Articles

Design

Development of a Concept
Player or Game Failure
Difficulty Curves

General

PC/Console Dev Differences

MMO Games

A Look at MMO Games
Gameplay Issues
Consider Systems

Working Practices

Introduction
Communication
Bug Reporting
Ambiguity

Game Critiques

Championship Manager
Gran Tourismo 3
Earth and Beyond

Difficulty Curves

What is a Difficulty Curve

A difficulty curve describes the way that the difficulty of play changes during player progress over a period of time. It can be game wide, or focus on a much smaller section, such as a level or particular set of obstacles.

It provides two main pieces of information - how difficult something is, and how the difficulty changes over the course of the period.  It's an abstract concept, so it doesn't have units, but it can be used to provide relative information to help create the correct level of challenge as the player's skills rise.

The main importance attached to a difficulty curve is the shape, although the level of starting difficulty is also worthy of note.  It doesn't matter if the curve is perfect if the overall difficulty of the game is ridiculously easy, or ridiculously difficult. 

Analysing the difficulty of a game also allows the developers to alter pace and mood during play.  After a period of intense difficulty and excitement, the difficulty can be toned down a little in order to give the player a breather, or the game can suddenly become significantly easier to encourage an "it's quiet... TOO quiet" feeling of impending danger.

It is important to note that when we talk about difficulty, we are including everything that makes something harder to do, whether directly or indirectly.  For example, if a player keeps getting lost, this may be included in a difficulty curve.  Exploring may not be difficult in itself (there may be no other obstacles for the player), but this would still effect the difficulty of the game.

Another important thing to remember  is that they are only supposed to measure the difficulty of the obstacles that the player faces.  Games exist to provide entertainment, and difficulty isn't the only thing that affects that.  For example, a section of your game might be intended simply to progress the story, and so may effectively have no difficulty whatsoever.  If you are not careful, and include this "easy" section in your curve, then it will look like there is a problem.

Although in this document I do use actual curves for display purposes, in most situations you would not do so.  The resources required to be able to construct a meaningful graph would probably outweigh the benefits.  You can get most of the benefits with much less work. 

Granularity

In order to be able to create a meaningful curve, the granularity of the measurements should depend upon the scope of the curve.  For example, if you want to find out how difficulty changes over the course of the entire game, measure the difficulty of the levels.  If you want to find out how it changes over a level, measure the major obstacles etc.  If you measure all the obstacles in a game, and try and create something based on that, your graph is going to be a mess of jagged lines.  Even easy areas will have more difficult parts and visa versa.

As a section is looked at in finer and finer detail, the shape of the curve will become less and less useful.

The Aim of a Difficulty Curve

The aim of a difficulty curve is simple.  It's to avoid negative emotions created purely by the difficulty of the game, and encourage the more positive ones.

If a game is too easy, it can become boring quickly.  If too difficult, it can become frustrating or boring.

However, if the challenge is just right, the player is kept engrossed and entertained.

Determining the Difficulty Curve

There are two main ways of determining the difficulty curve of a game.

Data Analysis

This method tries to supply a data centric approach.  Developers can measure important factors about the game, and use this to determine the difficulty of a particular segment.  The information measured will vary from title to title - for example length of time might be considered a factor in difficulty with a puzzle game, whereas it might be just part of the game experience in a more exploration or story based title - but they should generally be based around the concept of success and failure.

When these factors have been selected, developers can use playtesting to provide concrete data which can then be examined to see how the difficulty of the game is going.

In this analysis, the developers decide that the number of deaths, and the number of times the player gets lost are important factors.

For example,

Looking at this, a developer might say that levels 1 and 3 need to be made less deadly, whilst level 4 is far too convoluted and needs to be made simpler.  Of course, such conclusions in reality would be determined by how difficult the developers think it SHOULD be.  Maybe one developer might feel that all the levels bar 4 are too simplistic, and that levels 2 and 4 should contain more dangers.

This analysis should always be accompanied by an opinion based report.  The right difficulty level is what works with your customer, so if your testers like a particular facet of the difficulty, then keep with it.   In the above example, you might find the target audience LIKES the fact that they get killed a lot and like the fact that exploration is hard.

In the end, each case has to be judged on it's own merits.

The problem with this type of analysis is that it requires "vanilla" players each time it is carried out.  To create meaningful data, testers need to have the same amount of skill and knowledge as the target audience.  If a game is marketed to people that have never played a similar game, and all your testers are hard core gamers, then the results might not be helpful at all.  This means that internal test teams aren't that useful for this.  This is an important point, as relying on incorrect data is likely to be more damaging than relying upon the experience of the developers and testers involved.

Gathering specific data may not be necessary in some cases, but if this is not done, then feedback is only opinion based, and like all opinion based feedback, is subject to error and bias.

The other problem with this method is that the game needs to be fairly advanced before this can occur.

