To Err is to Play: Human Error and Game Design

[Music] game design just a little bit about me I'm Lu lead user experience researcher at Epic Games I have a PhD in psychology specifically in human factor psychology which is the psychology of how people interact with technology and I'm a Kiwi which means I come from New Zealand and also that I cannot fly if you like to get in contact with me you can reach out to me at that email or on Twitter or please I love to play games so there's all my various game IDs add me we can play together so today we're going to talk about era here's a academic definition of era it's a bit long-winded so we can shorten that down to things going wrong that you don't want to go wrong and that is an important thing about era is that it is that you don't want to go wrong and in

industrial settings when we talk about era we want to prevent all error from happening however the thing that's interesting about working in games is that we want to prevent some errors to happen the errors there are unintended that we think make the player experience worse but we do actually want the players to make mistakes that's part of the challenge different types of errors can be in fact not bugs but features and that's part of what I want to talk about today rather than just how to prevent these things so what do we know about error well it's been study for quite a long time it's an important thing typically it's been studied in transport to make sure that these things don't crash it's been studied in industry to make things sure these things don't blow up and it's been studied in the military to help these things blow up other things and there's been a lot of work around it and of course anything in psychology will have a nice model so let's look at a model that we can use to

talk about error as with most models in psychology this is a simplification but it's aimed at trying to talk about how people go about their everyday life and cognition they idea is when you encounter a problem for the first time what you'll initially try and do and see if it can be solved automatically without you really thinking do you have a solution to this that's fast and automatic and doesn't take much effort if you don't you'll move up to the next level where you apply if then conditional rules wishes to say if this happens then I should do this so it's not quite automatic it's a little bit more thinking but it is on the way to becoming automatic it's like this situation means I do this only then if both of those have failed you will you move on to what we would think of as thinking with logic reasoning rational thought this kind of slow processing that takes a while to do and has a lot of inefficiencies in it and the general

idea is humans like most other animals are satisficing we don't like to spend time and effort if we can get away without doing so so we like to operate at the lower levels if we can manage and it's really important that we do so these levels can be called the skill level the rule level and the knowledge level and within each of these there are different types of errors that can occur and what I'm going to do today is I'm going to go over the these types of errors and I'm going to talk about examples of how they can be prevented and examples of how they can actually be used and I do want to put a disclaimer upfront I'm gonna use lots of examples from lots of different games and when I use one that's in maybe a negative context I'm not saying they shouldn't have done this or this is the worst example it was just a good example the same way if I say hey here's a use of it this may not be the best use it was just a clear example I want to make that clear UX is sometimes accused of trying to make everything simple remove all the errors and that's not what we want to do we want to help games be the best so as

I said another title for this could be errors how to prevent and how to use them and that's going to be the structure of most of the talk also at the end of the talk I'm going to talk about a kind of holistic approach to dealing with this and why cheese is involved in that so to start with skill level as I talked about before this is the fast automatic level performance if you think about it in terms of fighting games it is what is often referred to as muscle memory where if you're operating at the skill level you're in that flow you can just hit the combo buttons and get the combo off and keep on playing you don't need to think about it that's skill level operation when you're playing at that kind of muscle memory level so when we talk about errors within skill level there are generally two types there's inattention errors and / attention errors and attention ones are the ones

that come to mind very easily when we think about people making mistakes and that's that there just was enough attention paid to something and something goes wrong / attention errors we think about less often but there when you're using these higher level cognitive processes that take a lot of effort but you're trying to apply them to something that should be automatic so again a fighting game is a good example of this when you're first learning a fighting game you're really really thinking about which buttons you're pressing very aware of all your actions and that actually makes it harder for you to learn and harder for you to execute them then when later on they become automatic and you can actually basically not think about them and go with the flow so the first type of inattention error is double CAPTCHA slip I'm gonna use these scientific names for them because it's helpful but you don't necessarily have to remember them it's more about what they're about this is where there's a couple of potential fast things you could do and because you are

distracted or overloaded the stronger one takes over so let's go for an example say you're playing fortnight and out of the darkness comes this terrifying pink beer now what you want to do is you want to hide from the beer so you want to you want to crouch now in fortnight crouching is pressing the right stick click it in but you know this beers really scary and you're quite distracted and under load and instead what you do is you press B which quite common is Crouch in a game that's a stronger automatic rule you Crouch you press B that goes into build mode not crouch and fortnight and then the beer takes you out this is an example of making a double captcha slip to prevent double CAPTCHA slips it's relatively simple this is why there is so much emphasis on standardisation and consistency if your game is consistent within itself or within frameworks set up by the industry then

