Gamasutra: The Art & Business of Making Gamesspacer
The Designer's Notebook: Difficulty Modes and Dynamic Difficulty Adjustment
View All     RSS
October 21, 2014
arrowPress Releases
October 21, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

The Designer's Notebook: Difficulty Modes and Dynamic Difficulty Adjustment

May 14, 2008 Article Start Page 1 of 4 Next

[In Ernest Adams' latest Gamasutra column, he digs into difficulty levels in games, interestingly suggesting that player-set difficulty can, in many cases, be preferable to dynamic difficulty settings.]

I just finished reading a book called Interactive Storytelling, by Andrew Glassner. While the first couple of hundred pages contain useful introductions to both storytelling and game design (for the novice, anyway), the book has some serious flaws and I can't really recommend it. But along the way, Glassner digresses into a variety of other subjects, and one of them is settable difficulty levels. He's against them. He thinks they ought to be banned.

This really pulled me up short when I read it. My opinion is exactly the opposite: settable difficulty levels should be mandatory in most game genres. In fact, I regard the lack of a difficulty setting as a Twinkie Denial Condition, except for a few kinds of games where they aren't suitable. Since I feel so strongly about this, I was especially interested in Glassner's objections to them. I'll take a look at, and respond to, his points before I go on to talk about his preferred alternative, dynamic difficulty adjustment (also known as adaptive difficulty).

These are Glassner's complaints about settable difficulty levels.

  • The player has to decide too early. Games usually ask the player to choose a difficulty level right at the beginning, and at that point the player doesn't actually know how hard the game is going to be because he hasn't played it yet.

    My response: This isn't really an argument against difficulty settings. A game could easily give the player an optional training level at medium difficulty, and then allow the player to decide if he wants the rest of the game to be easier, harder, or about the same. But even without that, many players can make an educated guess about how well they'll play based on their experience playing similar games (I know I'm lousy at platformers), or they may choose a difficulty setting for other reasons.

    I always start every game on easy mode, because I'm very busy and I want to see as much of the game as I can in as little time as possible. On the other hand, hardcore players carrying a heavy testosterone load routinely put every game on its hardest setting; that's how you get bragging rights.

  • The options are too coarse. What if medium mode is too easy, but hard mode is too hard? The categories are too widely spaced.

    My response: There's no reason that player-settable difficulty has to be limited to three or four options; it can be a slider. Actually, any well-designed game varies its pacing so that regardless of its difficulty setting, it has easy periods and hard periods. Even arcade games give the player a breather now and then.

    The difficulty setting isn't intended to determine the difficulty of every single challenge, only the maxima and minima at any given point in the game. I don't feel this is sufficient reason for banning them. I do have one requirement, though: easy mode is supposed to be easy. So easy you can win the game by pounding the keyboard with your forehead while rolling the mouse with your elbow. If it's not, the designer has somehow failed to understand the meaning of "easy."

Article Start Page 1 of 4 Next

Related Jobs

Nuclear Division
Nuclear Division — Sherman Oaks, California, United States

Senior Game Designer
Rumble Entertainment, Inc.
Rumble Entertainment, Inc. — San Mateo, California, United States

Lead Designer/Creative Director
Nintendo of America Inc.
Nintendo of America Inc. — Redmond, Washington, United States

Visual Development Artist
Filament Games LLC
Filament Games LLC — Madison, Wisconsin, United States

Web & Interaction Designer


profile image
Interesting article, though I'd wonder about a combination of the two. Not like what he described in lego starwars, but rather you'd set a dificulty 'block' of sorts, so easy mode would ramp itself between two values within easy but never higher than a maximum or lower than a certian minium. And to keep the pacing you could always have the minium value go a little higher after each level. Just a thought.

James Sterrett
profile image
A simpler response to the objection that, "The player has to decide too early":

Let the player change the difficulty setting at will.

(Personally, I consider the absence of this feature a Twinkie Denial Condition. Far too often I have little clue what the settings mean to the developer, and if I cannot change setting in mid-game, I can wind up stuck between an overly-difficult challenge or re-playing large chunks of the game. Much better to let me adjust the difficulty at will!)

profile image
James: There are cases where this is harder than it sounds. A good for instance would be if you were making an rpg and the way you varied the diffculty was, say the various monster populations around the world. If someone switched the settling mid game it could mean needing a lot of accets you might have put aside and not loaded for easy because they wouldn't be needed. Or in some cases some levels in games are only open to the harder difficulties, if someone played though on easy, then in the last level just as they killed the boss, switched it to medium to access one of the 'bonus' levels, it'd be a rather cheatish way to get something you're meant to earn though learning to play well.

profile image
Another good example I just thought of would be starfox. The multiple paths represent the diffculties so while between stages you can to a degree modify the diffculty you can't do it mid stage because it wouldn't make any sence.

