Quote:

Original post by Derakon While I appreciate anecdotes about how people approach game difficulty, I can't help but suspect that we, as game developers, approach games in different ways from the majority of the populace. Unfortunately, I have no idea how "normal people" like their game difficulties, and I suspect it varies from person to person there as well. I suspect that no matter what you end up doing, be it making the game easy so people can see all your content, or making it dynamically adjust, or providing difficulty levels, there's always going to be some set of people who will complain (and some other set who will silently just lose interest). The question becomes which gamers like which "difficulty styles", and answering that would require making a survey.

Well, I've never had any intention of making games with the largest market (casual-gamers) i heart - I've always wanted to make games that I want to play. The casuals want to play games in a market dominated by EA (who I hope will burn and die . . ., but not before releasing Spore, just immediately after), I'm looking more at an audience keen on obscure and innovative games that fly right over the casuals heads. I know that's a bad marketing plan if I ever get to the level of COMMERCIAL games, but I'll burn that bridge when I come to it. (I am so far removed from the casual gamer, like god from man-kind, that I'm completely irrelevant.)

Quote:

Original post by Alpha_ProgDes Why can't difficulty take a "learning" approach? Basically the if a user's favorite or most used move is a Shoryuken, then the AI/NPC should guard or immobilize that move so the Player will be forced to do something else. Forcing the Player to play differently is just as a challenging as making the NPC 4x stronger than its Easy Setting, if not moreso because the Player is conscious of the fact that his moves and actions actually make a difference in the game. And the Player will actually care about what he does as he progresses.

Still talking about dynamic-adaptable-AI eh? Sure, it could get tougher, but that would make the game easy on some people and hard on others. It would create uncontrollable variables and may create an unfair environment.

On the other hand, used in the right setting and with precise control on why and how it gets hard it could rock.

On the other hand, I have no real argument, for or against, this as there are just so many situations that it would be good and where it would be bad.

Quote:

Original post by sunandshadow Quote: Original post by makeshiftwings I personally despise dynamic adaptable difficulty. I hate the idea of the game dumbing itself down to make it easier on me if I'm doing poorly, and I cry inside every time I hear another designer talk about this system as if it's the holy grail of game development. I can tell when a game is "adapting" to my difficulty, and all that makes me do is want to "game" that part of the system: I quickly figure out that the fastest and easiest way to beat the game is to play really poorly on purpose until the key points where I need to overcome something. This is counterintuitive, and not fun. To me, it's like winning a race and knowing that the other person "let you win". I stopped being amused by that behavior when I was four years old and figured out that's what my parents were doing. I want to overcome challenges, not have challenges lie down and let me walk over them if they start to assume I'm not doing well. I also like to become better at games. There is no incentive, and sometimes no actual way, to become better at games with adaptable difficulty, because they constantly keep making themselves easier to the point where you assume you're good at the game, when actually you are not. Not to be instulting or anything, but this strikes me as silly. Let me explain why: a game, by definition is an artificially imposed challenge. It fails as a game if it is so hard you keep losing and thus can't keep playing. So all games are intended to be winnable. All puzzles have built in solutions, all monsters are weak and/or stupid, the game as a whole is a challenge which is meant to 'lie down and let the player walk over it'. All adaptable dificulty does is move the guesswork about the player's abilities from the design phase to the play phase where these abilities can actually be tested instead of guessed at. Also, adaptable difficulty games do not keep getting easier and easier, they generally will get adapt to the player's ability once near the beginning of the game and then stay at that level of difficulty because the player's ability won't change that much because most of a player's ability is based on instincts and subconscious strategies, which in turn are based on reflexes, IQ, and personality, which are permanent qualities of the player.

Actually, most dynamic difficulty systems continue to function throughout the game. I remember playing ratchet and clank 3 and finding the early stages beyond easy, but getting a few challanges latter on to rack up the deathcount. I didnt mind this, because it gave me a challenge, to get better so i could get past a certain point, unfortunetly, you could see that there were gradually less enemies at that point, and eventually, they became super weak too. So i passed the point easily and was disappoionted. I had accomplished nothing. I hadnt obtained that desirable blue key to open the blue door, no, the game decided "You suck, so hears the blue key!". Further, to add insult to injury, when i returned to a previous level to get some items i had left beyond i was rather displeased to find that early levels were devoid of enemies. Levels i had carved through without a sweat were now easier, because later levels made the game think i sucked.

Dynamic difficulty is a bad idea. Rather than implement dynamic difficulty, a game should instead use cumulative effort to help you past hurdles. What does this mean? Say you collect currancy as you kill enemies. if you keep dieing at a certain point, eventually you will rank up enough money that you can get past that area more easily, but you do so without feeling the challenge has died off.


