(This blog post was originally a talk I gave at the winter world of love event, then it spent two years in my blog’s ‘draft posts’ bit, I found it tonight and tidied it up since it’s still something I get super excited to talk about 🙂 )

(warning: links in this blog may be spoilers, including THAT spoiler from FF7)

Surprising an audience is something games can/could be super good at, but I think when we are making games we don’t think about surprises enough. I’m not saying it’s something we should all focus on but it’s pretty interesting and I think we should keep it in mind from time to time 😉


What is surprise?

First things first, we need to know what we are talking about when we say ‘surprise’. People often don’t realise it, but surprise is an EMOTION!

When we are talk about making something in a game surprising, we are talking about making something that results in this emotional response.

The second most important thing about surprise, is it only lasts for a very short period of time, otherwise it’s something rather different that we call ‘shock’.

Off-topic but don’t-want-to-delete paragraph! – I’m not really going to go into shock here, not because I think surprise is better or more important (IMO shock can be even more valuable), but because there are no shortage of games trying to shock anyway, and it’s much harder to shock right than surprise right. I’m pretty sure it’s not hard for you to think of a few games that are trying to be shocking, and whilst there are some where the game is better for it, I feel they are in the minority. (also just a minor thing, I think sex shocks can be good, IMO the sexuality displayed in mighty jill off is probably shocking to some, but it’s valuable to the game, Jill would do anything for her queen, wheras you would never see some GTA protaganist climbing a tower of death for a hooker.)

anyway, that’s enough about shock, just remember that it’s easy to do wrong, tricky to do right. 🙂

So, where was I… oh surprises don’t last long, BUT, the memory of them can last a very long time, sometimes even for a lifetime.

They are wonderful because they takes us to new places emotionally, places we never even expected to go. At the very least, this is good for keeping things fresh. In games we often find ourself in the position where players will get bored after a while, but with surprises we can take what the player is expecting and subvert it. This changes the way the player relates to the game. From their perspective, the game has actually changed, and they are now less certain what the game will present them next, it has more possibilites!

It’s important to note that I’m not saying you should take a boring game and put in surprises here and there to keep the player glued to it, that’s actually EVIL. (and yeah, I’m going off topic again here, but whatever) your priority should always be providing a valuable experience for the player. If your aim is only to hold the players attention, then you don’t really care about your player and you should do something else with your life. Ultimately you are creating something that will take up people’s time, and time is a limited resource for humans, so make sure if you are taking it you give something valuable back. (so just don’t be an asshole k? thx)

Surprises are not about random drops (though you can surprise people with random drops), they’re not about some unknown reward (though sometimes they can be), what THEY ARE about, is making people feel something.



Let’s get academic, or at least pretend to (since academic types are way smarter than me). The University of Southern California had some clever people to look into surprise and they then had some clever things to say about it.

First of all, they demonstrated their exceptional coolness by deciding that 1 unit of surprise should be called a ‘wow’. They then proceeded to measure people’s wows per second when exposed to various input data… none of that is really too relevent to us but they were very clever about it so I trust them when they say things that do apply, such as:

For a definition of surprise, we need 2 things, There must be uncertainty (this is extra important to us because we actually have a strong influence over this) and we must realise people react to the same situations differently (less important to us, but do keep it in mind)

in other words, so long as a person isn’t omnipotent you can surprise them, and when exposed to even the exact same stimulus different people will react with varying levels of surprise.

The USC boffins then go on to say “an event is surprising when the distance between posterior and prior distributions of beliefs over all models is large”. And if you translate that into English: Surprise is a big change in what you expect.

For example, I can show you a black screen. The longer the screen is black, the more you expect it to remain that way. When the screen turns white, it’s something you don’t expect and you are surprised! OK… you aren’t surprised much, it’s not a great difference in what you were expecting. but if the black screen instead sprouted limbs and started doing the macarena… the odds are that would be an example of a greater change in your expectations and therefore a greater surprise to you.

Here, I made a graph for you:

surpriseGraphAt the start, the person has little or no expectation of X occouring, then at some point X does occour, their expectations change rapidly and they are surprised. Then following the occourance of X, whilst it is not occouring the person expects it less.

I’m aware we seem to have moved away from games, but we have just gone over the most important sciency bit that is most relevent to us. You see the axis on that graph, with ‘expectation’ and ‘time’?

We can influence both of these things.

So my point, (yes we have finally reached it!) is that:

Engineering surprise is achieved by engineering people’s expectations.

But what if people are expecting the surprise? This is what we in the human-feelings business call ‘anticipation‘. I’d argue that anticipation is also a type of surprise, just with different timing. since you are expecting something that doesn’t happen, each moment it doesn’t happen is a change in expectations, a tiny surprise. But since it’s many tiny surprises drawn out over a period of time, it can have as much of an effect on players as a sudden surprise.