Jonathan Hartley
profile image
'Thief' and its sequels increased difficulty by varying the goals and constraints for each mission. Notably, as well as having to steal more loot, it would also forbid you from killing anyone on the harder difficulty levels ("You're a thief, not a murderer.") This is similar to taking away a player ability (ie. the freedom to sword enemies to death) worked well. Perhaps because the ability to use your sword (etc) was still present, but players had to use self-restraint and not invoke it?

profile image
DDA need not make the player feel as though they are cheating. A well designed DDA will still make the player feel like they are barely making it out of the situation.

There lies the key to tuning the game once, it is all about the scenerios you put the player in, each subsequent scenerio should be addititive and challenging. (sans the imperative puzzles described in the article)

Leaning towards an indicative gameplay is always much more powerful than leaning towards imperative gameplay.

Without a DDA, you are left with most obvious of tuning knobs a) add more hitpoints to the enemies b) enemies dish out more damage c) add more enemies. All of these are an engineers first reaction to tuning a game. A and C are always the more damaging to a game and B is the type of solution to more simulation type games like Rainbow six.

We have to remember that when the player dies that means something bad has happened. You as a the game designer or GM should keep the player alive and challenge them at the same time.

For instance,

Player's health is low = increase % drop of health pots by enemies

Player's health is low = increase % chance of critical, with a cool down of 20 seconds.

Regen players health when not in combat after 8 seconds

Create an adrenaline system

Just a few examples...

Also the arguement that the art of Level Design is going to dissapper is ridiculous, for anyone that has had to tune an entire game from higher level would understand that CHESS PIECE design is far superior than allowing each individual Level Designer to tune hp and dmg at their own whim.

Jorge Barros Cabezas
profile image
This excellent article makes me remember when I was younger, and I used to use the games' bugs to make it easier for me to end games.

Well, I agree DDA is not the best solution for all the game genres, but I agree using it for rpg, rts, and action games.

A bad example of DDA: nfs underground series (rubber band effect)

A good example of DDA: I haven't found any of my type.

For the DDA problem of a experienced player simulating he or she is a rookie, I think the game should do real dynamic adjusting of difficulty, this means, a simple rule "if the player could do that, I (the game) can argue he can do this (do a loop here until the game ends)". This can be implemented with some action-reaction, plus goals that can make a player distinct than the common, so a player will end the entire game within a range of points that recognize him as a rookie, intermediate or expert, or something like that.

profile image
"Players want settable difficulty levels"

Sure some players like difficulty levels, but I tend to think that to a large percentage of people, they are confusing and often ignored (i usually select the easiest setting, if it's too easy, i'll restart on a harder setting)

Actually it seems like alot of recent titles haven't been giving options for difficulty at all. Wii Sports was DDA wasn't it? Mario Galaxy had no difficulty settings and I have no idea if it even did DDA (anybody?). GTA4 (supposed best game ever) doesn't have any options for this. And with each of these games I really see no harm at all.

Honestly, I'll have to go with Glassner on this one. Difficulty settings almost seem cliche and old. Sure they had them in every game before 2000, but does that really mean we have to continue using them?

To tell the truth, I always thought difficulty settings were sometimes used to pad the game. (ie play it on easy, then medium, then hard, and it triples the amount of gameplay, right?)

profile image
The examples Anon 6:10 are putting out though (Mario Galaxy, GTA) are in fact what Mr. Hartley spoke of: difficulty not through arbitrary values (more monsters, more HP, less damage, etc) but difficulty through intelligent values (smarter AI, additional tasks within the same constraints, limitations to follow if the player wants greater reward without taking away ability).

Mario games have a 'easy' difficulty - the basic critical path. The player is rewarded with more stars, more coins, more lives, etc by intentionally pursuing harder but optional challenges.

And while I agree that there are many ways to avoid strict difficult levels, there may be some genres that a DDA may not be best at. Turn-based strategy games for instance rely heavily on the ability to take past actions as well as current events and respond to those. Telling an AI to suddenly see more or less or act differently mid-game may produce undesirable results.

profile image
Good article, but I wanted to bring something else that was overlooked about game difficulties. Often times the hardest difficulty setting is not even available at the outset of many games. I hate when game developers do this because I like to play games on the hardest setting, but I rarely replay a game once I have beaten it. I don't play games on hard to be an elitist or anything I play them on the hardest setting because I enjoy games more on this setting. Take a shooter for example on easy I can just breeze through it take damage and kill things without having to use cover at all. On the hardest setting I have to think more as I play and use cover. The game gets really boring to me if I can just run through it with no resistance. I think game developers should always allow all difficulty levels from the start if they are going to put difficulty levels in the game.

leon lubking
profile image
While a very interesting read, I feel one aspect is being overlooked in this article, allthough it's not applicable on all genre's.

DDA's effect on immersion.

