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
ever played UT? 
I'd Like To See 
some grunt style enemies. (humans usually)
get impatient with you hiding away and stop running for cover themselves and just blitz you sometimes. Like a full on charge at you, pistols firing, maybe ready to cock you in the jaw with the butt-end of their shotgun.

If this Sin - Episodes incoporates something like that as part of their 'system' I wouldnt be as skeptical as I am of things.

As for the Noob vs Seasoned vet. I don't think you can find a flawless method because there are new players and then there are people playing who can't even install the game themselves.
As much as I admire my father as I watch him struggle with Doom 3's inhospitable hallways and dark dark dark crevies. I can't help but watch him get extremely frustrated simply by the lack of control and unfamiliarity with things. Which is also kind of peculiar to me because he managed to make it through Quake using only the keyboard. Then I showed him how mouselook works and to my surprise I found him the next time in an effort to use it. He then managed to pass Medal of Honour and I think he quite enjoyed that one. So I gave him Doom 3 (remember that game I played as a kid dad you always yelled at my for playing late into the night...well there's a new one) His level of skill is still way below that of the easiest level setting. Nearly every combat situation he has to reload over 3 or 4 times, usually a lot more when there's a tricky set of stairs for him to navigate first. I guess my point is there are players, and casual gamers, and noobs, and then there are the extremely casual gamer that struggles not only with the game but the actual implements required to play the game. There ought to be some ways to keep these players interested enough that they can get past these hangups and grow without patronizing hints and over abundance of goodies.

On a side note: My dad in his constantly losing his direction in Doom 3's dark has found almost every secret so far simply by trying to find his way back to where he's supposed to be. Except he doesn't even realize they are secrets and instead thinks he's going a new route he hasn't explored yet.

mind boggling 
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

Nethack. Save game, game exits. Reload saved game, savefile is deleted. You die, you stay dead. Utterly awesome game. Has been keeping me occupied for years on and off. 
That sounds, ummm, yeah right.

It would keep me occupied too....right up until the first time I died, and then I would be more occupied with uninstalling it and binning it. 
Oh Come On 
Getting killed in nethack is an artform. So many ways to do it. Actually when it comes to difficulty tuning, it's a rather unique example. Incredibly hard at start, but middlegame is rather easy. Endgame very hard again.

This combined with one-off characters most likely wouldn't work well in FPS games. But RPGs with proper dynamic content creation.. that might just work.

I think my point is: Game can be silly hard if it stays fresh every time you play. 
is fantastic and weirdly addictive.
of course you CAN abuse the save games by renaming a copy of your savegame file but that takes out so much fun of the game.

It's all about taking risks, but not being stupid about them. And you will die so many ways you can't help but not laugh. 
Random Map Generator + Adaptive Difficulty = Potential Solution? 
One thing I wonder about is having the amount of gameplay correlated to how well the player is doing. More specifically, the amount of difficult gameplay. For example, if the player is not doing well, remove portions of the game; in particular some difficult portions in upcoming areas. Then let the player proceed, and if he gets better, add some difficult sections back in. On the contrary, if the player is doing above average, add new areas so the game appears to be both expanded and more difficult, thus taking the same amount of time to complete as the below average player. As an added bonus, if the player has improved their skill by the end of the game, then this will add plenty of replay value.

Mind you I'm not sure how this would apply to some traditional games such as racing games, unless the race track was just a simple point-A-to-point-B style track where the player never completed more than one lap.

This also reduces the intricate layers of transition that designers use, and can limit the presentation of the story; though I suppose you could include the essential story elements in easy sections, and add sub-plots into the more difficult sections.

I can also see a potential problem where if the player is not given a challenging situation, then they don't have the opportunity practice advanced skills, and this may result in a self-perpetuating situation unless the designers are careful. 
it would destroy what makes levels great: good composition.

But generated levels are a nice idea that just need good algorithms to work. i hope ;) 
I'm Cureently Playing Enclave 
and the single most retarded feature is check points. not only do they get more scarce as the levels get tougher (bit of a paradox if you ask me), it keeps breaking the atmosphere when i have to retraverse half the level to get back to where I was. A reload form a quicksave is different because you pretty uch start off from near where you died (unless oyu quicksave rarely).

But checkpoints are crap. 
In the game 'Sly Cooper and the Thievius Raccoonus' for ps2, the player moves forward through setpiece puzzle-ish areas, that all have lots of ways to die. But in traditional console game style, the player will respawn at the beginning of the area he or she is in. So, press button 1, 2, then you die -- you respawn to the place just before pushing button 1, but not any further back.

This was also used in Mario64, but you always started at the beginning of the level, which was harder.

Seems a sensible way to deal with player deaths, as long as the game doesn't take itself too seriously (could ruin the atmopshere of a ww2 game). 
Changing Monsters Vs Ammo And Health 
Maybe this isn't a direct response to the original questions, but it might be worth some thought.

