Sensing the Emotion in a Gamer's Head

David Perry

Game Developers Conference, 2000 In 2000, David shared his experiences after 1000's of hours of design meetings... He revealed his visions of the things we always wanted to be able to do in the 'future' and now finally (thank goodness!) have the technology to do it. Gamers' realtime emotions was the major topic, but he will touched on many other important psychological aspects. His goal in this high bandwidth talk is to inspire you to stop asking your programmer to code new gaming concepts and demand it instead!

(Editor's Note: You can order this speech on Audio Tape from the Game Developers Conference 2000 by calling: (858) 483 4300 and ask to buy Tape #91, or online at


This year David will be sharing his experiences after 1000s of hours of design meetings... He will reveal visions of the things we always wanted to be able to do in the 'future' and now finally (thank goodness!) have the technology to do it. Gamers realtime emotions will be a major topic, but he will touch on many other important psychological aspects. His goal in this high bandwidth talk is to inspire you to stop asking your programmer to code new gaming concepts and demand it instead!


Many new 'real' ideas that you could add to whatever you are working on right now.

Follow-up documentation:

  •  EMOTIONS (Ongoing list of emotions)
  •  SENSES (Ongoing list of senses)
  •  Books to read:
    •  MIND AT PLAY "The Psychology of Video Games." (I highly recommend)
    •  FLOW "The Psychology of Optimal Experience."


For this talk, I Planned to follow the gamers mind from purchasing a game until it ends up in his collection or in his trashcan.

It’s also a bit of an analysis of the things I keep forgetting to check in our games.

In my study I suffered through descriptions of 'Cognitive Dissonance' (pretending to like a task to avoid conflict), 'Intrinsic re-enforcement' (praise and admiration), things like that… I will be avoiding 'terms' as much as possible, and keeping it as real and useful as possible.

Before I start…

We MUST remember that our market is expanding rapidly

- People are playing your games in more and more countries

- They are playing on all sorts of machines with random configurations

- They have all sorts of random computer literacy

- They have all sorts of mental ability

- They have all sorts of random physical ability

- They have random frustration levels

Marketing people call them mass market. To a designer, I call it a MASSive challenge.

As Peter Molyneux said in his speech today, our goal should be to stop making "People Beaters", and to instead focus on making "People Entertainers".

I think, if we take their money, we have duty to take care of them...

Let's Start at BUYING...

- They have probably seen your adverts

- Read preview articles

- Waited whilst the game slipped :)

- Now it’s finally available

- They either pre-ordered it or rushed off to the store

- Parts with cash (if you are lucky, they got hint book also)

Let's say it's a PC game...


- He just spent his cash, he is now committed to your game.

- He can't WAIT to experience the game

- But there might be some bugs, so…

- He needs to go and check for patches

- Then of course we want him to fill those lovely Internet registration forms

- While we are at it, we can force him to watch trailers for other games

- Now it’s time for our logos, in-fact EVERYONE’S Logos

- And then all the legal and copyright messages

- When he gets to the menu:

- Default settings are often very poor

· Many games GUARANTEE you don't get the best experience on YOUR hardware using their default selection.

· Some will actually offer to CHECK the hardware… Then they report back… Yes your hardware is excellent, but don’t take advantage of it.

- Needless to say, if the programmer KNOWS what he has, then make intelligent defaults for him.

If he has a 400Mhz PC + 3DFX Voodoo I card = 600x480

If he has a 800Mhz PC + Nvidia GeForce, then = 1280x1024

If you are the designer, then it’s up to you to make sure this gets done.

You should guarantee the gamer the best experience based on NO selection.

You can even start getting clever if you like, like kicking up the brightness a notch if the clock says it’s daytime.

Then again, you would be surprised how many PC owners don’t know how to set the clock.

Now with getting into the game… Remember speed is of the essence…

It’s like renting a movie and having to fast forward through the FBI warnings and logos, it’s just annoying.

- Studies have shown when bored we can estimate up to 1.5 minutes.

- However we see 2 minutes as 4 minutes


- Remember we Used to ask for DMA and Interrupts (Yuck!)

- We ‘seem’ to have got past that with plug and play

- Yet we still use words like gamma, isp, DNS, URL, baud, direct3d HAL

- If you HAVE to have buzzwords, offer help from in the game, don’t make them have to head for the manual

- Your task for your programmer is to make it so that 90% of people never to need to look at this stuff.

That takes time and creativity and that’s your job.

So overall,

(1) Don't make people feel it's a cumbersome lengthy process just to get your game going.

(2) Don't let them accept bad defaults just because they don't know what they mean...

So now the GAME has STARTED

- Now watching the FMV intro...

- How many people here have a focus test for their intro?

* Where you ask gamers what the intro told them?