Oblivion, for example, lost much of its immersion to me due to the random, pointless feeling it gave me when exploring its many locations. I knew in advance what foes the locations would offer and that their rewards would nearly always be slighlty above mundane for my level. Little to feed my curiosity after having seen them all once.

Where it have been so that i could have discovered an ancient site with fierce foes and a worthy relic, i would've been fascinated. Moreso, if they'd brutally slaughtered me at the time this site would function as a goal for advancement, feeding my curiosity and immersing me.

By using DDA, the latter is impossible.

Also, as a player i want to be able to perceive my mastery and advancements; there should be weaker and stronger foes for me to test my mettle against as to judge my advancement.

By adjusting the game to always make the foe a little better geared and stronger, the player never feels theyve actually grown any stronger.

While this is might be speaking in extremes (as the above can be achieved with DDA and a bit of compromise), i hope I've brought my point across clearly.



Brett Williams
profile image
This was a good read. I have to agree that I like selectable difficulty. I don't see any reason why it can't be coupled with DDA for the sake of making things fit the right zone. However, removing the players ability to choose if they want to be challenged.

I don't believe that you can sufficiently challenge a player at the level they want without allowing the player to input that level.

The pure DDA model seems to be based around the concept that a player does not want to lose. While some players don't, some also do want to lose because that's a large motivational factor for improvement.

For many players choosing a higher difficulty and failing is a sign that they have a place to improve. If the player does not fail and is able to complete all challenges with very little failure they may not feel as fulfilled.

Players also can use this to monitor their own progression. If a player never notices themselves getting better its' difficult for them to know that in fact they are getting better.

Beat matching games such as Guitar Hero are a good example. Players start out on Easy, as they get better they see that in they are able to complete songs without mistakes, as they increase the difficulty the challenge comes back.

When I play through a game on an easier difficulty and then go for the harder difficulty it's for that challenge, it adds replayability and lets me know concretely that I have progressed as a player.

Burnout Paradise does dynamic difficulty but in an interesting way. The scale of the difficulty actually has a fixed beginning and ending point. There are several challenge types Road Rage, Marked Man, Race, Stun Run. The difficulty of each challenge is independent and based on the number of those that you've done. If you race a lot and complete a lot of races the races are harder, but the stunt runs are still as easy as they were. This allows players the challenge for the events they do often and progress in, without bypassing difficulty of challenges they haven't done.

profile image
Very well-done! A lot of great points here.


profile image
What might be interesting to consider is instead of the idea I had previously where easy would have some sort of dynamic setting between 0-x and medium would be x-y and they could scale then, maybe you utilize DDA but have the 'normal' dda setting sitting on medium diffculty. then if you selected easy, it would hover always below this setting by a little and if you chose hard always a bit above. so if you die a lot even on hard it would get a little easier on you, but even so it'd still be much harder than the easy mode would be.

Tadhg Kelly
profile image
"My response: There's no reason that player-settable difficulty has to be limited to three or four options; it can be a slider. Actually, any well-designed game varies its pacing so that regardless of its difficulty setting, it has easy periods and hard periods. Even arcade games give the player a breather now and then."

So you want sliders AND variable pacing?

And do you want fries with that too?

It's a nice notion that a game is a totally customisable user-inferred or controlled system, but in what I shall refer to as the Real World for a moment that sort of solution is usually a cumbersome beast at best that effectively throws balancing back onto the player.

In otherwords, it's like giving him a set of tools and saying "now go make your game". Fine if you're talking about a toy like Sim City. Not fine if you're talking about a pick-up-and-play game like GTA 4. Most gamers don't actually desire in their hearts to be game designers. They're hear to actually play the thing, not play with the thing that makes the thing.

In short, as with many prosaic designer articles, this is a nice intellectual "what if" sort of piece, but it has pretty much nothing to do with the world of actually making and releasing games.

profile image
Interesting article. Although I can follow your arguments against Andrew Glassner's points, I just don't agree that settable difficulty levels are a good thing.

My arguments against settable difficulty levels are these:

It forces the player to make a painful realization: When a game opens with a dialogue asking me to select a difficulty level, I usually just select "easy" because I want to get to the end of the game. I paid good money for the game and I want to be guaranteed that I get what I paid for. Regardless of whether there is a settable difficulty level or not I will feel hustled if I end up in a spot that is so hard to get by that I have to give up. Anyway when I make that selection, I must confess to myself that I'm not very good at playing games. The whole experience of the game starts out with giving me a sense of defeat.

It breaks the suspension of disbelief: When I turn around a corner to find two enemies, I expect that there are two of them because that's what exists at that spot in the world that I have entered. If I know that the number of enemies is just an expression of the difficulty level I chose then it feels fake. I'm constantly reminded that the challenge I face is designed to match my selected difficulty, rather than an inescapable condition in the world I am immersed in. In the end I will not have beaten "the game"; I will have beaten a specific difficulty level. On easy level I will prevail only because I was allowed to win, and if I select the hard level, I risk getting stuck midways. It's not unlike the "rubber band" problem you mentioned in racing games. It just doesn't feel right.

