Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
November 1, 2014
arrowPress Releases
November 1, 2014
PR Newswire
View All
View All     Submit Event

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

Difficulty Curves Start At Their Peak
by Jon Brown on 09/22/10 08:55:00 pm   Expert Blogs   Featured Blogs

The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.


A large part of game design, especially towards the end of development, is concerned with what we call "balancing". This term refers to two distinct areas of the game:

Firstly, there is the requirement to balance various elements in the game against various other elements. If you're making a driving game this means all the vehicles in a class should be competitive against one another, although they will require different driving styles to make the most of them. Making an RPG? Then you have to make sure that each character class can play a meaningful role in a larger party. And there are examples like these for most genres.

Secondly, and this is the lion's share of the work, there is the requirement to balance the player against the game. This process is normally approached as one where we look for difficulty spikes in the general flow of the game, levels, areas and encounters that the player finds hard, in comparison to the surrounding experience. Our aim is to create a smooth(ish) difficulty curve that climbs towards the sky. But do we really understand what we're trying to accomplish with difficulty curves? And why is there one spike we don't design out of the system often enough?

When you sit down to design a console game there is one given that's taken as red, and that's that the game will get harder. As an industry we are all about the challenge, and in console games that means a progressively tougher world for the player to interact with. Through the introduction of tougher, smarter, faster enemies and more cunning, increasingly lethal hazards, game and level designers dial up the challenge as the game progresses, creating a planned difficulty curve like the one below.


Standard Difficulty Curve


It's well recognised that difficulty curves should have regular peaks, often representing boss battles, which are crescendos to shorter passages of play. These peaks are actually created by dips in the level of difficulty immediately after them. These dips allow the player time to enjoy their victory over a boss and give them a heightened sense of mastery in the afterglow of the encounter.

Of course, the player isn't a static component in all of this. They become better at the game as time goes by, learning skills and techniques that they can use to progress. In fact, we bank on them getting better, if they didn't then the jumps and shoot-outs later in the game would be a source of instant frustration. If you put these challenges at the beginning of the game then controllers would be thrown, rooms would be stormed out of and the game would go unloved. We also equip the player better in many games, giving them faster cars, more lethal weapons and upgraded version of their avatar.


Player skill graph


This is part of the balancing process and if we subtract player ability from the game's challenge then we see that the difficulty curve that we are really aiming for is like this:


Difficulty Minus Player ability


What we're aiming to accomplish is an experience where the player is always challenged just enough for the tools they have been given and the skills they have learned. If we don't increase the complexity of the game then the player will soon gain the upper hand and have total mastery of the world. And lack of challenge equals boring to a great swathe of the vocal hardcore gamer types.

However, there is one part of the game where we often load a considerable amount of challenge, and its the worst place to put it; at the beginning. This would be fine if the challenge was a result of a flash bang James Bond beginning, but the cause is the most mundane and dull part of picking up a new game - the controls. At the beginning of every console game there is the Satterthwaite Peak (SP) - named after Andy Satterthwaite, who pointed out that my challenge over time graph was missing it.

The SP is the point at the beginning of a game where you know none of the controls. The height of the peak will depend on how similar the controls are to something else you've played, how well the buttons are mapped, whether things operate the way you would expect and, most importantly of all, how many buttons there are to learn. Learning is tough and takes time to accomplish, it's not an activity that most people associate with entertainment.

Factor in the SP and we end up with a difficulty curve like this.


The Satterthwaite Peak


Obviously, most teams do their level best to ease the learning experience but many still make the fundamental mistake that their control schemes are too complicated, with too many buttons to remember in the first place. Not only that but the real joy of many games is tied to learning and wielding these controls with a good level of skill, which is hard to do when you can't remember what all the buttons do. Driving games have long been popular for the simple fact that they offer considerable depth of play with relatively few controls.

However, a new kid arrived on the scene recently and made driving games look like the fat kid with glasses (me) when the sports team was being picked. These are social games, of course. These games manifest all their challenge in the form of grinding, so there is little in the way of texture to their difficulty curves and the only real balancing to be done is making sure that the player can accomplish tasks regularly enough to keep them interested.

But it's the controls that make these apps killer. If you stumble upon a browser based game then you already know how to click on things, otherwise you wouldn't have got there. That means you already know all you need to know to control the game. There is no SP:


social games have no SP


Looking at the graph you are probably thinking two things:

Firstly, why isn't challenge at zero? As I have already stated, social games have considerable challenge built into them, the challenging of playing the game regularly. If you think there is no challenge to this then get a garden and keep it tidy. Although there is certainly some skill in gardening if you go for tricky plants to nurture, the majority of gardening is grind, and many folk find this fulfilling.