Anyway, on to my personal views on difficulty.
There are 2 catagories game mechanics fall into. Driven and sandbox.

A sandbox should only be as challenging as you make it, because you can challenge yourself to, say, get a million dollar mansion in The Sims, or choose to lead a normal life with less of a challenge. Sandbox games, or games with sandbox elements, should always ensure that the player has some control over the difficulty, by challenging themselves or not. A previous example, Burnout, exemplifies my point. It had a sandbox element, in my opinion, because one of the driving characteristics is kickass crashes, and racking up crash points in certain modes. This causes the player to want to design great crashes for self satisfaction, which are rewarded by both points and visual payoff. For this reason, the player challenges themselves, and they are rewarded (as opposed to playing a game without using something and not being rewarded).

Drvien gameplay elements should always challenge you, even in story based games, because a story is degraded/cheapened if the protagonist doesnt have to work to achieve his goal. A story driven game, however, should never build up, build up, build up, then have the entirity of the payoff in the last segment of the game. This causes players who do not complete the game to feel the game is lacking resolve. Instead, plotlines should meander. You would acomplish resolution to plot elements throughout the game, but the number of open plot elements should outway the resolved ones, until the final resolution. This allows the gamer to constantly feel they have accomplished something, and feel a greater drive to accomplish more. There is no drive to accomplish anything in easy games.

Story driven easy games are like the grind in an RPG. You simply work through the game, without any challenge, to get to the resolution. This isnt fun.

I also do not like multiple difficulties, unless there is a normal one, and an unlocked harder one (commonly known as Hero/Legend Mode). This allows you to feel accomplishment in normal, and further accomplishment in Legend, without the cheaponing of either accomplishment.

One final important note on difficulty is that a challenge should never, ever, be drawn from bad controls, lack of intuitive control, or frustrating gameplay mechanics. You should only ever loose/fail a challenge, if you arnt good enough, not because of a glitch, not because of undetected input, not because of a glaring weakeness in the user interface, and especially not because of an SNK Boss (a boss that is made to cheat so that it is difficult, rather than propose a genuenly difficult challenge)

I don't see any reason to ever reduce the difficulty because the player cannot deal with it. It would be infinitely better to just start the AI out on the lowest setting and work it's way up to the player's level. And by working up, that means slow adaptability to fight against the player's style. The AI should only grow stronger where it needs to. If the player is beating the hell out of policemen with his bare hands, giving them better armor isn't going to do anything but annoy and frustrate the player when he finally does use a gun. Not because of the heavy armor, but because the armor is a direct result of him using his bare hands at the beginning, which makes no sense. A real police force would likely train their officers better in melee combat.

Quote:

Original post by Alpha_ProgDes Why can't difficulty take a "learning" approach? Basically the if a user's favorite or most used move is a Shoryuken, then the AI/NPC should guard or immobilize that move so the Player will be forced to do something else. Forcing the Player to play differently is just as a challenging as making the NPC 4x stronger than its Easy Setting, if not moreso because the Player is conscious of the fact that his moves and actions actually make a difference in the game. And the Player will actually care about what he does as he progresses.

Absolutely. AI should get harder. But never easier. Not unless the player has directly caused them to become stupid, like wearing them out, or getting them drunk. But as you mention, tuning individual tactics is a lot better than just making the AI harder. If it's possible to have the AI learn to combat the player's specific styles, then you don't need to make the AI easy or hard. Just make sure it doesn't get carried away, and stays a bit generically prepared. Don't make it go all anti on one single tactic, or the player can use that to crush them every time. Witty reverse psychology on smart AI usually never fails.

I find your example of Ninja Gaiden interesting, Splinter of Chaos -- did you ever play the re-release of that title? Ninja Gaiden Black to Westerners -- it was a more difficult varient of the original title that included the live content that was available for the original version of the game. In any case -- Ninja Gaiden was not so hard for me -- Black was significantly more difficult.

I suppose the argument should be this: difficulty level shouldn't be about not trying to insult people by either how difficult/easy a game is -- if you give people the opportunity to scale their experience (from super easy to nightmarishly hard) then some people will, naturally, play the easier version. And -- some of those people who play on easy will be quite happy doing so, just as many people have said in this thread.

Then of course, you cater for the hardcore people who like limited saves etc etc. I think it would be interesting for games like FPS -- where difficulty level is -supposed- to make a difference to allow unlimited saves on the simpler settings and limited on the more difficult ones. This of course, is just a variation on the old theme of limited Vs. unlimited -- and is of no consequence really.