In a game where the level designer implements the difficulty settings (e.g. Quake), is it better to simply vary the number and types of monsters, or should the level designer change the amount and availability of ammo and health?

It's pretty uncommon to find a map that changes the ammo and health instead of the monsters, but it occured to me that changing ammo instead of monsters may require just as much skill from the player. So changing monsters may not be the best approach.

Maybe instead of replacing the fiend (on hard) with a couple dogs (on easy) I should remove the SNG just before the fiend and force the player to use the shotgun. Alternatively, maybe I should remove some nailgun ammo and hand out more ammo for the grenade launcher, since that takes more skill to use effectively.

Or is this forcing the player to play in a certain way and limiting the gameplay? Is the best approach to find a balance between the two? 
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.

They're not - they set the mark at 20 instead. :)

I was down to 15 at one point, and happened upon a good spot where I could slaughter guys at a distance with the pistola as they came down a long hallway at me, and since I was at 15 the whole time every one of them dropped a medkit for me. So, it's not without it's problems - if you're low on something you can 'sploit the system by killing a ton of guys before picking up whatever you're low on and you can generate stacks of it.

The easy fix for ritual would be to consider player's stats PLUS items that have dropped for the player already, but then you get the hairy problems of how old do those dropped items have to be before they no longer count, what if they fall into unreachable places, etc. 
So, there's a game called "flOw" that's supposed to be designed based on this guy's thesis on game difficulty and fun.

But what's interesting is this page where he talks about how to implement difficulty:

Summary: he rejects both static difficulty tuning (quake's skill settings) and adaptive difficulty (sin episodes) and suggests creating systems where player choices affect the level of difficulty.

This is pretty much what I was suggesting earlier in the thread, where the player has choices that determine exactly what obstacles they have to face, and has multipule approaches available to each obstacle. 
Almost everything on that site is dead, thanks no doubt to the fact that /. just posted a direct link to download flOwoOWowowow.

It makes sense though. Tuning static levels is half guesswork, as anyone who's had to flag entities in a quake map would know. Adaptive difficulty almost seems like the designers don't trust that they've made a game that can actually be finished unless there's software in place to catch cases where the player winds up SOL. 
Me and my friend played this today - we had a lengthy discussion about it, which I will try to convey if I have time.
Short version: chen's 'subliminal dynamic skill adjustemnt' doesn't seem to solve a problem as much as create one. The game is certainly addictive, or at any rate attractive, to play. But that is not at all related to the skill adjustment. I remain unconvinced. And that graph comparing challenge to ability is a little too James Edmund Pritchard for my liking. 
If the game's too easy for you you get bored, and if it's too hard for you you throw the controller at the wall and hopefully go outside. Someone just graphed it as a function of one thing over the other, which you can theoretically do with just about anything.

Below is a graph of what has been called "fatassedness" by world renowned diet doctor and professor and doctor Mikhailnikov Cznelzignyschkzovkyia, graphed as a function of metabolic rate against how much shit you eat:

According to this graph, it can be seen that eating too much for your metabolic process to handle leads to a rise in body weight, or "fatassedness", while the opposite leads to a corresponding drop in such. Proper amount of food consumption for one's metabolism leads to a perfect wavy red line within a transparent oval, within which you are always hot and sexy. Only choices on the part of the eater can maintain position within the ovoid of sexiness.

Of Course His Most Famous Chart 
was the one in which self esteem tied to metabolism equals a deep dismal feeling of unworthiness that eventually
works its way out by either a warm 357 Magmum and thousands of skull fragments on the wall, or a lane jump in front of a freight truck. 
And One More Thing 
doesn't the name Flow on that graph look like the logo for a feminine hygiene product?

That Nails It! 
I was wondering why I was getting a creepy feeling with this game. It's the ol' fear of vagina issue. 
Oh My God 
that's what my fake graph should have been about. Aunt Flow. 
so i actually played through the game (at least I think i did ... i beat the entire blue level sequence, and then did the entire orange level sequence, and then at the end of that there was nowhere to go.)

I think it's a fun game with some interesting design touches (like seeing the next level's bad guys as a blurry, faded image in the background -- this is especially cool when you first see a really huge manta-ray type creature, and builds anticipation.)

But, I don't see much here that seems to be related to the guy's thesis. The only thing I can connect to his thesis is the red/blue gateways to the next/previous levels, which are available at any time. And since fighting guys makes you stronger, skipping guys means that you can get to the end sooner, but with a weaker creature. Just like any RPG (though RPGs are less skill-based.)

Problem: looking for the red/blue portal things is ITSELF boring, since it involves waiting for a sonar ping thing, then travelling in that direction, then waiting for another one, etc. So you can't really skip boredom entirely, even though that's what the system was designed for. 
Another Bump 
For 2019. 
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.