Secondly, you might regard the lack of undulation as a lack of excitement. Well, this isn't a graph of excitement over time, it's a graph of challenge over time, a graph of excitement would show many, many peaks as players clear tasks and acquire various items they have been gunning for. What you can infer is a lack of frustration in the system, as there quite definitely aren't any snag points on a flat difficulty curve.

Do not, for one minute, think that I am being patronising or condescending to social games here. I like social games, I play social games and most importantly, I think they're games. This graph does not represent boring, it represents optimised, they are games that deliver an experience of the utmost smoothness. Players can glide into these games and start playing them with the skills they already have. At no point is their progress barred by a jump they just can't make or a race they can't quite win. Their challenge is to keep coming back.

However, I am not suggesting that all games should be like social games, there's enough money, talent and players floating around for us to make all kinds of games. What I am saying is that the more complicated your controls are, the harder your game will be at the beginning. So why not cut out a few of those buttons, save yourself time and money on your tutorials, and spend it on the rest of the game?

Related Jobs

Rensselaer Polytechnic Institute
Rensselaer Polytechnic Institute — Troy, New York, United States

Assistant Professor in Music and Media
The College of New Jersey
The College of New Jersey — Ewing, New Jersey, United States

Assistant Professor - Interactive Multi Media - Tenure Track
Next Games
Next Games — Helsinki, Finland

Senior Level Designer
Magic Leap, Inc.
Magic Leap, Inc. — Wellington, New Zealand

Level Designer


Andy Satterthwaite
profile image
Great article - and not just 'cause you mention me.

However, I would like to point out that the "Satterthwaite Peak" (lol) is not just based on controls and control familiarity, but also on paradigm familiarity (though the two are linked of course)

The driving genre you mention is simple to grasp, not just 'cause of a low number of controls, but also because Left steers Left & Right steers Right; (and a couple of arbitrary buttons control accelerate and brake) it's not much to learn, and if you can actually drive then it's just a remapping of controls (move stick left, instead of turn wheel left). The paradigm is familiar so the learning curve is smaller.

By contrast a beat-em-up is nothing but a nightmare of controls and arbitrary mapping. X means kick, Y means punch except when X&Y together mean something else, or a third thing if I'm pushing back at the same time .... AAAARRRGGGHH!! But not only are the controls complex, but the paradigm is unfamiliar. I don't know how to do BlackBelt martial arts, and even if I did, it wouldn't be through pressing buttons on any form of controller/steering wheel etc.

With that in mind, I'd argue that social games haven't got it spot on. Certainly the controls are simple, but understanding how to farm using a combination of coins, horseshoes, energy and god knows what else is not instinctive, and therefore there is an SP for newcomers.

Future social games need to be careful not to fall in to the trap of assuming that their users understand social mechanics, 'cause if they do that SP may be even larger than before.

Jon Brown
profile image
That is true, social games to tell you where to click these days and the current format that they generally follow would fall down quickly without these tutorials.

Charles Stuard
profile image
I'm not really familiar with Andy's ideas of what makes his "peak", but I wanted to throw out there the "interface" of a social game can readily replace the difficult controls seen in console games. Most console games, especially nowadays, are going as "Hud-less" as possible, taking away confusing menus and random indicators. Social Games, on the other hand, are filled with these.

I imagine that any "social gamer" would have difficulty running GIMP, at least initially. Sure, they know how to click a mouse, but that doesn't mean they can use the program that mouse controls. The UI, the menu flow, all of these can be painfully done, and I have seen some really bad examples of it in social games.

...anyway, I'm basically saying I disagree with your final chart. I think social games have a very similar initial difficulty, scaled by how intuitive it is (similar to console games and their "intuitiveness").

Jon Brown
profile image
It should be noted that Andy is of the opinion that social games do have an SP at the outset. Like all good ideas, there is some contention in these early stages of debate.

Christopher Landry
profile image
Actually, in most social networking games the required amount of grinding increases with progress through the game, so I think it would still involve a rising line, complete with peaks (as it would be very easy to continue grinding when you are near a goal, but very difficult when they are distant). Furthermore, I would put forth that players' skill at grinding doesn't improve with time, so the composite graph would remain an increasing line, and thus be MORE variable than classical game design.

Jon Brown
profile image
That's a very interesting point, and right now I agree entirely. This is an evolving thing so expect to see me update it as time goes by.

Vin St John
profile image
Love this post, also agree with the above points.