Difficulty is not important. FUN is: I don't buy a game to have my abilities tested. I just want entertainment. If I find that the game is so easy that even a monkey could play it, it doesn't matter too much if it on the other hand is fun and interesting. If it is so hard I can hardly get anywhere, it BETTER be the most fun game ever for me to keep trying to get through it! Don't get me wrong. I want my games to be challenging, but you have to realize that "difficulty" is not the same as "challenge". Presenting the player with a big "Select difficulty" menu at the start of the game is giving that particular property of the game way too much attention.

You give some "solutions" to the issues that Glassner mentions. Let's say we implement all of them. What kind of game would we get? Let's see: First off the game would start with one or more "try out" levels, where the player has to try each of the skills that will be required in the game. Then the player has to select a difficulty level for each skill. Of course it won't be as simple as selecting "easy", "medium" or "hard". You have to tweak a slider, which requires an intuitive understanding of what "75% hard" means. The player will also be required to understand that even though he has selected a medium setting for "driving", that doesn't mean that the driving will be equally hard all through the game. The difficulty will still go up as the game progresses. That means that there is really no point in letting the player try out the training levels again with the selected difficulty in order to let him tweak the difficulty further, because he can't expect the same experience in the game anyway.

So, We have succeeded at making a game where a lot of the play time will be allocated to the subject of selecting difficulty. We have spoiled part of the experience by revealing what is going to happen in the game, and we have put the player in the absurd situation where he has to tweak different difficulty parameters, so that they are just right for him: "Hmm... Today I think I want a 25% risk of getting killed for every 500 yards I progress in the game. Let's see if a setting of 40% difficult will do that for me". Ridiculous...

Making the experience of the game just right is the job of the game designer, not the user.

So what options do we have? Well, obviously there is DDA, and there are many ways of implementing that. We also have some other tricks available:

Let skill in one area be exchangeable with another personal quality: One example I like to emphasize is Diablo. Here it is possible to restart the game with the character that you have already built, if the game becomes too difficult for you. In effect you can exchange skill with patience. As far as I remember Diablo also had a selectable difficulty level, but with the other mechanism it really wasn't necessary. The player is guaranteed to be able to complete the game eventually. In my opinion every game should come with such a guarantee. The customer MUST get what he pays for.

Likewise we can apply the principle to other kinds of games. In a first person shooter you can for example let the player choose if he wants to steal a jeep and drive around a mountain, or if he wants to fight his way through a tunnel system to get to the same destination. You get to choose not only how tough a challenge you want to be faced with. You also get to choose which skill you want to employ. You get to choose on a case by case basis, and it is going to feel natural. The down side is of course that more content has to be created, but you do get replay value, and it will make the game feel less linear.

Give the player clues and other help when he seems stuck: One example is "Shadow of the Colossus". When the player has been running around for a while without achieving anything, a voice will call out giving you a hint. Similarly if the player continually gets hurt by a specific type of opponent, a game may suggest that it displays which buttons to press, whenever you encounter that kind of opponent. If the player keeps missing what he aims at, a game may suggest enabling "aiming assist", so that he only have to point the gun more or less in the general direction of the bad guys and the assist will snap to the nearest target. If that makes the game too easy, you could change the sensibility or make the assist only help with the x-axis, and let the player focus on the horizontal aiming.

To conclude I think that there are so many things a creative game designer can do to get around issues with difficulty that there really is no reason to resolve to primitive solutions such as asking the user to select difficulty level at the beginning of the game.

profile image
What if the game allowed you to select the difficulty of an action as you did that action in-game? I don't mean a big dialog box coming up every time you want to kill an enemy. I mean being able to see an enemy, realize how difficult he would be, and based on that be able to either approch him or ignore him (or maybe approch him in a different way that would be less difficult). If the player is hardcore, he will approch him and be rewarded further for that. If it's a casual gamer, he could easily slip past him, but wouldn't be rewarded as much or at all.

I guess my point is that maybe it would be better to think of difficulty in moment-to-moment gameplay rather than a big box coming up before the game asking you for a tall, venti, grande selection.

Ishaan Sahdev
profile image
Really, I think it depends on the genre and target audience of your game. Genres such as racing or fighting would probably benefit the most from dynamic difficulty adjustment, where the AI tracks your progress throughout the game and adjusts itself accordingly. The point of both those genres is to get better over time. Getting too good means the game loses its appeal. DDA adds lasting appeal.

But take a third-person game or an FPS for instance. What would the proposed "slider" do exactly? There's no point if all it does is adjust HP and damage values on the player character and enemies. But for it to do any more than that -- to actually change the way the AI behaves -- would require a considerable amount of time and resources.

