penalty. Since gold is coupled with advancing much like everything else in an rpg, it holds you back. So in this case, the good players that charge ahead are much more likely to die than who spend longer building up their character, and then are forced to play longer as well to recover their lost gold.

I think a combination of all of this could work well: infinite resources + no quicksave (as diablo) + selectable characters that are fundamentally different in how fast they can acquire things (and thus represent difficulty levels!). The casual newb can choose "barbarian that learns fast and has good starting stats" and the genre fan can choose "8yr old wizard apprentice with no possesions, no stats, and no friends". The only problem here is that characters don't easily allow for skill changes half way the game, though there may be ways to make this possible too.

In the end everyone wants different things out of a game, and wants to invest a different amount of time. Yet we want to feel that we have accomplished a skill level of a certain standard. A difficulty system should allow all of this. 
(part2 ;) 
What it comes down to, is that a game can always be seen as overcoming a set of challenges. Now the game has to decide what happens if the challenge fails. It can't let the game just continue as normal, because that would make the player feel like there was no challenge, thus no game. So the gamer has to be punished. Sofar, ALL methods of punishment all amount to extra time spent (anything from repeating the level, to just doing more of the same). Does anyone know of other ways of "punishing" the player when he fails other than extra time spent?

I guess what I am saying here, that if the player needs to spend extra time, it might as well happen in the smoothest, least annoying way.

The problem is also in the formula:

game skill level / player skill level = time

That means that I, as a seasoned FPS player, if I want a game to last long, I have to play it on ultrasuperhard mode, which just makes it last long thru being very annoying. Or conversely, if I am playing a new FPS whose gameplay I don't particularly care for (has happened to me many times), I may play on easy mode just to shorten the game, but it also makes the game no challenge, so worse than it already is.

So ideally, time and skill should be decoupled somehow. The rpgish idea above does this to some extend, in the sense that an easy character doesn't necessarily make combat easier, you just need to kill less monsters to progress. But this can't work as easily for an FPS. Anyone any ideas? 
You're Overthinking This, Aard 
"Time and skill should be decoupled"? Bollocks. It's natural that beginners take more time to do everything. Twisting a game's design to try and align both noobs' experimentation and veterans' rapid confident play with some "ideal" rate of progression is silly.

Your first post makes much more sense. 
Here's A Couple Of Thoughts... 
What it comes down to, is that a game can always be seen as overcoming a set of challenges. Now the game has to decide what happens if the challenge fails. It can't let the game just continue as normal, because that would make the player feel like there was no challenge, thus no game. So the gamer has to be punished. Sofar, ALL methods of punishment all amount to extra time spent (anything from repeating the level, to just doing more of the same). Does anyone know of other ways of "punishing" the player when he fails other than extra time spent?

How about having a tricky section (like a jump, for example) where if you make the jump, then you get a reward, like a weapon that you would be getting in a couple of rooms time anyway. But if you fail to make the jump, then you fall down (and take damage from the fall) but there's a teleporter at the bottom that takes you to the other side of the jump (or the beginning, but with a bridge extended) but the weapon isn't there, and you will have to wait a couple of rooms to get it.

Also, and this is more coming from a coder's point of view than a mapper... If you have several monsters which are similar in most ways (except how they attack you), then have some way of randomy replacing some of the easier monsters with more lethal ones, and the changece to upgrade the monster increases with the difficulty level. 
Time VS. Success 
"Time and skill should be decoupled"? Bollocks. It's natural that beginners take more time to do everything.

Not always true. It would take a beginner less time to get to the end of a computer chess game than an experienced player. 
Well 
In chess the main time cost is thinking through the game's possibilities, and beginners can move quickly only by doing far less of that. Besides, good chess players will not always move slowly - they might rip through the first twenty moves of a well known opening in mere seconds.

IMO an FPS shouldn't be time regimented like chess (except for effect), but be playable at any rate the player sees fit to take. 
Eh? SiN Episodes? 
So ideally, time and skill should be decoupled somehow... ...But this can't work as easily for an FPS. Anyone any ideas?

I read a recent interview with some guy from Ritual (possibly Tom Mustaine) who said that they recently had a novice player and pro gamer play through SiN episodes, and they finished within 15 minutes of one another. To me, that suggests that they might have taken the balancing to the extreme (if some OAP finishes the game in the same time as Fatality), but I do like the ideas they have for making things tougher for veteran players.

Mind you, you could argue that increasing the difficulty automatically if the player is exceptionally good could be seen as punishment for skill, which is the opposite thinking most game design has had since the first game was ever made. Skill is usually rewarded. Then again, if a player likes to be challenged, the extra challenge could also be seen as a reward, and is similar to the standard difficulty curve that should apply to most games (but often doesn't) where the game gets harder to compensate for the player getting better.