double CAPTCHA slips are less likely to happen because the strongest rule is the right rule so if we look at this fidget toy the green button down the bottom if I asked you what it does a lot of you would say that it jumps or it confirms right because that's pretty much a standard not in all cases but it is in a lot the red button a lot of people would say that it cancels because again that's pretty much a standard the other approach you can take here is to allow customization customization is great for accessibility but it also means that people can bring along their own strong rules and set it up as they like so on the next slide very quickly what button would you press to jump in Nintendo it's different from the other games for them the other platforms and then it's not necessarily a bad thing but it is if you're not used to it but for Nintendo it's good because it's been consistent that way for a while so how can you use double capture slips well they get used

a lot of the time in any game that requires this fast-paced twitch action something like rock band will be throwing a lot at you and it's part of the challenge that you're to choose the right action under a lot of pressure a more kind of direct example was in Dead Space 2 the original dead space when you fought the boss in that the boss would grab you at one point and pull you upside down and when it pulled you upside down your control is inverted now that's gonna cause a lot of extra difficulty for you because these double capture slips are gonna take over you're so used to aiming that now you have to think about something that used to be automatic much increasing the load all right so it's double capture slips the next is emissions following interruptions or the walking out of your Martian outpost without your keys problem this is kind of like you're going along and you've been in ads Oh killing monsters for a while and you decide you need to go ahead in a quest but on the way to the quests you get interrupted by some more monsters and

then two hours later you're killing those monsters and you've never handed in that quest because that automatic thing you've been doing more often took over when you got interrupted so to prevent it we can of course put in are you sure prompts for you're trying to exit right now have you saved have you remembered to do this thing because exiting can be quite an automatic thing you can have in world reminders of activities like quest prompts or where to hand in objects that are pushing the player forward there's not much not many cases you'd want to use this so I'm going to kind of skip over that delay reducing control this one's kind of a very familiar error that a lot of us have imagine you're you're a wizard in Skyrim and you're in your lab and you're making a potion and you work out you need some Nirnroot and there's none there and you go out into the field to collect it and then you get out in the field and then you're like why am I here again and the reason

that this happens is that queue that you had back in your laboratory of no known route is now missing when you're in the field were there's plenty of it so that you walk into a room and you're like I don't remember why I came in here but I had a good reason that's this error to prevent this error that it's very common nowadays to carry affordances carry queues with us wherever we go in a game so an assassin's creed for example they always show the in context action you can take it any time what your current mission is this kind of thing in fortnight we also can do the same thing where you can pin quests or you can pin activities that you want to do this is to reduce this error you can take these cues with you so you always know what you need to be doing you can use this of course if you do want memory and people to have to rely on these higher-level processes in your games this is often used in adventure games or hidden object games

the games where you don't want these cues carried with them because part of the fun of the game is having a good memory and being able to overcome this usual problem of operating on automatic and therefore forgetting things perceptual confusion is a straightforward one it's essentially this thing looks like that thing I know this isn't a problem for lots of people work for me personally one dude in camouflage well it's quite like another dude in camouflage like the American soldiers on the bottom look like the Russian soldiers on the top right probably made someone really mad with that Americans on the top Russians on the bottom it really to me it doesn't stand out but it does to a lot of people again these examples are not hey this is bad of course you can make military shooters just an example of how someone could get confused and of course in battlefield they overcome this in some ways with the spotting mechanic where if you see an enemy you can hit a button and they get highlighted for your team and makes them stand out so to prevent

perceptual confusion you can use the sync shape silhouettes and animation so in in fortnight there's not much case that's chance that someone's gonna mistake that black night for the bright bomber they have very distinct silhouettes they stand out they don't look too similar from each other you can also use animation to make enemies move in ways that look evil or threatening and allies move differently or you can do highlights such as in Paragon when you target an enemy they would have a I red highlight through on them or you could see them through the world these are all ways that you can overcome these problems you can of course use perceptual confusion though you can add load spot the difference games do this all the time hidden objects games do it lots of things that look like other things and spies and ambush enemies are an example the bob from the old marathon games was a li you could have on your side but at some times they could be taken over by the aliens and they would run at you and explode they look the same so therefore