Another thing we haven't covered is the immersion aspect. To find yourself constantly baffled by a particular challenge, only to jump into a menu, dumb it down and then beat it makes the player lose a feeling of accomplishment that I believe is a very important part of games.

Really, I think time a developer's time would be much better invested trying to make a well-balanced game overall or one with DDA. The concept of difficulty levels, to me, need only be used when dealing with games whose challenges are only within a certain range.

Very interesting read overall. Both points of view have their merits, and each can be applied to a certain of genres. :)

Simon Lepine
profile image
I'm pretty sure the reason GTA doesn't offer selectable difficulty levels is mostly cost-cutting and not a pure design decision. I'm sure a lot of people would play in easy mode if it were available. I know I would.

I must say I agree with Ernest here. Of course we can't make blanket statements and for some specific games DDA might be the best option.

As other have noted, nothing prevents you from offering difficulty levels as ranges and offer DDA within those boundaries. Or having the DDA affect just specific gameplay elements. For example drop more ammo and health packs with DDA, and the difficulty level changes just the AI and combat stats of the enemies. This could be done to achieve having the player on the ropes regardless of the difficulty of the AI.

For me, it all comes down to giving the player the choice of his experience. Some want to be challenged and some just want to play through the story. Not offering difficulty modes just gives 1 possibility to the player. DDA doesn't give the possibility for a good player to kick ass in easy mode if he feels like it.

Henrik Axelsson
profile image
The best difficulty settings i have ever used have been in diablo 2.

Firstly you have the three corse grained setting of normal, nightmare and hell. Each setting makes the game noticeable harder, even when taking into account the rising power of the character attempting them. Also the difficulty increase between nightmare and hell is much higher than normal and nightmare.

Then there is the pacing of the difficulty levels. Normal has very obvious pacing, it starts out extremely easy and gets noticeably harder by the end, the curve in nightmare is similar but then hell is challenging the player every step of the way.

Then for fine tuned adjustment, you can use the /players x setting. I must say this is the best invention ever. The player can fine tune each of the corse grained difficulty settings by setting this from 1-8 (the setting increases the amount of hit points monsters have). Also the player is rewarded when increasing this value by increased experience rate gain and more item drops. The real beauty of it is it allows for fine tuning of the player experience based on the challenge they want or how good their character is. There are some character build that are just better at certain points of the game than others, so when using a powerful build the ability is there to jack up the difficulty. you also might have twinked out a new character with really good items, so you set /players 8 and reap in the bonus of faster leveling.

Then as a final point, there is hardcore mode. Where if you character dies you can not play them anymore.

The combination of these factors makes for a highly customizable challenge for the player, and greatly extends the life span of the game. I have not seen any game do a better job of giving the player the ability to customize their gaming experience. (i still play it too much these days, damn hardcore characters....)

Rando Wiltschek
profile image
In my opinion it all boils down to how high the player's tolerance for failure is. I personally don't mind having to play the same "checkpoint" 20 times until I beat the challenge. So I usually select the highest difficulty mode, because of that.

Someone else might be frustrated after failing a mission once, maybe because he's only interested in the story or has limited time.

Therefore I believe the initial difficulty selection should not determine the difficulty of a challenge, but the rate at which DDA adjusts up and down.

So "easy" would guarantee, that the game becomes easier once the player is heavily challenged (tracking damage taken, getting close to dying, etc.), while "medium" would allow the player to die a few times before getting easier and "difficult" would only slowly become easier but eventually be easy enough for the player to beat the challenge without being stuck.

So if the player passes a challenge without getting hurt or threatened, etc. it obviously is too easy. Depending on his selection of difficulty it needs to become harder (on easy only slightly, on hard a lot).

The other way around, if the player fails a challenge the game needs to be easier (on easy a lot, on hard just a very little).

Although there needs to be a cap how much harder or easier it can get during a fixed interval of time.

Of course this only works for DDAable games, but in my book these are a lot.

Personally I'm against dumbing down AI on easier modes. I'm a poor strategist, but I'd still want the AI be able to surprise or outsmart me - and on "easy" rather have an edge against him / his troops in power.

profile image
The art of gaming has been lost to many developers and the almighty dollar seems to rule these days. With that said, DDA works against the spirit of gaming. If a game is meant to tell a story that the customer is entitled to see to the end, then create challenges that add to the experience and let the main goals be obtainable via lowering Difficulty Settings. Being able to change difficulty when the going gets tough is a great compromise. God of War implemented this very well. What meaning does getting to the next level have for a gamer if it is just going to adjust to his "peak"? Games are supposed to be rewarding. My favorite games are those that are extremely challenging, but have the gameplay to support the challenge. Zone of the Enders 2, Ogre Battle, Onimusha 3 and Virtual Fighter Evolution have great gameplay and offer an insane level of challenge at higher difficulties, but are playable by those with less skill (maybe not ZOE2). If they had DDA the way games are implementing it today, I don't think I would feel the same way about them.

