Obsidian Entertainment - Chapman on Difficulty
 

Obsidian's Nathaniel Chapman opens his blog account with an article on game difficulty settings:
More information.

I simply couldn't agree more.

He speaks truth, obvious as it should be.

Very well said.

There are examples of bad difficulty scaling in other genres too, such as strategy games - some strategy games increase the skill of the AI on higher difficulties. I like that a lot. Others simply give the AI the ability to cheat (i.e starting with more resources, producing units faster than what is normally possible, etc). Never been a huge fan of that.

I agree too with the article , most devs are taking the easiest way using well tested models ; personally i support Paradox's approach where the user can modify AI and self settings on bonuses , aggressiveness etc .
An INI file with easy to edit +% to HP +% to damage +/- level scaling on NPCs / loot (and other) variables is the way i see the best since it let user take control of everything

I advocate the opposite, actually.

Zero difficulty levels. Make the game you want, and let the player decide if it's something he wants to invest in, if it's a hard game. If the game is good, he'll make the effort.

Naturally, that's not going to make you rich - but from my point of view, it's the purest kind of experience and as such the one I will prefer.

Nathaniel Chapman is working as lead designer on an unannounced project George Ziets (MotB) is writing the story for.

I wonder why no-one has yet implemented different AI as a means to increase the challenging ?

I mean, AI that works more aggressive, uses more spells, enemies that "attack the wizard first" and so on.

But I fear that AI has always been and will be so for a long time, still, the far weakest point in games - at least in RPGs.

DArtagnan: I wouldn't mind if turn based games did that but with real time games I wouldn't be able to do it because of physical difficulties I have.

PS. I only play rpgs and if a game is too hard to play because of the problems with my hands that would be one less game in a genre where it is hard to find good games.

Ideally I prefer difficulty controlling how intelligent the AI is, but that is often hard to implement.

Regardless of what the difficulty setting does I prefer to have a custom difficulty level. Being able to fine tune exactly what difficulty-altering features that go into the difficulty level is nice. Take a game like Baldurs Gate where I might want to have a 100% success rate for scribing scrolls and full HP on levelup, but still want the other features that come with a higher difficulty level. This should of course be combined with pre-defined "packages" that make up a few default difficulty levels. The Imperialism strategy games did this well.

I feel for you, but unfortunately I don't think it's reasonable for an artist to take into consideration physical problems of that kind.

Unless, of course, it's part of their vision.

Personally, I think games would be "best" if they were as close to the artistic vision as humanly possible - with complete inconsideration for anything but the "art" - which will include monetary gain or mainstream appeal.

Unless, of course, the whole mainstream appeal thing IS part of the art. I actually think Blizzard has such a vision, and in that way - I can't really fault them.

I think that's because it has the risk of 'low difficulty' being confused with 'dumb ai'. Low difficulty isn't a 'bad' thing by default - dumb ai is.

I'm with DArtagnan on this one, however I'm thinking that pacing might be, up to some extend, relevant to each player's speed, in which case offering a lower difficulty setting for those who can't keep up wouldn't hurt the game.

Well, I'm not going to tell developers what they should do - but I think the whole concept of difficulty levels is a long-lasting misunderstanding. It's metagaming and it can truly hurt the vision or the experience.

Look to Demon's Souls for the perfect example of why this could be true.

If you "can't keep up" you either adapt or you stop playing. But the game remains true to its vision no matter what.

That's the kind of game I'd like to see more of.

mount&blade has "good" & "bad" AI , minimum difference (at least in "good" you don't see arrows flying over your head while you are admiring the landscape) but if an indy can do it everyone can.

The thing is, you see, that I've got this theory about difficulty.

It's not really about how difficult a game is, but about how invested as a person you are in the game.

Sure, there are games that are nearly impossible to "beat" unless you're heavily invested - and today most games are (on "medium") total pushovers for enthusiast fans.

But, in the "old days" developers WERE enthusiast gamers themselves and that's how the bar was set. It has nothing to do with being "better" as players, but simply being more invested.

That's what I believe in, in general, and that's why I want developers to set a bar that matches their own artistic vision - because then we don't have to metagame, and we don't have to mess up the experience by making it into anything but what it was intended to be.

I know that's not realistic to expect - so it's just a wish that won't come true ;)

If I understood correctly, you're saying it's best if either the game is really close to the artistic vision or if it's as far away from it as possible? Doesn't make much sense to me.
Hard difficulty should play exactly like how the developers intended their game to play ("closest as possible to their artistic vision"). So, if you want to experience it the way it was meant to be, choose hard. But, if that's too hard for someone, I don't see a problem in letting that individual to change the difficulty.
I'm no programmer, but it doesn't sound too hard. Just make the smartest AI you can, name that "hard" level of difficulty, then take away some of the its tactics and call that "normal".
I understand your point of view, but I (and I'm probably in a majority) don't really like to challenge myself. I do, however, like to be challenged, so I'm all for predefined difficulty levels.

No, that's not what I'm saying.

I'm saying I prefer games to be as close to the artistic vision of the developers as possible.

Is that hard to grasp?

No, I don't think it should.

I think there should be one difficulty level - the correct one.

Anything else will require metagaming, and it will compromise the vision (if there is one, beyond simply making money). That's why we see it fail so often.

I'm sure you don't see a problem, but I do.

It doesn't really require programming skills to understand, but probably a lot more experience with gaming than the average gamer has. That, or simply an understanding of what difficulty levels mean and how they affect development, and how they affect the player's experience - within the game and in terms of metagaming.

I really don't think you understand, and it has nothing implicitly to do with challenge.

It has to do with "purity" for lack of a better word.

You say that the majority of players do not like to set difficulty levels themselves but play with predefined ones ? i just don't think so .

I think he is saying that most people are more comfortable with a few pre-defined difficulty levels (we can call them easy, normal, and hard), rather than tweaking a few dozen parameters.

Personally I dont see why a game cant allow both, just add a "custom" difficulty level that allows you to set as many parameters as you like, while the predefined difficulty levels are packages of these parameters. Plenty of games have done this.

But isen't it possible for developers to do that even now. I mean, you could assume that Normal difficulty is what they have envisioned for their game. And if you want to play closest to the developers vision you would play on normal mode.
I don't see that the ability to increase or decrease the difficulty by individual players would have to hinder that.