a lot of players would just kill every bob that they saw no matter if they were friendly or not but this is a way of using perceptual confusion an Xterra is interference error this is where two thoughts and actions enter and a kind of mixed up combination emerges so for example when my uncle was at a wedding and wanted to welcome and thank people he stood up and said I would like to wank these are often called Freudian slips but they aren't anything to do with wanting to sleep with your parents they are just part of how cognition works with if you have several things that could be a possibility they sometimes combine into each other so in this really it's having low task load and only one choice for tasks at a time is the best way to do it again this tends to help with other errors as well but in this case this is one of the most effective things to do but to use it is

slightly more interesting we do this again in games with a lot of similarity or high task load if you're in a quiz game what you'll often do is the questions get harder as you'll make sure that the options start looking very much like each other and you'll increase the time pressure so that this error is more likely to happen where someone's going to pick something that looks right fast quickly automatically without really thinking because they're under pressure to do so that's all the inattention errors so now we'll talk about the over-over attention once is only really three there's a mission which is you're paying so much attention to one part of a task that you miss another there's repetition which is where you'll repeat multiple parts of a task so maybe you'll go and open a door multiple times even though you didn't have to do it because you're just so focused on what you're trying to do and then this reversal

which is a little bit rare but you always feel really stupid when it happens it would be something like you know you have to throw you you have to throw some guns out but you throw the gun out and then you're because you're thinking about that gun so much you pick it back up again off the ground and then you're like I need to drop it again and then you put it out and a mission examples is where I've got this picture up here and that would be you're lining up a sniper shot and you're really trying to account for bullet drop and the way the characters moving and then you pull the trigger and it goes click because you didn't think to read you didn't think to kick that and hopefully you're paying so much attention to me you also didn't notice that that wouldn't count in this because it is reloaded but to prevent over attention onboarding helps a lot here and onboarding ramp that's smooth that doesn't overload the player too much this can be interpreted some ways by don't use tutorials or don't give a lot of information to new players but actually new players tend to need more

information it's more about gently introducing them to it rather than hiding it or saving it for too far later flow state is kind of the opposite of over attention it is the point you want people to be in if you want to get rid of over attention you can give skips here's a nice example from a recent game deadbolt where they've got this message popping up saying like you'll have a harder time to play and frustrated they're pointing at this over a tension problem but there may be can be problems with that so if you want to use over attention if you really want to make your game hard for people and they're supposed to be operating an automatic level and they're failing tell them to try harder do the opposite of use the Force Luke throws stuff at them point out obvious failings add time pressure Metal Gear Solid had this thing that you played on lower difficulties and died they put this hat on you that probably actually made it harder for people because it's reminding them that they're not very good god of war it had this

nice feature in it that if you died a lot of times it would suggest to lower the difficulty that's great except if you were dying from a jumping puzzle and you couldn't do the jumping puzzle lowering the difficulty wasn't going to help you so these suggestions popping up may have made you focus a little bit more and not relaxed so moving on to rule level errors here's an example of a rule level system from Dragon Age where you could set up your tactics for your squad to say if an enemy is elite then this squad member does this and that's what to think about how rule level operation works if-then rules again there's two types of errors here one is mistakes caused by good rules which are rules that generally work in situations but then you encounter a situation where in this case for whatever reason the rule wasn't appropriate it wasn't that you had a bad rule it's just the situation changed or you hadn't encountered this situation before

the other instances if there's you've formed a bad rule so there's been something poorly explained to you or something miscommunicated and the if-then rule you've formed is just inappropriate so the first type of good rule problem is first exceptions now this is simply that exceptions are exceptional and so when you encounter something that's different from what's usually worked up until now you can't have taken into account in your rule system and so things can fail we've all heard of Murphy's Law but actually languages law is more commonly encountered which is that everything that can go wrong actually usually goes right the world actually usually accepts your mistakes and accepts your errors and so you can sometimes get along with nothing going wrong and then suddenly things will will happen and the exception will come up everything falls over to prevent first exceptions have no