DDA also should never be used in RPGs. A decent RPG should allow players an alternative way to beat the primary goals and offer secondary goals with incentives to beat those challenges. Final Fantasy 8 may have been the worse RPG ever due to adapting enemies.

DDA may have its place in games, but should be limited to those meant to measure ones performance, such as Big Brain Academy or twitch type games.

Growing up, there were many games I was never able to beat due to not being able to overcome the challenges and not once did it stop be from buying more games of the same genre.

Stephen Kalb
profile image
Although Devil May Cry and its ilk certainly give the player a much-needed DDA in the form of unlocking Easy Mode if you die too much, my favorite example has always been Pirates! It does a few things against the above recommendations that work well: It offers a few discrete difficulties, and offers them at discrete times: It may knock the player down if he gets captured/stranded, or the player might get the option to go up a notch when he split the gold. But the big thing, in my opinion, is that the game always tells the player when it's doing so. The feedback is really vital in helping the player progress.

profile image
I believe that the reason why there is argument and uncertainty about difficulty settings, is due to the partial incompatibility between story-based and challenge-based gameplay.

If you are playing for the story, the challenges are something to entertain you until the next part of the story. They should never keep you from getting to the end, but they should also not be boring. In these cases DDA might work well.

If you are playing for the challenge, then your goal is to see how high a difficulty level you can overcome. Therefore, choosing a difficulty level may be seen as a way to tell the computer how far you have come. Or, in the case of rogue-likes for example, the difficulty is always on maximum, but the difficulty curves in the game, and death is permanent, so your score is an indication of your skill.

In general, it seems that difficulty is easier to apply in challenge-based games. Especially when winning and losing are not the only outcomes. In many games you can end up in second or third place. Sid Meier's Civilization, for instance.

One story-based game which handled difficulty well, is "Cave Story". It is a platformer and it does not have a difficulty choice at the beginning. At a certain point in the story you have to decide whether or not you exchange your weak gun for a much more powerful one. If you do, the combat is much easier and hard-to-reach places are easier to reach (you can fly with the gun), but the ending is a bit sad, and a certain bonus stage is locked. And if you don't, the game is harder, but you also get a better ending and more challenging skills to master throughout the game. You can also escape the island and leave everyone to die, which is the worst ending, but also the easiest to reach... kind of a super-easy mode.

One has to be careful with this though, a player might in the above case think "Why would I not want a stronger gun?" and not realize that it will have story-line or difficulty-level consequences. This is of course caused by the fact that game story-lines have thus far (almost) always been (almost) entirely linear. (Also, they manage to have different story-lines playing through the same maps, so that very little additional resources are needed.)

It seems to me that handling difficulty in this way would cause a stir, but would also eventually become an accepted and more immersive way to handle difficulty level.

Whichever difficulty model is used however, it has to support the game 'feel'.

A true story-based game, like an RPG, would likely work best with the story-choices method, while a racing-career-game will benefit from DDA, a strategy game could use a directly chosen difficulty level and a survival game might be better off without multiple difficulty levels.

Tom Newman
profile image
This is way too general. Difficulty settings have always been controversial, but the rules are different based on primarily the genre. There are many games where I do not feel a difficulty level is appropriate. Many games are designed well enough to let players of all skill levels enjoy the game, like many adventure and role-playing games, while other games like first-person shooters actually require a learned skill that some players are better at than others, and a lower difficulty is justly warranted for the average player to complete the game.

I prefer the designer to give me as little choice as possible, so I oppose difficulty levels for the most part, but if a game requires a learned skill set that improves with experience (like FPS, many sports games; etc.) adjustible difficulty is welcomed.

Video games have become way too diverse and wide spread to say all games should do one thing or another. I understand both positions, and agree with both, but only in specific circumstances.

Jonathan Hartley
profile image
I'd like to reverse the idea suggested by 'Anonymous 15 May 2008 at 2:22 pm PST', that hardcore players might choose to attack more enemies and be rewarded for it, while giving softcock player the chance to sneak around them without fighting.

Maybe the softcock approach should simply be to wade in with big guns. Hardcore players might choose to find an alternative (more skillful) solution. (sneaking, without using weapons, using ammo to achieve some other goals, etc)

Chris Crowell
profile image
For Rumble Racing, we had what I considered to be a very believable DDA system (which was used in conjunction with a global front end chosen difficulty level), generally based around the very human concept of of making mistakes. The chance of the AI making a bad decision was increased when and where we wanted. For example, we created ways for the AI to make bad decisions by laying down multiple racing lines and flagging them as fast or slow. The sliding variable affecting the choice decision was dependent on the chosen difficulty level, where the player was in the race and in the game progression. To the player, the other cars simply went around corners smoothly or clumsily, which looked just like human drivers making small errors. This and many other hidden systems created a game that constantly felt challenging, but not condescending.