* Then ask them what the GAME is about based on what they saw?

* Ask them what gameplay they would expect after seeing the FMV?

Sometimes I feel that the FMV is the teams version of what they wished their game looked like.

Sometimes the FMV sets a high expectation that the game will definitely not deliver.

(1) FMV is very expensive per second, so it’s often too short to tell a story well.

- But it CAN set a really good mood!

- It CAN give great game tips (eg. Don't be scared, you can actually fly on the back of the dragon)

- So you are aiming for (Story / Atmosphere / Tips)

- If FMV does not deliver ANY of these points, forget it.



- He WILL make lots of mistakes...

- Make it obvious what he did wrong! – Let him learn quickly

- The key is BLAME... They need to blame themselves for mistakes.

- They should KNOW that they are smart enough to beat this.

A study was done by Kahneman & Tversky, it showed that the blame they place is based on their length of mistake. If they feel that they did 10 things wrong to make this mistake, their desire to re-play will be greatly reduced. If they think "Damn" I should have just put on the radiation suit, and they can re-load and try that quickly, the replay appeal is enormous.


- Offered a worn path in an open forest, people will tend to follow the path instead of exploring the forest.

- Use this to your advantage!

- It’s easy, just look though his eyes.

- Take the time to Imagine what he seeing.

- When the gamer gets creative and you the designer are already there, you make the experience RICH.

- If you are not there, the game will seem shallow

Play with things like, like voyeurism...

If you let the gamer think he is hearing or seeing something private, he will love it.

- Remember: People are like magpies, they are attracted to interesting objects and situations

- Use that to your advantage!

- Let them feel free, but anticipate all the things that they will try.

- Tease them to come and explore things that you have ready for them.


- Be careful to keep objectives clear and simple!

- Remember they could be playing in a Hotel, Bus, Train, Airplane, Naked..

- That means they can have distractions, at any time, for any length of time.

- So keep an eye on them, sense frustration...

- They might have forgotten an important message objective because they had to answer the phone.

- Look for inactive periods... If the person is playing, then hits pause for 40 minutes or leaves the controller alone, assume a distraction happened – YOU EVEN KNOW HOW LONG!

- TRICK: Timestamp game saves, so you know it's a week since he last played... Help him recover quickly.

- To touch on that point, make it quick to save the game and get out of the game. (Dinner is ready or the plane is landing...)

- Gamers pick the games they KNOW they can get in and out of fast to fill those moments when they have spare time.

- That is one of the many reasons why death-matches are popular.

- Other things you can spot from within the game... If your character travels faster by jumping than walking... You can sense that. Do you really want people jumping everywhere?

- If they keep pressing on a control terminal in your game, but get the message that it’s already active. Realize that they are running out of ideas.


- Flow is like driving when you go into auto-pilot…

- You end up a mile down the road but don’t remember the journey?

- A guy with a really hard to pronounce name called:

- Mihaly “C s I k s z e n t m I h a l y i” in a book called “Flow, the psychology of optimal experience” writes about a 20 year study he did.

- He put Pagers on people, checking to see their emotional state at random times during the day.

- The study showed that to achieve FLOW, you need feedback moment to moment

- Like in Rock Climbing... Watching hands move…

- Each movement gets you closer to your goal…

- And you can always see that goal.

- This means you can assess your position constantly

- Your mind will take control of the task for you

- This can lead you to the flow state of mind

- On the other hand, feeling lost or distracted instantly breaks the flow

Playing a song on an instrument is very similar... Once in the flow, the music and coordination just come out of you. But stopping in the middle makes it hard to just continue from that point... You often have to retreat back to recover.

You will find that the average human has the SHORT-term memory capacity of about 7 things. So you can give them 5 objectives and they will probably do just fine, suggest 12 and they will start forgetting their goals. It takes repetition to put that information into LONG-term memory. A great example of this is when you are expected to learn 20 icons to get going in a game.. It's just overload! Getting them out of short-term memory and into long-term memory takes time and usage, so hold their hand.

I have to agree with the findings of this research... I am a great believer in gameplay that lets you know your goal, and then lets you easily assess your progression to that goal with great feedback.

Most gamers will find that out of 2000 games on the shelf, only a few will get them into this mind state where they feel so comfortable that 4 hours go by in what seems like minutes.

NOTE: It's important that the challenge is serious and consistent for a gamer at YOUR level of ability... Climbing a mountain that then feels impossible breaks the flow. So does a brainless climb... The mind will wander and lose focus.

So, ask yourself what does YOUR game do to give the player a feeling that he ALWAYS knows what his goal is and then know moment to moment if he is getting there.

Along with FLOW goes re-enforcement (seeing your goals being met / pats on the head)… Studies have shown that positive partial re-enforcement is more addictive than continuous re-enforcement. So keep the pats, un-predictible, but semi-regular. Not consistent, anticipated or expected.