exceptions be consistent be standard similar to at the skill level the Dark Souls games in the soul series for all their reputation for being difficult are fantastic at this they are deeply consistent games which is how people can play them and overcome them enemies move in the same way they attack in the same patterns that always in the same place your character moves in predictable consistent fashion that does not generally have exceptions this lets you overcome this problem and learn good rules you can of course use first exceptions by adding exceptions this can be a way to break up the gameplay pattern you know the classic example is having fire type enemies or ice type enemies that are resistant to different things to force the player to change up their tactics an example from destiny is the vex enemy that gets shot with criticals in the stomach rather than in the head the swaps up how the game is

played how the flow of game the game goes depending on what enemy that you encounter and makes the game feel a little bit more dynamic and means that these exceptions are shown to players and become part of their repertoire another reason why a good rule can get misapplied is due to signal retention or counter science no science and information overload detection if you look at these pictures to the right there the llamas in there are kind of hidden by a lot of noise and if you're color blind you may not be able to see llamas in all four squares this is a colorblind test that we have in fortnight for the colorblind mode so you can adjust things until you can see the llamas and this is generally the idea of the taking a go signal from a no-go signal which is like an alternate activity you could take or from noise so you come into a room in a shooter the objective in that room is to go press the button to open the door but a lot of enemies come in and previously in the shooter maybe you're used to killing

enemies to clear rooms but these enemies are going to keep coming until you press that button the press the button would be the go signal the enemies would be the no-go signal don't spend your time shooting enemies and then all the textures and lighting and everything would be the noise that you have to detect these things amongst so to prevent this clear teaching making it clear what the player needs to do prompting in the world highlighting information making your go signals the signs really stand out from the alternatives to use counter science you do just add noise make it a noisy environment do small differences between right and wrong again this is very common hidden object games but also for invisible enemies that often have a noisy pattern and move amongst the textures and don't stand out another reason that good rules can get

misapplied is because of rule strength general rules and rigidity these are three categories but they're very similar to each other essentially the rule that you use the most becomes the strongest and therefore becomes the thing that you rely on that's because generally speaking if a rule works and if it isn't broke don't fix it like good enough is good enough and if you've ever seen this in your games it's relatively common that say players will come up with a dominant strategy very early on and then they'll stick with it they'll use 400 ammo in the gun that they like rather than swapping to the gun that would have been more effective in a situation that has just come up so to prevent this you can either make multiple viable paths so this never comes up like whatever tactic they want to take always works or if you do want people to task which you need to provide really clear feedback one issue we had in Paragon is that generally speaking

the rule was you use your automatic attack to attack things and do a reasonable amount of damage but for a lot of characters who used your abilities if you wanted to do a lot of damage that was a good rule however when it came to towers your abilities didn't work on towers but people would try and use their abilities and they would waste their Ultimates and waste their 'high cooldown abilities and we had to even though no damage numbers came up even though there was this feedback that was not enough to stop them using this otherwise good rule so eventually we really just had to add huge big immune feedback whenever they did that like this does not work really getting their face to do it another way you can do it is offer help on failure you can have hints pop up so they're trying their dominant tactic and you should take that and you can gently steer them towards other tactics to use full strength you can add exceptions you can also understand and subvert expectations you can essentially see it players up to

really demonstrate to them hey cool you can play how you want here's a dominant strategy use it and then you can surprise them by mixing things up so Metal Gear Solid 5 had the adversary system I believe what was called where the enemies would actually start to adjust to your tactics so if you got a lot of headshots that start wearing helmets if you snuck around they would wear night-vision goggles these kind of things another reason why good rules can go bad is redundancy this is essentially if a piece of information when it's presented to you is not useful you won't integrate it as part of a rule system so this can often happen in games where maybe in a tutorial or in an early onboarding we teach something and then we're like okay we taught that but then the person doesn't have to use it until an hour later or until some significant time later and the teaching or the object was not introduced close enough to where it is useful example of this from fortnight is the Outlander

fragments these are objects you find around the map and what they do is once you have an outlander hero you can use them and it gives you special abilities however they are introduced with very little explanation way before you can ever have an outlander hero meaning that a lot of players just kind of learn to ignore them because they don't do anything and even once they get an outlander they continue to ignore them to prevent it you can give extra volume on rare information you can only show information when it's needed or remind them about it later so maybe you give this is the example of the Metro vania games right you give someone a new tool and then you immediately show how to use that new tool and then that'll trigger them to go oh wait I've seen those rocks with the cracks on them earlier on I thought it wasn't important but now I know it is important or maybe when they walk past those cracks late earlier they shimmer it or they had a had a message that said come back later to help prompt and remove this redundancy problem you can also approach it with unflattering