Its not easy, but the essence of games is discovering and overcoming challenges. Its core to the experience and worth a lot of work to make those challenges feel right to the player.

Josh Collins
profile image
One problem I find with difficulty settings in games is that they often adjust setting which strongly effect the immersive feeling of the game. An (perhaps extreme) example would be a war game in which on easy mode you can be hit with a rocket launcher and survive whilst on hard your own rocket won't kill the basic foot soldier. The Entire experience of these two modes will be completely different. A better solution would have the player always feeling like a normal, non-super human soldier fighting other normal soldiers in which the odds that your up against get worse as the difficulty level increases.

Alexander Filipowski
profile image
RPGs and Sandbox games such as Grand Theft Auto are best played without a difficulty slider, because there are usually ways of making your character stronger for difficult sections of the game. Difficulty sliders give greater choice, but unfortunately sliders are largely meaningless, because the player has no idea how easy/hard it would be until they give it a try.

The Matrix: Path of Neo had an interesting opening level in which by beating the level, you would gain the ability to play Hard mode. This way, it makes sure that if you can't beat this level, you wouldn't be able to stand up to the challenge of Hard mode.

Adaptive Difficulty sounds like a nice idea, but it also has its downsides. In Lego Star Wars for example, I can't even notice the adaptive difficulty in play, because it is just too subtle. With Wii Sports, I always end up with the hardest possible AI opponent in every sport. This makes it a challenge, but it's somewhat annoying that you can't just 'relax' and play a somewhat weaker opponent. It always tests your maximum capability, which you may not always be at.

Mike Lopez
profile image

In the past I have found many of your articles on design *theory* thought provoking but respectfully I think your opinions here are dangerously misleading as you try to argue against DDA when you have not *designed/implemented* a DDA system and have not even programmed for games in nearly a decade. Those of us who have designed and implemented levels, systems and mechanics might be better equipped to weigh in on issues of design production and execution (vs. theory). My apologies if it seems like I am hijacking things here but I think it important to espouse the dramatic benefits of DDA and refute the unfair assertions that do not make sense.

Although I have not read Andrew Glassner’s piece in question (I have seen him speak), I tend to agree with much of his statements re-printed above about the limitations of manually tuned difficulty settings alone. I do think that your arguments have fallen into the same pitfall as his in that this column appears to assume that DDA and difficulty settings are mutually exclusive; while historically it has most often been the case that they have been separate, their is no implementation reason why that must be so. Still I believe that with the properly implemented, multi-dimensional DDA system the designers can maximize the satisfaction of the most players and separate difficulty settings are probably not required.

What qualifies me as an expert? A: 16 years of level, systems (including DDA), and mechanics design. Back in 1999 I designed a multi-dimensional DDA system for opponent speed management in Road Rash: Jailbreak, but I also coupled that with different physics settings (arcade and sim settings to cater to different skill and preferences of handling) so in that sense Jailbreak had both the DDA and difficulty adjustments (albeit only 2). I also was able to leverage the DDA system to control the pacing and intensity of opponents during a single race so the intensity would build dramatically towards a climax at the finish line (assuming the player was doing reasonably well) and this ability to leverage a DDA system to improve the pacing is precisely why DDA is so powerful and why all my experience leads me to believe it is the wave of the future for top-echelon design teams.

Let’s review a few of your arguments against the Glassner piece:

“There's no reason that player-settable difficulty has to be limited to three or four options; it can be a slider.”

ML: It appears you are refuting your own point here against DDA. The only option for a difficulty slider is through a DDA, or possibly through a manually adjustable difficulty system, which would be at least 75% of the work of a full DDA IMO without most of the benefits. I am not aware of any games that have used a manually adjustable system without DDA and I see little value in implementing the manual system without DDA.

“Players want settable difficulty levels, and removing them for purely theoretical reasons is not a good way to serve your audience.”

ML: Players only want to set difficulty levels because the game balancing alone does not completely satisfy their interest, needs and abilities. A well balanced multi-dimensional DDA system can provide a much more satisfying level of competition, challenge and engagement and so a player should have no need to want to change the difficulty under these circumstances; even if you choose not to believe a player will not want to change them then you could still give a slider to allow adjustments the overall maximum difficulty in the DDA system although I submit this effort would somewhat wasted as it would only be used by a small fraction of players.

Let’s also dissect your arguments against DDA:

* Some players hate it.