A GREAT book on this subject is "Mind at Play", "The Psychology of Video Games" by Geoffrey and Elizabeth Loftus.


- This phrase will be overused...

- You will get tired of hearing about it over the next few years…

- Especially now that Sony have their Emotion Engine…

- As designers, we WANT to stir emotions like Fear, Anxiety, Excitement, Surprise etc... At our control!

- Recently at an MIT talk I gave, there was a lot of conversation about making gamers Cry.

- I believe that Crying will come from truly CARING, not just LOSS.

- I think there is so much shooting in games today because we just don't care... Characters are so dumb and annoying!

- It’s more fun to shoot them than to hang out with them.

- Like a movie, feelings for characters must evolve slowly and that takes work.

- Obvious goal: Characters to do smart things (not walking into things and saying the same lines over and over)

- Characters need to learn from what they see or sense (eg. If they see you shoot out window, and climb through it, they also learn that new exit)

- If they smell smoke, they know to look for fire.

- They should also respond to what they see (If you act like an ass, they should say, “What the heck as got into you?”)

- Most importantly....

- When humans make decisions, they make them based on past experiences. SO SHOULD THE GAME CHARACTERS.

- The game characters should help you THINK...

- The key here is that you feel gratitude when THEY help you out of a tough spot or help you advance or make YOU look smart.

- In your game, characters should speak based on

· What they have learned about you (you are a psycho killer)

· What they have experienced with you (what you have done together)

· What that think about you (dangerous guy to be around)

- The point is that they have to LEARN...



- That said I am not going to get into AI discussions, that’s a topic in itself.

- The clear goal is to make computer controlled characters that are FUN to be with and that make the game more engrossing.


It's actually very very easy, but extremely detailed work!

- I am going to give you a real solution you can go back to your office and do tomorrow... This is not an academic research study.

- I call it a ‘Personal Response Matrix’....

- This matrix means that the characters can be as smart as YOU are willing to imagine situations that THEY can get themselves into.

- Imagining all the possible situations is really the challenge

- However there is a simple low tech solution...

- Prepare a spreadsheet for example a COP

- Along the top of a grid, put Yes and No questions:

* This character is ALIVE?

* This character is WEAK or DISABLED?

* This character is ALONE (with me possibly watching)?

* This character is with SIMILAR CLASS (like a cop with other cops)?

* This character can see a FRIENDLY CHARACTER?

* This character can see a FRIENDLY CHARACTER (looking threatening)?

* This character can see an ENEMY?

* This character can see an ENEMY (No Threat - broken legs)?

* This character just saw someone DIE?

* This character just heard a massive EXPLOSION?

* This character has a WEAPON?

* This characters weapon has BULLETS?

Just this example list has over 8,000 combinations...

Many are repeats and can be ignored, but consider the combinations over a few days...

- A healthy Cop feeling safe in a room alone with a weapon?

- Make him play with his weapon like a cowboy.... Give him personality.

- A disabled Cop in a room with a threatening enemy...

- Make him plead for his life.

- A disabled Cop in a room with a friendly character...

- Make him say… Help me, call a medic..

Now add memory to the equation...

What if you came into a room, killed lots of people and left... When you return, the people should all remember you...

- Now mix that with Touch, Taste, Smell, Hearing, Sight...

- "What was that? - Ah it's him again, that psycho is back!” as they run to hide.

This gives them personality by actions...

So going forward, you the designer need to really worry about - Senses, reactions, memory, personality, speech and common sense.

- Good news is that a few days of considering the combinations this way, with some imagination, you will have a longer list of cool ideas than you will ever have time to implement.

- Do not expect to CODE common sense… It has millions of parameters, even MIT has failed to solve this problem so far.

- So DO NOT expect one of your programmers to produce some brain simulator that will save you lots of design work…

- You the designer need to work out the experiences that can happen in YOUR specific game world and make sure they are creatively dealt with.

HAH, and you though that was it...

It will take more work than that to make a gamer begin to emotionally care about your characters…

Now you need a ‘Vocal Matrix’ also for that character...

- This lists all the possible things that this character can experience with or without your help and their vocal response to it.

For example

- Getting set on fire…

- Falling and breaking an ankle…

- Being saved

- Being captured

- In a bad smelling area

- Seeing something hiding in the shadows

If you use your imagination, this can be a VERY long list.

You must remember that it’s a TOOL to make you think about these combinations.

In the WAY they respond, you can inject TONS of personality...

For example if a leg is broken, a really tough cop would get on the radio and request help :"Man down, need assistance"... What would a scientist do? What would a soccer player do, or a referee or a manager.