onboarding where things are generally taught over time and as close to when they're useful as possible it's generally best to not just teach something once and have the player use it once but to teach something and then have the player use it a couple of times in a useful beneficial way there's not really many cases you'd want to use redundancy so I'm going to skip over that and move on to encoding in action deficiencies which is the primary reason why bad rules exists this is a if then rule you've created based on bad information generally because the system or the game you're interacting with is not giving you the information you need it or is letting you progress via an incorrect tactic you can think of these as superstitions like don't open an umbrella inside there are a rule but there are an invisible rule we have an example of this in Paragon with due to a weight the way that we taught players how to level up their abilities and then how to use their abilities on controller a lot of Clair's believe that the button

press you had to do to level up where you had to hold down a trigger and use press the button that was the ability was also the way you use that ability and it did work you could hold the level up button and press the ability button to use your abilities but it was inelegant you didn't need to do that you could just press the ability button once you had it learned the weak points in this is when you're harvesting in fortnight sorry that's gonna yeah when you're harvesting for tonight these blue circles appear and then you hit them to harvest faster this was a massive problem for fortnight probably one of the longest-lasting kind of low-level UX issues these weak points went through massive number of iterations to get where they are today and people were still not picking them up because what people were doing was

learning a bad rule associated to them that hit an object a spot would show up on the object and that like okay this is just telling me I've successfully hit this object and this object is now marked so I know which one I'm harvesting from that was the bad rule that they learned and the only way we were able to eventually overcome this is rather than give this ability to people right at the start you play a little bit without it so it's not there and then it unlocks in the skill tree and then there's a little video and then suddenly we saw even though this has gone through massive things with numbers on it and bright lights and all this kind of stuff that was when people got it and of course if you look at people playing fort night battle royale where there is no unlocking of this you just do have it from the start a lot of people don't know that it exists however in that mode it's less important so again onboarding which was an example there you progressively unlocked and clutter your feedback really how stop these bad rules being formed and they can become good rules right away

there's not really a reason to use them so let's move on to knowledge level errors again knowledge level is this higher level of operation what we typically think of as thinking here is at this level are from referred to as cognitive biases here's just a logic part puzzle from the very original sin here they're gonna want you to think about it think it through and often a lot of game developers are people that like to think things through they'd like to optimize like to be rational and they feel that this is the the major way that players also encounter things and it can be the case but what is more often the case is players will look to there will be some players that optimize your system take advantage of your system then other players will learn off them it's not that all players will try and do this most of the time players are just trying to get by rather than to optimize across all the knowledge level issues and the previous ones but particularly impacting knowledge is work space limitations this is really where

the fact that we only have limited cognitive capacity comes into play working at the knowledge level is very inefficient it takes a lot of cognitive capacity but it is also what's fun like it's that that like intense concentration that really struggling with a game and thinking it through it has a Mitter impact on all these things so every example I'm gonna give it's like caused or could be helped by reducing load and I'm gonna try and not talk about that because it just would come up every time so the first of these errors is seal activity that's attention to psychologically salient things in the environment rather than to what might be logically important so a straightforward example of this would be the cake is way more psychologically salient it's got a motion attached to it you want to eat that cake but logically probably for your health it's better to eat the vegetables which probably don't have the same emotional impact they don't have the same psychological salient see attached to them now an example of how this can happen in

games is is the Luke cave and destiny this was not the optimal way to get Luke and destiny but look at all that glowing loot like that's such an emotional such a psychologically salient thing and players will do this they will do activities that are not optimal there are grindy that are otherwise not really particularly fun if the psychological salient see is high enough so to prevent it make the logical psychologically salient in fortnight the chests and the the safes and different things will you find objects and materials in the world they glow they have this lovely sound to them you can hear the sound through walls that attracts you to them and they have real-world analogues right a chest is where valuable things are the same thing for our weapons in fortnight Battle Royale you can look at a weapon it's got a nice glow on it it lets you know its rarity it looks kind of like