ML: Players hate poorly implemented systems of all klinds, not just DDA and they especially hate heavy-handed DDA systems that obviously appear to favor the AI (I often read comments about how the AI "cheats" and poor DDA is often one of the more obvious problem cases) or that clearly weaken the player in relation. But with a well implemented DDA system the progression of difficulty can be controlled to deliver a much better experience, without the pitfalls of difficulty spikes or flat interest and the pacing/intensity of the gameplay can be leveraged through DDA as well. My 16+ years of level/systems design and Design Direction experience tells me DDA is the path of the future and I believe that only through DDA can designers deliver truly exceptionally paced gameplay (I will write more on this in one of my next Gamasutra articles in the next few months).

* Players can learn to exploit DDA by pretending to be worse than they are.

ML: So what? The GRAW team, in their GDC 2007 lecture said they spent a ton of development time, money and effort attempting to thwart the player from exploiting their dynamic difficulty systems only to come to the realization that it really did not matter since it only pissed these players off that they could not manipulate the system. The conclusion from the GRAW team (and one of the most successful DDA implementations): let the players play how they want. Period. [That includes gaming the system if they wish.]

* DDA doesn't work for all kinds of challenges.

ML: Agreed, but it can work for the majority of genres including shooters, driving/racing, platformers and probably even RTS and I would submit the compatibility would equate to the vast majority of published titles. DDA is an extension of AI and as such it should be leveraged to improve the competition and engagement and make for a much more satisfying experience.

* DDA can create absurdities.

ML: Fasle. I would correct that to read Obvious, One-Dimensional difficulty systems can create absurdities. I submit that rubber banding is one-dimensional and therefore not truly a DDA system and racing companies that are still using simple rubber banding are definitely in need of evolving their system to improve the sense of competition from AI opponents and remove the horrible anomalies.

For Road Rash: Jailbreak I designed and implemented a multi-dimensional system of opponent speed management that was subtle and exciting and delivered excellent engagement and competition for players at all skill levels without the obvious anomalies of rubber banding. Furthermore, I leveraged this DDA system to control the pacing of the race and to build the intensity of the race as the player did reasonably well or better and as they approached the finish line. This is something that cannot be done with a traditional fixed difficulty system and this is the reason DDA presents a unique opportunity to increase engagement and pacing in the future and the subject of a Gamasutra article I will be submitting in the next few months.

* DDA ruins pacing and obviates good level design.

ML: I suspect it has been far too long since you programmed games, Ernest. As one who has designed both levels and systems I agree with the comment poster above that this notion is utterly preposterous and I submit that manually tuned levels will always be inferior to well implemented DDA tuned levels unless the investment in level tuning is very, very significant and well planned/coordinated (e.g. Bioshock / Halo 1). Levels and missions of today require so much effort that teams often have 6-12 level designers/scripters making levels for the game, each typically implemented and tuned by different individuals. This manual level implementation process is highly prone to difficulty spikes and poor pacing (especially between levels and designers) as the effort required to coordinate the tuning of all the levels as a whole is so cost prohibitive. With DDA the system should be in place first and a high level difficulty plan should be created up front to understand how the difficulty should progress from level to level (and with which mechanics or weapons should be available for each time - see my Gameplay Progression article for more details). Then, when the level designers are implementing levels, they call difficulty values from the DDA system based on their position in the overall level plan. When the levels are all complete it is then easy to tune the levels as a whole from one to the other. The level will still utilize fluctuations in pacing through natural means (e.g. shooter battles are sequenced with lulls) but the amplitude of pacing and difficulty is controlled through the DDA system and the result will be a significantly better experience for the player.

Regarding your closing statements I definitely do agree with your points about setting a maximum difficulty and about the need for DDA to be subtle and hidden and to only affect the AI and never, never appear weaken the player (even if only in relative terms). However, as should be obvious by now I definitely disagree with your opinion of making DDA optional and feel the only reason a designer would have to make DDA optional would be because the DDA system was one-dimensional/obvious or poorly designed, implemented and/or tuned and did not properly engage and satisfy the player.

Stay tuned for a more thorough analysis of the benefits of DDA systems to come in the next few months when I find the time to finish a Feature article on this very subject.

- Mike Lopez

Road Rash, James Bond, Scarface, Baja

Mike Lopez
profile image
For reference to my previous article on Gameplay Progression, here is the link:

Ernest Adams
profile image
Mike: And can I mention again that you have 16 years' experience in the subject, and can I say again that you're writing an article about it? ;) I'm still waiting for that article, by the way.

My column isn't an argument against DDA. It's warning to developers that they'd better know what they're getting into before they commit to doing it. You have 16 years' experience; bully for you. Not everyone who reads my article will.

It's also a counter-argument to Glassner's demand that difficulty settings be "banned." They should not. Even you would have to admit that it's easier to build a difficulty setting for many games than it is to build a DDA system. *I* have 10 years of experience consulting for companies building casual games with small teams on short deadlines, and under those circumstances you bet I warn them not to get into DDA unless they have someone experienced on staff. They seldom do.