If we all did this all well... We would have intelligent reactions to lots interesting situations, you would surprise the gamer that you are there with him...

This is not AI, the intelligence is yours injected into the world...

The AI is the character having this library of reactions and ideas and comments, that can make it much more rich as it learns and experiences things with you.

Now back to emotion... Leading to motivation...

If you save a girl from certain death (SHE KNEW SHE WAS GOING TO DIE)... Then that girl comes over, like a professional actress, throws herself to your feet crying... and with tears in her eyes, in photo-realism thanks you... I mean REALLY Thanks you... (so that you FEEL thanked)

I think you will feel... Good.

The reason it will play with your emotions is because it's all correct.

- It's not happened to you before....

- You did try to save her

- She would have died

- She knows she would have died

- So she is now emotionally attached to you...

- From now on her speech pattern and reactions to you will have altered.

If done right and is truly convincing, I think it will give you feelings that the "Well done, here is your next objective" message never gave you.


Remember that you the designer (with help from the programmer) ALWAYS know what is going on.

- If I save that girl, she comes over crying and thanking me and I shoot her in the head... Then learn about the gamer. Make people that saw this VERY scared of him.

- If the gamer shoots dead bodies over and over... Hmmm....

- Learn from watching the gamer interact with your situations, is he doing good or is he back-tracking a lot?...

- If he kills friendly characters and robs their body, then you see how his mind is working.

- If he is hacking the heck out of other players in a soccer game... You see he really enjoys playing dirty. (In a soccer game, I would love for the biggest guy on the other team to come over and really threaten me)

- If he is in a driving game and seems intent on smashing into other cars he is probably getting bored of your simulator. He is looking for new challenges.

- If you have a re-charge point and he keeps going back every two minutes even if he only needs 10%... He is showing little confidence in his ability.

Good example of emotion sensing

- Disney Imagineering once did a cool VR demo on a Silicon Graphics machine....

- You held a Flying Carpet joystick in your hand for Aladdin...

- Kids that rocked and worked the controls hard, got a different ride than grandmothers that just held on for their lives.

- Other benefits of people playing their way....

If you have a gamer shooting up your rooms... Like he has a fetish for shooting ever monitor and window and killing everyone... Make the game react to that... The idiot is leaving a clear trail of exactly where he is... Remind him that you know what he is doing. (He can hear guards talking about the vandalism) But give him some more bullets for the ones he wasted.

Actually while I am on this... A WARNING: If a gamer is having fun in your world... Let him do it... Make the game react to it, but don't PUNISH him for it. Don't make him a loser for having fun.

If shooting up the place is fun to him… Let him do it… Let him enjoy it… But then maybe make him realize that getting seen doing it is a bad thing.

I could go on and on with situations that you could pull a meaning from... But they are kind of obvious... You just have to watch how people are entertaining themselves inside YOUR game.

Then think of ways to make that fun that they are having a flexible part of the game.

If you want to get really advanced, you can start testing the gamer from within the game..

- For example if the player is standing still and an explosion gets generated beside him… You will find that his average response rate is about 1/5 of a second… You can time him and rate him.

- Over numerous tests, you can actually learn about the person playing your game and tune the game to their needs. Learning also about a gamers physical limitations is interesting…

- The eye can only sharply focus on a very small area (like a few words in a book)

- Periphery vision is great and will be attracted to movement, but is not good for detailed decisions.

- However, movement of their eye takes about 1/13th of a second (called saccades)… Then upon focusing the gamer cannot move his eye again for about 1/5th of a second (called fixation).

- As a game can run at 1/60th of a second here in the USA or 1/50th of a second in PAL countries, you can see how easy it would be to carefully distract and manipulate a gamers vision.

- You find that they end up focusing on the most important thing to them. If you experiment with games like House of the Dead, you can calculate and monitor the mental and physical agility of the gamer in surprising detail.


When the gamer wins the game, make sure you make them feel good...

- Don't rush to show them your credits

- Try to end on a revealing moment or with some humor

If they lose, make sure they feel that they screwed up...

- A great way is to pull back to show them just how close they were to that power-up or just how close to the finish line. That can be a great tease to start the “just one more try” cycle.


If you gave a gamer some really surprising or key moments, he is likely to tell people. THIS IS IMPORTANT.

- Half-Life did a great job of this because it created conversation

- Warm conversation about enjoyment is what REALLY sells your games

So before you save your design document out... Ask yourself:

What are the KEY moments that people will choose to talk about when talk about this game?

What are the things that they are never going to forget?

What have you learnt about them? How did that knowledge help you craft the game around their personal needs and afford them the maximum amount of entertainment possible?

Thanks very much for coming.


Post a comment

Thanks for signing in, . Now you can comment. (sign out)

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Remember me?

David Perry's Game Industry Map Game Design Book