the real world weapon you can have the psychological impact of this gun is better than this gun from how it looks you can add tips gameplay clues you can highlight new information make if you can't make it really psychologically salient if you can't make it emotive and attractive then at least point it out through highlights or other ways you can use steel activity against players using traps and distractors a mimic is a really good example right players see it chests this is a very emotional thing for them they run over to the activator it's jumps up and it starts attacking them they didn't spend the time to think about it they didn't notice that it's in a funny situation or whatever it might be you can take advantage of the fact that these items will stand out you can also use them as this detractors distractors sorry I should say in hidden object games you put like nice glowing things and really pretty things amongst where the players actually need to click

and drag their attention away from where they need to be if they're supposed to be thinking it through out of sight out of mind is also sometimes called the availability heuristic which is just simply the things that come easiest to our minds when we're encountered by a countering an issue will be typically the things we think are most important and also the things that we tend to focus on whatever pops into your head first to someone with a hammer everything looks like a now and again this is similar to the rigidity problem of players will use the tools that you give them and unless it's very clear that it's not hammer time they should they'll stick with that so to prevent out of sight out of mind you can encourage solution switching by mixing up enemy types by making certain things immune and giving strong forcing functions to move them away from root or you can provide failure based hints again similar to the rigidity you can pop up oh the enemies enemies that are

red take more damage from fire item up from ice items or whatever it might be that you want players to think about that would not usually easily come to mind you can of course use out of sight out of mind by presenting a lot of now looking problems we do this in games all the time you give someone a big gun that's good at shooting a lot of people and then the very next room they go into has a lot of people in it this is using the availability heuristic by proving that it's true I have this thing I've just walked into a room with a lot of ammo a lot of crates I know there's probably a fight through that door you can also use it as kind of like a fun part of puzzle games portal would do this in some cases where you'd come into a room and there'd be all this complicated stuff in the room that looked really important really can like something that you should be thinking about yourself this puzzle but the actual solution was very very simple and this could be a amazing thing for

players to have worked it out this complicated way but then maybe see this really simple way to do it confirmation bias is so well known I almost lift it out but it's essentially that people put efforts towards what they expect the joke example here is that tilted towers and in for an open RL if you don't know is somewhere there where a lot of people like to land it was it's a high loot location a lot of players land there at the start of the game we introduced the new location called lucky landing and this was kind of a meme that come up as haste people wise still people landing it at tilted they could go to lucky landing where it's a little less hectic but there's still good loot there and that's because they know that like that tilted Towers is the place to go they already know that so they're just going to keep doing what they expect to work they can whatever confirms their expectations and when they do land there there is loot to prevent confirmation by us understand your players expectations make sure

they're met and strongly predict contradictory information if you if you want them to not have this bias which is a very natural one so another example from Paragon is that when Paragon first launched we had these really lovely towers it's it's a MOBA and they're these beautiful art crafted towers they look amazing with this huge bubble on top and when you're approaching them they target you and they food at you it looks really awesome but the thing you have to destroy to take out the towers is that little crystal down in the bottom left there so what we would see is players would actually try and attack this huge big thing and it does nothing it was what they expected it's where they put their effort even with contradictory information coming in of them getting shot of doing no damage they would go for these big towers so eventually over time we moved to this design where the tower was the crystal and this big tower was replaced by a turret that would open up and have a gun in it and track you around because this lined up expectations with what the

players were after all right another one is overconfidence now this is typically an error and I think this is one of the strongest examples of one that's not an error in games in most cases because overconfidence where we disregard signs that are saying hey you're doing the bad thing is actually what we want players to be doing nearly all the time in our games it's really natural to be overconfident in our games when they see this I have a gun I'll shoot you if you come in here we don't really expect our players to turn around we want them to keep on going when we see when they see these messages written and blood on the walls if you do really want to prevent this if you feel this situations they're people you want to knock someone down or you want to show that maybe their character is not very powerful there are ways to do it by put forcing functions in or having the player be defeated you do have to be careful about them this is a classic thing that many of the Bioware games have done where they give you a boss fight partway through the game and you

get beaten by that boss and there's nothing you can do about it there's no way to beat the boss and this is to try and reinforce that maybe you're not powerful enough at this point but if handled incorrectly it can feel like a real cheat to players so you do have to be careful with it because remember all the time your players think they can do it that's why they keep playing your game if you really if you really convince them they can't they're probably going to stop but you can use it every game every day game design relies on this things getting harder throwing challenge spikes at players they'll keep on going because generally speaking we're overconfident in Dark Souls this this huge boss you meet early on but players are going to attack the boss because well why would it else would it be there like we can take it and then they get killed which introduces the death mechanic in the game you do have to be careful around this to make it clear what's happening make sure that it's fair but it is just part of everyday game design bias reviewing is simply you're thinking