Speaking personally though -- to contrast two similar games I just got done with -- Doom 3 and Quake 4. Same engine (essentially), similar idea (run along, blow $%^@ up), but key differences: Doom 3 you're all alone and stuck in Hell by yourself....Quake 4 you've got guys who can help you out; sure, they're never around when the major stuff goes down, but it's so nice to have friends to help you out.

An example of how people adapted from perceived 'unfairness' could be the flashlight in Doom3 /Quake 4. In Doom 3, -- without a user-added hack -- the player was stuck switching between an ineffectual flashlight and their weapon in order to progress. There were areas where things were pitch black. Holster your weapon -- grab the flashlight...RAWR. Monster jumps out, you dump the flashlight and whip out the weapon...-- now you can't see anything and you get the crud beat out of you. In ID's defence, there was nearly always some kind of lightsource available to help out (i.e., the crud the monsters fling at you gave light to the scene), but overall it felt...unfair. Would a seasoned space marine not duct-tape a flashlight to his standard-issue machine gun? Of course he would! But, the flashlight dilemma was used to create tension.

Contrast this with Quake 4 -- two weapons have flashlights attached -- the more powerful of the two actually has some decent stopping power in it. Hence, no more being afraid of the dark -- period.

Can you guess which game I finished more easily? If it isn't immediately obvious, Quake 4 (on the second to top difficulty ratio, say ultra hard and not nightmare) was significantly easier than Doom 3. Why? Because it was designed differently! Doom 3, you have predictable gameplay elements -- if you pick up that medpack, chances are a monster will spawn and attack you -- usually out of thin air. This is not unfair because that's part of the system -- demons and thin air go together!
Quake 4, the villains don't waste much time trying to scare you -- they just try and kill you, so less jumping out of holes in the roof -- and little or no materialising out of thin air.

To sum up, it depends on how you want people to experience your game -- build unfairness in if it serves the purpose of atmospheric immersion, story progression, whatever...or alternatively -- make things as 'fair' as possible, and build some sort of progression into the system that allows players to face more complexity/challenges the further they progress.

In the end, no matter what kind of system you decide on - some people won't favour it, and you may have to compromise on 'what works' rather than what you like best.

~Shiny.

Let me just add my 5 cents to this discussion :)
I see a game as entertainment, that is the player should be entertained while playing and the game should help this move along. If the player wants to play it though and not be challanged (s)he should be allowed to do so. If (s)he wants to _beat_ the game and face a super challange it should also be allowed.

user-challanges: great

dynamic adaptable difficulty: bad.

Quote:

Derakon said I've seen are to include your playtime and item collection rate at the end of the game, so that players have some way to compare penis lengths. :)

Show 'em at level end aswell to help the player see how well he has done. x/y monsters killed, x/y secrets found, x headshots x minerals collected etc. PErhaps displaying a graph to show that the player has improved his skills?

Quote:

makeshiftwings said I play because I legitimately want to see how good I am at racing.

You are aware that you actually need to drive a _real_ car to see how good you are :)

Quote:

Kest said Oh, and down with cheat codes as well. If you're adding cheats to your game, you might as well go into the magazine review office and put the 0 on replay-value yourself.

I actually replayed sof2 recently because I could cheat. Reaction gameplay with 4x speed, or knife throwing party at 0.25x speed.

Quote:

Original post by Shiny Would a seasoned space marine not duct-tape a flashlight to his standard-issue machine gun? Of course he would! But, the flashlight dilemma was used to create tension.

I agree that they could have gone about that in a different way. But damn, it sure was a lot of tension, wasn't it? Every new grunting sound made me spin a 180 and fire off a shot. At many points, I really didn't even want to advance. I just stood at the entrance to a hall, running scenarios to make sure I was fully prepared. In other words, I was afraid. It takes a lot of umph to create an atmosphere like that. And that games takes the cake. IMO, in terms of scariness, it even blew RE1 out of the water.

Quote:

Doom 3, you have predictable gameplay elements -- if you pick up that medpack, chances are a monster will spawn and attack you -- usually out of thin air. This is not unfair because that's part of the system -- demons and thin air go together!

I think you're mistaken. I don't remember any demons coming from thin air. There were a few lame trap doors, but there was usually somewhere for them to come from. Well, except for the flaming skulls. They cross dimensions. Heh.

Quote:

To sum up, it depends on how you want people to experience your game -- build unfairness in if it serves the purpose of atmospheric immersion, story progression, whatever...or alternatively -- make things as 'fair' as possible, and build some sort of progression into the system that allows players to face more complexity/challenges the further they progress.

I wouldn't say it's ever good to add unfair elements. Even if it does add something good. You can always find something else that does the job.