One more point of my own to add to the mix:

Many developers feel it necessary to introduce all controls at the beginning of the game (or likewise, in the manual or help screen that a new player would refer to before they play). Of course, many more developers have gotten smart about only introducing controls as they are necessary - to some degree, this spreads that initial peak out into many tiny peaks throughout the game, as the player adds to their repertoire of available actions.

But game manuals and help screens never seem to do this. What I'm proposing is that these "quick reference" diagrams of an Xbox Controller or whatnot come in two flavors: the "beginning of the game" version and the "end of the game version", so I can refer to one when starting off and one when I've forgotten something I already learned.

Jon Brown
profile image
That's a really good idea, I expect I will use it at some point in the future.

Benjamin Hanes
profile image
I'd just like to mention that, when eliminating buttons, if you place their functionality back onto other buttons, you are making your game more complex, rather than simpler. Having to remember that pressing A can make you run, duck behind cover, or jump over a wall depending on the situation is more complicated than remember that X is run, B is duck behind cover, and A is jump, because you have to remember all those situations, as well as the button mapping. It also reduces precision, because someone might be trying to run, but winds up ducking behind cover instead.

Jon Brown
profile image
That's a really good point, I should include that in an expanded version of this post. Also, Tony Ventrice mentioned a few points about this (in relation to Mario 64) in his excellent article Evolving the Social Game: Finding Casual by Defining Hardcore (

Dean L
profile image
I don't think mapping two actions to a button makes it more complex if you think logically. If A is jump and A is climb, while B is crouch and B is drop when hanging, I don't see a problem at all. They're connected and they're not contradictory. Essentially A means upward motion while B means downward motion.

If buttons were mapped as per your example, then yes I would wholly agree. If A means move forward, then it doesn't make any sense to all of a sudden to stop me outright.

Jon Brown
profile image

It's important to remember that every action adds to the to total control complexity. Secondary functions on a button, even extremely logical extensions of functionality, like the examples you give, add to the complexity, and the number of things that must be remembered. They certainly don't add as much complexity as adding a completely different piece of functionality, or even mapping that extended functionality to another button, but they do add.

There's an excellent section in 21st Century Game Design (
4504293) that goes into the detail of calculating the complexity of a control system.

Dean L
profile image
I agree that adding more functionality accrues more complexity, however I do not think it adds anymore complexity than having a button for each function as long as the functions are logically distributed.

Jon Brown
profile image
Many thanks to everyone for your contributing thoughts. I plan to expand this into a much broader and complete theory at some time, especially in light of all the things I haven't covered or factored in.

Also, I think it should be called the Satterthwaite Spike instead of the Satterthwaite Peak, at it actually describes the thing better. So expect to see its name changed in future revisions.

Paul Preece
profile image
Games optimize for challenge and/or fun. Social games optimize for monetization and/or viralness. If you showed a 'normal' game's challenge graph alongside a social game's monetization graph I think they'd have a good chance of matching up quite well.

Jon Brown
profile image
Just to clarify on this point: Are you suggesting that Monetisation and Virality should be interchangeable? If so, that's an interesting point in itself - they're certainly related as many people spread the word of the game for rewards, rather than paying for them with cold hard cash. As for the shape of the curve over time, I expect it would be a bell curve, the playing figures for social games certainly suggests this. Good food for thought.

Shay Pierce
profile image

Great blog post, and great points. A great lens to view what social games are, which is something I think a lot of game developers are struggling with these days.

I for one think you hit the Farmville issue on the head: yes it is a game, yes it does present a challenge... and yes that challenge is: "showing up." Show up to tend to your crops on time, or you will fail. It's actually a challenge we all constantly face in our real lives; and as you note it's exactly the "skill" that is most important in real-world gardening/farming.

Jon Brown
profile image

Many thanks. I think the thing that social games probably lack at the moment is the genuine "I got lucky" victory. If there were more times when I just got lucky in these games then I'd be even more addicted.

Chris Cook
profile image
Excellent blog post, especially for a level designer who is trying to introduce the player to the controls in a tutorial level.

Do you think the Satterthwaite Spike will still be a problem with Kinect games? Will the spike perhaps change from mapping of controls to the limitations of the player-to-game interaction that the developers implemented? E.G.: "Can I interact with that?", "Will the system recognise what I am talking about?" etc.

Jon Brown
profile image
I think Kinect games certainly do have a spike at the beginning. Jason VandenBerghe spoke at GDCE about the problems the Red Steel 2 team had with the simple action of swinging a Wii Remote. You'd think that would be a simple action, but players have different interpretations of what 'swing' means, in the context of sword play. Any action that is well known to a player won't push the spike up by much though.