through what you need to do and you go I've thought of everything that I can think of so yep cool but you can't think of things that you can't think of or that you don't know like unexpected information so of course to prevent this very simple it's very well used in a lot of games you provide actually check off lists that have nothing left out that the player can carry with them they then don't do this bias reviewing because there's no bias they have the review list they go through it you can use bias reviewing against players by not providing progression logs if you do think it's important for your game that part of the challenge of your game is to actually think of all the things and keep them in memory and be able to do that or you can provide progression but leave information out or change things up and the shadow of mordor games you build these examples of orcs you find out what their weaknesses are and then when you go to fight them maybe in other auction up that you don't know about or maybe partway through the fight the orc its weakness changes of their tactics mix up so this is using the fact that

the player feels prepared and then throwing a bit of a loop Helu effects that are that good attributes in one area rub off on to others we tend to think that people who are attractive are also smart and kind people who are less attractive are dumb and mean this can also apply to games you can use this in your game if your game looks nice and polished people we actually be a little bit more forgiving to it than if it looks rough and and unpolished because of this halo effect we can prevent this by submerging and challenging expectations halo effects exists because of the society around us so for a long time books were bad guys then orcs started to be shown as good this was server ting the halo effect and probably nowadays the good orc has actually unexpected things so this may not apply over time but this is one example of kind of going against halo effects and of course we use them all the time by making good characters look good make evil characters look evil

we always want to punch Nazis the last two things here are related to each other one is problems with causality this is essentially that players tend to underestimate future array regularities which means they think tend to think how things are going now is how things will go in the future they have the representative heuristic which is the idea that small a big outcome should come from big causes and small outcome should come from small causes so they should expect if they if something blows up there's something big that causes it to blow up but if something small and subtle happens they expect something small and subtle to have led to that illusion of control is essentially that if you give players control they'll feel like they can have more control over random outcomes and they actually have correlation as causation people as not causation people tend to associate

things that happen at the same time together and hindsight bias is simply that once something has happened and you think back on it it becomes kind of very obvious that it would have happened and it feels like you should have known beforehand it can be a bit very big problem in UX and for people giving talks about these kind of things because they seem very common sense after you've heard them prevent problems with causality this is why we put so many feedback systems and games this is why we put progression systems power growth mapping systems we match players expectations we help overcome these cognitive biases by providing systems in the game that accurately let them see how their progression will go that things are going to change in the future and we of course use this all the time by just giving players control and feedback by doing that we involve them in our games it makes them feel like they have more power than they actually have when we really make them their actions matter

through strong feedback so match three games are a really good example of this you have very little control over what's going to fall in when you make a match but because you're the one making the match and the game makes so much big deal about it with all these popping things it feels like you're really smart and in control there last one is problems with complexity this is essentially we're really bad at delayed feedback so if we press a button and nothing blows up or nothing obviously happens we're gonna have problems working out what happened things need to be close to each other for us to understand them in most cases we tend to think in causal series we like to think that this causes this which causes this which causes this which we know isn't true in game design right we we make systems that have all these emergent properties where systems all interact with each other and are necessarily causal but your player is likely to think that kind of very causal and we have difficulties with Exponential's for

much better with linear growth than we are with exponential growth which is partly why expert exponential progression can feel very and rewarding to players and it has been a recent move towards more linear progression systems as they're more understandable and more in line with player expectations so to prevent it feedback visualization and clarity really letting the player know what's really happening in the game same with the the similar thing at the last issued here's an example from Ubisoft scheme for honor where they're just showing the monetization a lot better than a lot of games do where they actually just put a simple bar graph in to show hey when you mouse over things it shows you how much base you'll get and how much bonus you'll get this is so much better than making the player do maths by adding things up by like 20 percent off and plus 100s and whatever this actually clearly shows the benefit of going at each stage all right so to finish off I'm gonna talk about a systems view which is kind of a holistic