...I wonder if Sin episodes will complete itself if the player is REALLY bad. 
Wow 
great thread. comments later 
 
You can punish lack of skill with time, but I would hope it's sometimes possible to reward skill with satisfaction, or at least a warm glow that comes from knowing you did something the "right" way.

Hints in puzzly type games are like this -- it's nice to only read them _after_ you've solved the puzzle. This is difficult to apply to FPSes in the large, I guess, though there's an element of it in quake when you take out a shambler 'properly' with the DBS rather than wanking the corner. 
Damn. This Is A Good Thread. 
Some thoughts:

When adjusting difficulty, I must be weird or something, as I tend to start in the middle (normal skill), and work to both ends once I have a sense of what I want "normal" to feel like. If I'm at all hesitant, I always, always err on the side of "easy" being too easy. There are people out there who just aren't very good at [insert game here] and I don't want those people to be hopelessly frustrated. When ramping up to "hard"...well...I don't know that I really do anything different from what's already been mentioned (e.g. new/different/more enemies, fewer weapons/ammo/health), but, like metl, I try to make health less of a variable, since it will tend to adjust itself.

Adaptive difficulty: I hate this with a passion

Amen brotha', preach it. I suppose if it were done in a way that isn't hamfisted and insulting, I might tolerate it a bit more, but generally it's done in ridiculously unsubtle ways, and you can almost hear the designer saying, "Awwww...you're pretty damn retarded aren't you? Let me turn that off for you." *pats head*

Re: decoupling time and skill...IMO the trick to this is to make the exploration FUN. So what if the skilled player doesn't have to find that extra health/ammo; if you accompany finding that with finding something mildly cool that doesn't really affect game mechanics all that much (minor eye candy or Easter egg or something) then it will be enjoyable regardless, and the skilled player won't feel like they're wasting their time or jumping through hoops when they don't need to. 
Progression 
Okay, a few people brought up the secondary issue of difficulty progression. So in addition to the problem of shifting the challenge curve up or down, there's the problem of deciding how steep it should be.

The intent/assumption behind the traditional action/platformer type game is that everyone is equally good by the end, right at the moment when they beat the last level. Going into the game, some people will struggle all the way through becuase even the first level was somewhat of a challege for their pre-existing skill level. Other people will breeze through 60% of it and then hit the first level that forces them to improve their abilities. This experience can actually be MORE off-putting than the novice's experience, simply because the first part of the game created a certain expectation.

This in part depends on the difference between being "good at game X" and being "good at genre X." The more skills you can carry over from previous games, the better you'll be when you start playing. If we make the control scheme novel, then most people will find the first level mildly challenging, and nobody will be suddenly hitting a wall halfway through.

This also ties in with my idea that a game is the most fun when it's right at that threshold where if it was any harder, the player would be stuck. Slightly easier would be slightly less fun, and slightly harder would be no fun at all. By bringing people in at the bottom of the curve, they will be at the frontiers of their own abilities throughout the game.

But there's still the problem of players having a maximum skill potential. If there are some people that can only get so good at a certain type of skill, then what if those people hit that cap after playing most of a game?

Also, a steep difficulty curve probably gives people who succeeded a greater sense of satisfaction, especially when they see how far they've come in ability (the first levels will be very easy, where before they were hard.) 
Are You Suggesting... 
..to change the controls of a game to make it more difficult for players of a genre with established controls? The only reason I can see for doing that is if the controls could use improvements to start with. With the FPS genre there have only been minor changes since Quake, and Quake itself simplified the basic controls for FPS games slightly by removing the need for a use key (carried through in all id titles up until today afaik). Other games have added extra keys for things like leaning and going prone, but aside from that, I can't think of a single FPS title that has messed with the standard controls (though the default configuration has moved from using the cursors to wasd).

Anyway, with FPS games (on PC at least) there is very little improvement that could be done to the existing standard - I can't imagine anything anyway. Developers can only develop for a mouse and a keyboard, since that's what PC gamers all have. Console FPS game controls, on the other hand, are generally shit, so could probably be improved somewhat, even using standard controllers. I'd imagine that developers should concentrate not on the mapping of buttons to the controller (this should be up to the user imho), but on the way the game code interprets the players input to make things as effective as they can be.

I am looking forward to seeing how FPS games work on the Revolution :)

With regard to your original suggestion that introducing novel controls to established genres will level the playing field somewhat, I have to disagree. I think it would frustrate fans of the genre unless the controls were a huge improvement on what has already gone before. I also believe that the way to make it more difficult for these players is to bring something original to the gameplay and introduce situations that require a bit of strategic thinking rather than focussing on the reflexes of players. This way, I think players will only blame themselves for failure rather than the weird controls.

