A Tale of Two Curves - measuring Player Engagement over Time

Today I'm going to share some insights about how to manage a hypothetical user's engagement (read: "interest in playing") over time.  Although I'll be describing some things in terms of equations and curve shapes - there will be very little math involved, I promise.

Driving Factors

What determines a player's engagement?  Let me throw out a few likely culprits:

• Familiarity
• Competition 
• Personal Mastery
• Information Deficit  (Wanting to know What Happens Next)
• Sensory Submission

Note that most of these can be responsible for either positive or negative changes in engagement - some players hate competition, for example.  Sometimes players want the comfort of the familiar, and sometimes familiarity leads to boredom.  But these five areas are a pretty good cross-section of the driving psychological factors at work.

Too Hard vs. Too Easy

Many discussions about Engagement over Time center around the word "difficulty."  If a game is too difficult, people get discouraged before they get engaged.  If a game is too easy, people lose interest quickly.  (I don't agree with these positions - I'm simply stating the common usage.)

Why would someone make a game which is incredibly easy?  Usually it is about increasing accessibility in order to expand the potential player base.  A refusal to rely on established norms is a fast-track to the "too easy" label from experienced gamers here.  But it's hard to argue with a larger player base - because more money from this game means a better chance of another game next year.

Another route towards "too easydom" is to create a game which impresses players initially, but then fails to grow the core experience over time.  These games are labeled "flashy" or "showy" in many cases.

Why would someone make a game which is too hard?  Often it's a game which is meant to be an extension of a previous game, or group of games.  The thought is that instead of starting at the beginning, this new "elite" game can just start at the end.  That approach actually works well for games with huge fanbases, I'm sorry to say.

Another interesting idiom is the "pleasure delayer" syndrome, where a game keeps you weak while taunting you with a promise of future power.  Experience gamers, or fans, who believe these promises sometimes demonstrate an incredible ability to plow through unfair / unfun / boring content just to get to the promised land of power near the end of the experience.

It's also very possible to make a game which is too hard or too easy simply because the game was not designed well.  We'll be ignoring those cases, since "shoddy workmanship" isn't a design technique.

Building Two Engagement Systems

The interesting observation I've made about "too hard" games is that they always seem to emphasize the end-game experience over the early or mid-game.  Conversely, "too easy" games seem to make their push initially, but then stop trying to engage players once their initial push has dissipated.  I thought a good solution would be to create a static level of engagement throughout the experience - keep player interest moderately high for the entire length of the game.

That didn't work at all.

A rule of thumb about human stimuli is that if they aren't changing, they get filtered out.  Players NEED that upward or downward engagement curve to be present.  So I came up with a different approach to designing engagement: Plan for two overlapping engagement curves.

The first curve starts at y = 1, and follows a strict downward trend over time.  The second starts at y=0, and follows a strict upward trend.  These two curves are measuring player engagement in response to two prominent systems, so the total engagement of the player can be represented as the sum of the two.

Curve 1: y = 1/x

The first system is designed to grab players right away, but isn't meant to create lasting engagement.  Whiz-bang graphics can do this- so can a strong license.  A big cinematic moment at the start might set this up, or a rapid escalation of player power.  The most interesting moment in this system happens as soon as possible during the game.

Curve 2: y = (x^2)/2

The second system is designed to be completely absent at the start of the game, introduced shortly after gameplay starts, and then build in significance in an exponential fashion.  Diablo 2's skill trees fit this bill nicely, as does the weapon progression during single-player Quake 2.  Almost every RPG fills this curve with a level/advancement progression of some sort.  Story-driven games use their narrative to build engagement in this fashion.

Remember that what we are graphing here is not power, or excitement - we are graphing Player Engagement.  Merely increasing your sword damage each level doesn't build engagement - but providing progressively larger game-altering skill choices does.  Foreshadowing (or explicitly telling) about these future choices serves to build interest and engagement, even long before the future choice has any in-game realization.

Systems-Level Game Design

What we really need, in order to realize a two-curve model of engagement for our projects, is a view of our game as a series of interconnected Systems, rather than "moments" or "features."  A System is a set of grouped challenges and responses, defined by the rate of change within those challenges AND the rate of change within the available responses.

As an exercise, let's consider how we might construct a system which satisfies engagement curve #1.  We want something which will instantly grab and occupy the attention of our players, but which will very deliberately become less and less engaging over time.  It should never end completely, but player should pretty much stop caring about it after just 15% or so of our game.

A character vanity system can do this - players love to personalize their avatars.  We provide players with a bounty of vanity options at the start of the game, but then begin to water down the rate at which they acquire new choices for the customization system.

Systems which meet the curve-features of equation #2 are thick on the ground - so I don't see a need to describe one in detail.  Skills, bosses, story crescendos, combo-depth - lots of classic game systems fill this need.

The Inflection Point

Once your two systems are in place, the last check to make is about where those two systems cross on your virtual Engagement Graph.  That's the point in the game (remember that the x-axis is time) where players stop caring about curve #1 and shift a majority of their attention to the system generating curve #2.  In City of Heroes, this shift takes place at character level 12 - when significant new power options begin to open up.  In Grand Theft Auto this shift takes place the moment players stop following new missions - and just begin to play in the sandbox.  Players who cross this threshold are not just casual players - they are fans.  

This is what you are defining with your two Engagement systems - you're building fans slowly with curve #2, while you dazzle them with curve #1.  In a mobile game the transition might happen really soon - usually in less than 30 minutes or an hour.  PlayFirst made a good living for themselves by betting that players would become fans before their 60 minute trial was up.

When we talk about "hard-core" gamers, I think what we mean is that some people have played so many games that they trust that there is a curve #2 to take them to 70 or 100 hours of gameplay, and they hunt for it right away.  They are like the audience members actively looking to spot the magician's trick, instead of letting themselves be fooled.  When we talk about "hard-core games" we often mean games which have eschewed curve #1 almost entirely - making the game a very difficult Engagement proposition for anyone who does not commit to fan status immediately.

Invisible Systems

Engagement isn't the end-all of Systems Design.  There are many features and systems essential to a game's success which have little or no direct influence on the Engagement level of the players.  Players do not need to know how everything works - nor should they care.  But players will care about something - and it is an important task to define what they should care about, and make sure that those systems deliver the engagement players are looking for.

 

Comments

[HT]

I think you mean "Point of Intersection" not "Inflection Point".

[Simon Strange]

They should be equivalent in this case - the individual curves have no inflection point, but the sum of the two (the overall Engagement Curve) has an inflection point where the two intersect.

I wanted to emphasize the change in sign of the user's Engagement at that time, so I chose the term which emphasizes that change.

[Jeronemo]

Haha that's funny I was just going to say something about that.

I think it would be interesting to have a graph of player engagement and player enjoyment (for specific games of course). I can think of plenty of times where I played a game intensely for like and hour before I've realized that I didn't enjoy it, and also the reverse.

I don't even know if that's possible to measure someones "enjoyment level" that accurately, serotonin or adrenalin levels? ... BPM?

[Look for Trace-In Here]

The caption read "Marry me. By Nathan Archambault" As the images slowly sunk in, nate dropped to one knee and asked me to marry him. I was overwhelmed with happiness and almost forgot to say yes!