They could have easily taped that light to your gun, as the battery would still limit how often you can use it. That was just a bad design choice. As were some of the situations where zombies pop out of small closets, or the trap doors sliding open when you walk over. Those unrealistic situations actually took away from the game for me.

Quote:

In the end, no matter what kind of system you decide on - some people won't favour it, and you may have to compromise on 'what works' rather than what you like best.

How could something you like not work? I mean of course you don't put photon grenades into your 1942 world war game just because you like them. But when the decision can work either way, then go with what you think is the best. Don't rely on a group of game designers to direct you. After all, their ideas are already being put to use. As a designer, you want to design, not go with the flow.

I liked the Doom3 flashlight, and I felt that the people complaining that you should be able to tape it to your gun were in the "realism vs gameplay" camp. When they did let you tape it to your gun in Q4 and in the Doom expansion (i think?), it made it less fun for me. Not only did it remove that gameplay element of choosing between a gun or a light, but it also still wasn't realistic, as apparently our space marine was only smart enough to figure out how to tape a flashlight to a pistol, but for some reason couldn't figure out how to tape it to the guns he actually wanted to use, like the rocket launcher. From a purely realism-based viewpoint, you would think that a grizzled high-tech space marine would have night-vision goggles or ocular implants, or at the very least a light on his helmet. But then there would be no darkness in the game, which defeats the gameplay entirely.

I think the main problem with the Doom 3 flashlight is that the equipment and abilities didn't mesh with the character and scenario. They wanted to create a feeling of horror and dread, but your character was an unshakable space marine with access to tons of high tech killing machines. By contrast, the Silent Hill games go for the same feel, but put you in the shoes of an ordinary person, like a weak-looking teenage girl scared out of her mind, where it makes sense that she wouldn't think to tape the flashlight to her gun.

Quote:

Original post by Shiny I find your example of Ninja Gaiden interesting, Splinter of Chaos -- did you ever play the re-release of that title? Ninja Gaiden Black to Westerners -- it was a more difficult varient of the original title that included the live content that was available for the original version of the game. In any case -- Ninja Gaiden was not so hard for me -- Black was significantly more difficult. ~Shiny.

No, I couldn't afford the re-release . . . yet, although I do realise that a lot of people say that NG is the hardest game ever (never learned to block>roll>counter) and others thought it was over easy (either religiously used the former or ignored the timers in favor of a slower paced game), yet they did release an "expansion pack" as I guess you could call it . . .

Although, about the flashlight conversation that has come up:
I haven't played Doom personally (any of them actually), but according to the review I saw of it (XPlay) they created an environment that was too realistic and then ignored realism (like looking for ducked tape).

And in relation to dynamically adapting AI:
It struck me suddenly that due to differences in gaming philosophies, I have another reason why dynami... DA AI let's call it for long term's sake. In DA AI, the coder has to decide how hard a game should be. Should it stop you at every corner and try new challenges, or should you always, but barely, be able to pass? Different gamers would want it differently so DA AI doesn't work for everyone.

EDIT:
It suddenly struck me at some point after the original post that we hadn't actually defined what makes a game hard. Continuing with my Ninja Gaiden example, I consider any part of the game that forces my to use a health potion hard (I often max out on health potions), but others might consider it easy not because they use no health potions, but because it takes them less tries to complete a challenge. Is there really a way to universally judge how difficult a game is?

[Edited by - Splinter of Chaos on June 26, 2006 2:05:12 AM]

In my opionion, most games nowadays are waaaaay too simple; for example the new Prince of Persia installment, where all you really need to do is mash buttons on cue (screen flashes) to pull off insane, but boring, pre-scripted special moves on opponents.


Every game, no matter how complex it is, I start at 1 difficulty level above the default level unless that is the highest setting. (e.g. a game with easy-medium-hard I play on medium, a game that defaults to medium and has easy-medium-hard-expert, I play on hard; if it defaults to easy I play on medium).

It is interesting that you all seem to think of DDA as a "dumbing down" of difficulty, but does it have to be that way? What if it is the opposite, of increasing difficulty as you get better? I am thinking of a game where the AI starts out knowing nothing, and then evolves with a genetic algorithm as you play.

Here is an example, a Space Invaders game where the enemies you kill are replaced by variants of the live enemies. That way it has DDA, but instead of having an easily-exploitable system of getting easier when you do poorly, it gets smarter when you do well. If can't hit any of the enemy ships, well, they won't evolve and the game won't get any harder. But the faster you kill them, the faster they adjust to your tactics.

I think this is where DDA has potential. Another perspective is the game flOw, which I think was mentioned in the first post but was ignored in the DDA discussion. What do you think?