I do agree that a game is most fun when it is just at the point before being too hard, but I don't think that it is neccessary for a game to be like this all the way through. To me it's nice to have some short sections that are easy (although not short sections that are too hard). For instance, it might be fun after having difficulty with a certain type of enemy early on in the game to later on encounter them and deal with them very easily. This is probably a staple feature of grinding RPGs, although thanks to the time investment usually required of the player, I haven't really played any (certainly recently). 
Than: 
Are You Suggesting to change the controls of a game to make it more difficult for players of a genre with established controls? The only reason I can see for doing that is if the controls could use improvements to start with. With the FPS genre...

No, I'm not really suggesting that. I was actually repeating my earlier suggestion that making original games instead of following existing genres reduces the problem of balancing for both experts and newbies.

In the case of designing games in an established genre, I think you have to stick to some version of the accepted control scheme. Ways to reduce the expert/newbie gap in that situation would revolve around (of the top of my head) creating novel gameplay dynamics, like monster attack patterns that are not rehashes of familiar monsters from past games, weapons that behave differently, and perhaps interactive world objects that are not direct clones of past objects. 
<--Penis 
a game is the most fun when it's right at that threshold where if it was any harder, the player would be stuck.

That's a gigantic blanket statement, and the blanket has a lot of holes in it. A player's enjoyment of a game can vary wildly depending on a great number of things.

Some people just plain don't like difficulty and having to work at something very hard. Many people do like it, but not consistently (e.g. it could vary from game genre to genre, or even from game to game, or even depending on the particular player's whim and how much shit they had dropped on them on a particular day). 
I Tihnk 
you have to take into account what action is involved when talking bout the 'right' level of challenge. doing a very hard jump a hundred times is no fun, whereas outsmarting a very hard chess computer a hundred times is fun (or at least has a much better chance at being fun). 
On The Theme Of Adaptive Difficulty... 
what about having just certain areas with certain monsters have the 'adaptive difficulty' setting enabled? this way, the core parts of the level's combat will remain novel, but whatever more tedious parts can have adaptive difficulty enabled, thus saving the silly newbie from getting slaughtered.

just an idea, be sounds good to me 
On The Theme Of Adaptive Difficulty... 
what about having just certain areas with certain monsters have the 'adaptive difficulty' setting enabled? this way, the core parts of the level's combat will remain novel, but whatever more tedious parts can have adaptive difficulty enabled, thus saving the silly newbie from getting slaughtered.

just an idea, but sounds good to me 
Argh 
>:| 
Hmm, If It's That Adaptive... 
it should be toggleable during gameplay. Then again, many developers can't even implement a resolution change without having to restart the fucking game these days, so maybe it's a bit of a tough one. 
 
many developers can't even implement a resolution change without having to restart the fucking game these days

lol :) 
Well 
I first would like to say this is the better thread I saw since a long time ! kthx metl ! ..err... even drunk thread is not so good ;) Anyway... hum...

IMHO, it's not so easy to find the good trade-off between game skill and several player skill... I mean that playing game in medium skill can be either very easy or very hard depending of the experience of each players... (it has been already said)

Concerning mapping, it's completly subject to mappers experience both as a player, and as a mapper... On my side, as I'm using QuArK editor, normal skill is the default one when you launch quake from the editor... So normal gameplay is the basis from which I set other skill levels... From the "normal" gameplay, I simply remove/add monster in order to set easy/hard skills... Generally I use this method: if in medium you have 2 monsters, remove one for easy skill, and add 2 for hard... it also depends of the monsters strenght... (it's for sure different between soldier and shamblers..) Or sometimes replace monsters by powerfull/powerless ones... Maybe it is a "rookie" method, but at least it is very very simple... maybe too much, and it could explain why my maps are too easy for certain players sometimes... though.. ;P 
Sin Episodes Difficulty Tuning 
Courtesy of shacknews (great site, and Chris Remo awesome)

http://www.shacknews.com/onearticle.x/40410

Haven't read it yet. 
I Dont Like The Sound Of It... 
With short development cycles, the team cannot afford to do months of playtesting and balancing, so they opted to create a more powerful dynamic system that should provide every gamer with a truly unique experience.


So far the auto-skill setting I saw in games was was utter shit rubberband 
How Can You Rubberband An FPS? 
An FPS is not a race, it's a killing competition, them vs you. It's a bit difficult to come back in the last lap when you are dead.

This adaptive stuff is probably doing exactly the same thing as the difficulty levels in Quake/Doom/whatever, except the level is derived from characteristics of play instead of selected from a menu. That's not rubberbanding. Perhaps if the game spawns health packs every single time you drop under 40 health you might get some of the same why-the-fuck-do-I-even-bother feel, but hopefully Ritual isn't that incredibly lame.

IMO adaptive difficulty could work brilliantly as long as there is the tension, caused by real risk, which makes a difficult, balanced game such a joy to play. I also hope that players will be allowed to tweak or bias the difficulty system as they see fit in order to create that tension.

Why be so jaded? Let's wait and see. 
First | Previous | Next | Last
You must be logged in to post in this thread.
Website copyright © 2002-2024 John Fitzgibbons. All posts are copyright their respective authors.