way of dealing with error the systems view it it's hat is a safety culture of you versus what would be a more traditional blame culture and it's a recognition of the fact that errors are caused by system interactions as I was talking about errors arise out of a lot of complicated things and this this error that we make naturally to thinking cause also systems also applies when we think about error we like to think that errors happen because someone did this and then this happened and then that happen and then that happens but that's not how errors happen they tend to come out about out of much complicated systems and so the idea is rather than look to one person or look to one cause and then blame that and stop you look to the system you look to what caused that to occur blame stagnates things it stops you at that point whereas a systems view a safety view lets you learn that lets you improve and innovate where cheeses

involved is a common model used in this is called the latent failure model or the Swiss cheese model and while this is still kind of a linear line what it suggests is within organizations and also within games there's all these different systems interacting and most of the time they're solid that good systems but within each system there'll be holes and then when certain systems all line up with each other an error will occur when all those holes line up it wasn't just one piece of cheese all four pieces of cheese had to be in a line now a SystemVue on players is kind of the heart of usability in the heart of UX when we write UX reports we don't write the player was confused because that puts the blame on the player we write what the issue was so for example this color the color blue was not readable on the background then we wrote what the outcome was the player couldn't read it and that's how we present a UX issue it's not putting the blame on the player it's looking to what

the systemic reason is it's not to say get good or players are just done we look at the system to see what is causing it look deeper so it won't happen again in the future and one thing I do want to end on is we can also apply all this in workplaces that's where human error is applied generally speaking outside of games and within our own workplaces we should also be thinking in a systematic way and not looking to blame it can often be the case with our players right if something goes wrong in a game they may find just one dev and just pile on them and blame that one day of who they say was responsible for it but it happens within our workplaces too we sometimes have a tendency to just look at somebody and blame them for a mistake and that's not to say that people aren't responsible but within our own workplaces the errors that occur are due to systemic problems crunch as an example crunch happens because planning wasn't in the right place or whatever it may be lots of factors deliver into it not just one

individual and with my own team I'd lead the UX team when something goes wrong when somebody makes a mistake and it error occurs I talk to them about it and I talk to the wider team and I try and find out what were the circumstances around that what were the demands placed by different things what were the systems going on because if I can find that out rather than just shouting at one person I can make sure that that does not happen again and I'd never have to deal with it whereas if I just blame that one person they may not let me know next time something goes wrong they may leave the company they may not work better they may just become resentful and these errors may happen over and over again because they haven't changed what actually the problem was which is systemic so summary all humans make errors to to err is human errors are what makes human behavior interesting and predictable and fun we use errors all the time in game design and we want

those intentional errors the errors that are part of challenge that part of the fun of games to occur what we don't want is the unintended ones that make players frustrated and leave our games and to do that we can embrace fun error prevent unintended error and look to systems both within our games and within our workplaces to really improve and not just focus on individuals thank you very much [Applause] so again please remember the evaluation and I think I've got ten minutes for questions no questions okay [Music]

okay so the question was I'm quite catching the developer but is a developer that pathologic the void okay they have NPCs lie to you as part of a tutorial presumably because the NPC is unreliable is that what they're trying to establish they try and establish characterization okay right I mean fundamentally what you're trying to do there is you are trying to give players false information to form bad

rules and that's fine as long as the feedback is clear what's happening and what's happened once once they do follow that like I assume the NPC laughs at them or informs them that they've been tricked or do they never find out I'm not sure I haven't played these games but that again if that's the intent to make these NPCs unreliable then that's perfectly fine what you want it's okay for people to make errors and be tricked by your game it's fun you just don't want that to happen and them to feel like they don't know why it just happened there's a there's a mic there if you could use it sorry I portal to I think does a version of this where in the beginning when you wake up Wheatley is trying to get you to I do I forget what it is like talk maybe and he keeps you get the instruction on screen like press a to talk and when you do it you jump and then he's like no no what

you're doing is jumping and then again it says press a to talk or whatever mm-hm and he keeps saying like a you know no that's a jump and then eventually he's like oh let's move on but basically it's just taught you is jump and you're gonna use that from then on and it's hilarious so yeah that might be similar yeah and and also that's a really good example of of psychological salience II we tend to remember things that are funny or make us laugh and with that clear feedback loop of you press the button and you jumped and then there was him saying that you jumped and and and the repetition of it that helps it be something that you learn and and something you're not going to make a mistake about in the future any other questions okay doesn't look like it if anybody does want to talk to me I'll be up front here a little bit and then I'll go around to the area [Applause]