Guesstimation Based

The second method involves using the expertise of participants to create an estimation of how difficult areas are likely to be.  This involves examination of the game design, and all aspects of levels or dangers.

Once this has been done, the developers should have a reasonable idea of how difficult an area should be, relative to other areas, and should also be able to determine whether level difficulty develops in the right way.  Since the results are based upon the knowledge of the people doing the estimation, this is an imprecise system, but it can be done very early in the design process.

In order to try and provide some flexibility, the designers should also come up with alterations that can be made to adjust the difficulty level if (as is likely) there is any variance between the intended and the actual difficulty of a particular feature.  These should involve as little extra work as possible, as this isn't something that can be specifically scheduled, and can almost be seen as a gameplay bug.  In many action games, one of the simplest methods is just to alter the amount of damage given and received by the various entities, whilst strategy games often give AI opponents different levels of resources to change them.

Both of these methods should be used in conjunction during the development of a game.  Although the guesstimation approach isn't very accurate, it makes the final tweaking much simpler, and less costly.

Typical Shapes of Difficulty Curves

In the examples I give, the curves are fairly simple.  When trying to create a difficulty curve, you have to be fairly general - you are looking for an overall shape, and attempting to identify areas that fall a good way away from the intended difficulty. 

Increasing Linear Difficulty

This is probably the ideal difficulty curve - something that gets gradually more difficult as time goes on.

However, creating a game which has a constant and linear increase in difficulty is very difficult, and even in games which constantly get more difficult, you are more likely to see a convex or concave shape.  If these curves become too severe, then this becomes a problem.

The game gets massively more difficult very quickly, then levels off.  If this gets too severe, then some players will be put off, and not get the chance to develop their skills.  The ones that do get past the original, steep part will often find the game lacks challenge in the late game. 

The game starts off with a very flat difficulty curve, and after a period starts to get much more difficult, very quickly.  If this gets too severe, the game becomes dull early on, then suddenly the player hits a brick wall which is very difficult to overcome.

Early Peak, then Trail Off

Originally, this article was intended to be about the difficulty curve in strategy games, but as I considered it, I decided to expand it.  However, many strategy games feature a rather bizarre difficulty curve.

Games start off easy, and then there is an incredible ramp in difficulty when the enemy is encountered.  If the player gets through this, there is then a long tail off of difficulty, until the game is incredibly easy (as the player is mopping up the enemy).  RTS games tend to do this in individual missions, whilst strategic campaign games often suffer from this during the course of the campaign.  It is common for players to know whether they are going to win or lose within a quarter of the time it takes to play a game. 

Fluctuating Difficulty

In real life, game difficulty fluctuates somewhat, but some games have enormous fluctuations that have a major effect on games.  Having a huge increase in difficulty can be difficult to overcome, but this can be even worse if subsequent areas are massively easier, sometimes even to the point of being no challenge at all.

All of these curves can be seen if the game is looked at in different degrees of granularity.  For example, looking at a game like Starcraft, a difficulty curve for the whole game may look fairly linear, whereas examining most levels generally give a curve that follows the "Early peak, then trail off".

Other Things to Look Out For

Spikes

Sometimes a specific situation can be so much more difficult than anything else around it that it provides a completely unnecessary block to the player's advancement.

Troughs

The opposite of a spike.  If a section of a game is much easier than all areas around it. Then it provides so little game challenge that it is almost useless from a play perspective, and the time taken to create this area is wasted

Small Changes

All games will have variations in difficulty as the game progresses, and if you create a difficulty curve with a large number of small spikes and troughs, then you are probably analysing things too closely.  Difficulty isn't an exact thing - what is difficult to one may well be easy to another - and therefore trying to be too exact can just be counter productive.  Remember, you are trying to get "the big picture", and identify major problem points, not examine every little feature of the game.

Learning Curves

A learning curve is a description of the way in which the player has to learn the game.  If a game has a steep learning curve, it means that the player is expected to take on a lot of information or learn a number of skills within a short space of time.  If the learning curve is shallow, then new factors come into play much more slowly.

Although learning curves aren't the same as difficulty curves, the two do bear a relationship with each other.  Whilst the player is learning new skills, actually using those skills can make a situation harder than it would otherwise be.

Summary

How the difficulty of your game changes over the course of play can have a dramatic effect on how much people enjoy continued play.  The best games provide a constant challenge throughout the game, keeping the player entertained and engrossed.  Whilst this article is over fussy in some of it's analysis, and deals with ideals as opposed to more real life situations, the concepts involved are important ones.

Difficulty is just one aspect of a game.  Like most components of a game, judging difficulty correctly can help towards making a great game, but it doesn't make a great game on it's own.  An uninspired game with just the right level of difficulty is still going to be an uninspired game.  However, the opposite CAN be true.  An inspired game can be ruined if difficulty levels aren't correct.