When people are anticipating stuff you can give them what they are expecting to relieve them… or make out like it was never going to happen (the killer wasn’t really behind the curtain, nothing to be afraid of I guess…) and/or then throw it at them a second later anyway (the killer was actually behind the door! *stab stab stab*). This way you get all the cool stuff of building wows gradually with anticipation, and also dropping a bucketload of wows on people at once.

Playing with expectation and time is what it’s all about really, for both surprise and anticipation 🙂


Point made, The following paragraphs are disorganised! \:D/

Even the great so-called ‘masters’ of surprise, the ninja, are WEAK in comparison to our power. A ninja must adapt to it’s surroundings, play to whatever their victim’s expectations are, and make their move at just the right time. For us game developers however, we don’t have to adapt to the surroundings, we own the surroundings. we don’t have to play to our victim’s player’s expectations, we create their world, their expectations are built from the experiences we have already given them.

Therefore game developers are better than ninjas. >__>

You know who the *real* masters of surprise are? Magicians! A good magician is in complete control of their audience’s expectations (just as we can be). Every word, smile and flourish leads the audience a certain way, sometimes the magician gives the audience what they expect, sometimes they don’t. To a magician, the audience’s expectation is like a toy, or better yet, a tool. Magicians use this tool to engineer that sense of wonder the best magicians are known for.

We can do better!

Magicians are still limited by reality, I can pull a rabbit out of a hat in a virtual space, I can do it with a million rabbits in under a second. Our virtual world exists not in reality, not even in the computer, but in the mind of our audience. Magicians have to conceal the reality of their tricks, whereas we are kind of the opposite, we create a reality for our tricks. In this way our tricks can exist in any world we can make the player believe they are happening in. The conventions don’t exist until you make them, and you can break them later, just to make the player feel.

That’s not to say every player is a blank slate of expectations when coming into your game the first time, the reality is quite the opposite. But you can work this to your advantage too. There are many conventions players take for granted and you can easily have fun subverting these. Things like walls being solid (you can walk through that bit!), always jumping the same height (jumping in sunlight makes you go twice as high!), conversation won’t advance till you press the button (make NPCs act concerned about the player spacing out after a while!) and so-on and so-forth.

And let’s not pretend you don’t have any impact on expectations of players before they start the game! The name of your game, trailer, screenshots, description and so-on all paint a picture in the mind of people who are going to play, you can use this to your advantage by planting ideas and expectations that the game then subverts. Remember all those metal gear solid 2 previews with snake looking cool and doing snake stuff? then people got the game and half of the players didn’t see snake for hours! Kojima probably had a grin a mile wide from all the “wtf!?” he caused there.

Expert level expectation-engineering: make an entirely different game and send that out to reviewers, use it for screenshots and all other stuff! or give different players different games all together, tell no-one!

And you can do it too both in your game and elsewhere; you can influence players’ expectations and surprise them where it suits you most!

( Slight disclaimer/warning: getting people to expect one thing and doing another is a deception. It’s pretty easy to argue it’s OK for our purposes and all, but it’s still important to know when you lie, and make a judgement call on whether your conscience can handle it 🙂 )


  • Surprise is an emotion
  • It’s over quickly (otherwise it’s shock, not surprise)
  • Big changes in expectations make for big surprises
  • We get to decide what is expected
  • That starts before anyone even starts playing our game

Extra stuff to note:

  • Presenting things consistently will lead the player to expect them
  • Surprises don’t have to be big, you can use them as small accents wherever you like.
  • “The game isn’t over” is always a cool surprise, give people more than they expect 🙂
  • surprise is measured in units of “wow”, that is crazy awesome!

End of post! thanks for reading! 😀

I’ve kind of run off a bit with this blog, and missed some points so here is the original notes for the talk (.txt), the second draft (.odt), and the presentation slides (.odp)

(I promise I tried to present this post well, but I kept getting super excited because this is so interesting, it’s hard to stay on track and explain clearly when you’re in love with a subject, double-thanks for reading it anyway!)


2 Comments on “SURPRISE!”

  1. Mabi says:

    Great read! I always say that what I’m most looking for in games is surprise. Even when my expectations are not tricked, like being fooled by symbolysm or creating a calm mood to throw a zombi in my face. The simple act of discovery, let it be cool game features or art/story twists is what keeping me hooked from game repetivity. This can go from realizing you can high-five your friend in Tojam and Earl to finding an abandonned mine in Minecraft. I guess these kind of surprises should be measured in units of ‘cool’ instead of ‘wow’ if that make any sens 🙂

    This article will get me to think more deeply about building surprises for better impact, thanks!

  2. arkades says:

    Just as a small addendum: that bit at the end, about completely deceiving players being something to gauge with your conscience? I should also note that if people are paying for your game, and receive something entirely different than promised, you open yourself up to liability (fraud, specifically, I believe). I don’t think anyone’s ever charged a game dev with fraud before, but if one pulls that big a bait-and-switch on millions of people, it wouldn’t be entirely unlikely.

Leave a Reply

Your email address will not be published. Required fields are marked *