Useful Computer Games

Speaker A: 00:00:00

Hello and welcome to the Cognitive Engineering Podcast produced by Tell Me Studios for Aleph Insights. In this series of podcasts we take a look at interesting topics and discuss what we think they tell us about analysis and decision making. I'm Fraser McGruer and I'm here with Peter Coghill, Chris Wragg and Nick Hare of Aleph Insights and this week we're talking about computer games and their real life utility. Unusually I'd like to start off because I'd like to talk about a game that I used to play and I went through a period of playing it quite intensely and this is when I lived in Qatar about 15 years ago and I managed to access a free copy of The Sims. Now I had a lot of free time on my hands in Qatar and so I played it quite a lot and for anyone who's not familiar with The Sims and I'm sure you all are though, is essentially you have your own real life character, you build their life and you've got to look after them, make sure they're all right and that they progress in life, get a nicer house, better job etc. But it got to the point where I mean I was very invested in the character that I'd made and certain humorous things happen if you don't sort of send them off to the loo you know they'll wee themselves and if you don't feed them they'll starve, if they don't sleep they'll sort of turn into you know they'll go mad essentially and after a while as much fun as it was after a while I thought you know what I've got enough trouble enough work in my own life making sure that I don't wee myself and that I get enough sleep and that I feed myself without having to do this with a bloody computer simulation and so I just thought you know what this is just too much like real life and I've had it this is not fun anymore okay and so I pretty much immediately stopped playing that game. So these are my thoughts, what is your response to what I've just said? Well and then you had children presumably and started all over again. That's exactly what happened so yeah. Maybe you could have less

Speaker B: 00:01:52

free time in Qatar if you hadn't been playing this game so much, just a thought. You could have had

Speaker A: 00:01:57

friends in a life. No but you can't have friends in a life in Qatar, that's the problem. All right

Speaker C: 00:02:02

they're probably not as good as the characters in The Sims. Your characters were probably more interesting than other expats in Qatar. Quite possibly. I mean the thing is about there's an interesting question about why we play games which are close to simulations like The Sims. I mean obviously the first big sim game was SimCity. Great game. It's very unlike, it's very groundbreaking because you know you weren't really a character. Most computer games had made you a character and you were a person and in SimCity you were a very abstract kind of personality. You were if anything a sort of town planner saying well I want this area to be residential and I want a road here and you know you weren't a person in a game doing stuff. And actually you know so there's a question, it's an interesting question why we are entertained by that kind of thing. Why we like seeing simulations unfold and of course there's now these days games of massively more complexity. But of course then at the same time there's a limit to what we as a player are able to do in games. So you know there's this interesting kind of design, this design tension between giving the player a manageable number of things to do and at the same time having a vastly increased simulation with lots and lots of detail in it. And you know obviously people vary in their preferences but this touches on something else I think we're interested in which is you know what can you use simulations for in real life? I mean how are they, how is it useful to put someone in the position of a potential decision maker in a simulation? What's the utility of games and particularly of simulations?

Speaker A: 00:03:45

But is that the question we want to discuss? Which is how useful simulations are? Or do we also want to talk about if we especially look at it from a gaming point of view is how much fun is it if it's too much like real life? Maybe we can discuss both. But those are the questions you want to

Speaker D: 00:04:04

look at right? Yes? No? Okay. Who wants to jump in? Chris? Peter? Yeah well I think the sort of fun and realism kind of dilemma or you know how you balance fun against realism because obviously as you found from your experience something that effectively replicated your own life, not that your life isn't great fun but as a game that wasn't terribly entertaining. So if you can sort of raise up the level of abstraction and make yourself more omnipotent and omniscient and be able to see more of the world and become godlike effectively then that perhaps is more fun. Who doesn't want to be a you know an all-powerful deity sort of toying with people? But there you've got this balance between being able to see things on a detailed sort of scale versus being able to see the big picture and you know move mountains and sink whole civilizations and somewhere there's a sort of sweet spot where you get enough of the enough of the realism and enough engagement with what's going on and also enough power and enough breadth of vision. Well there's two separate things here

Speaker C: 00:05:36

which we need to consider separately I think or it's helpful to consider separately. One is the sort of decision environment that you're put in in a game. So what are the you know effectively how many decisions do you have? How much randomness is there? You know how much is it to do with your sort of physical skills sometimes called sort of twitch ability you know where you're sort of you know running and jumping over things? How much is it to do with sort of slow deliberative decision making and so on? All of which is completely irrelevant to the theme which is something quite different. So you know you can have you could have a very you know a very detailed complicated game about making a cup of tea or you could you know you could have a very abstract game where you are ruling an entire planet. It doesn't the theme that who you are or the decisions you have to make are very different to the structure the kind of decision structure that you'll put into. And this is I mean this is a very old sort of issue which has been dealt with in board game design before computer game design was a thing. So you know and it's a topic of quite a lot of discussion about what the right level is you know to make a game entertaining. At the absolute extreme you get these monster games. The classic example is Campaign for North Africa right? The play time in Campaign for North Africa is 60,000 minutes which works out about 40 days non-stop playing. And it's an incredibly detailed simulation of the North African campaign during World War II which seeks to track every item of ammunition, every piece of food, you know tracks the condition of every vehicle. It has rules that one of the things it's famous for having a rule where you know Italians require more water per head because they have pasta and you've got to take into account you know the water you need for cooking pasta. It's ridiculous. And I venture to say even though I've never played it that you could play an entire game of North Africa not ever actually have had to think about what you're actually doing. You know it's all process. You're all just managing the simulation. And in contrast a game like Memoir 44 which has it's a war game 20 page rule book and a friend of mine who used to design military simulations claims it's he thinks it's the most realistic war game one of the most realistic war games he's played. It's very very abstract you know it's very much sort of you know I'm going to move two squares with this unit roll a few dice because it simulates the decision environment of a military commander who doesn't need to know where every item of ammunition is. He can assume that someone is doing that. You know what he needs to know is just broadly where his units you know where is his command focus what are his chances of winning if he orders a particular assault. So that tension you know between sort of accuracy of simulation and playability is something which predates computer games.

Speaker D: 00:08:33

But presumably the in you know where we have now got to with the technologies is whereas before you might have wanted to have that level of detail from the first game you talked about in order to make the environment in which the simulation is taking place kind of representative. So you might you might have wanted somehow the amount of pasta the Italians were eating to factor into the outcomes related to the game. And the only way of doing that in a board game was for somebody to actually work it out and go through the process. Now we have technologies where that process side of the game can be taken care of and you know the results are built up over with that level of detail and that level of realism. But you don't a human doesn't need to engage with working out how much pasta everybody

Speaker C: 00:09:26

needs. Well yes so this but this comes down to what you need to simulate and what you can abstract. And I mean you don't need to have rules for the Italians using more water. You can abstract that completely by by you know having rules that say that the Italians have you know you might have an abstract logistics supply cost and you just say the Italians have to have an extra cube each each week. You don't need to know that that's because you know they are using more pasta and to have an actual rule about it. Or a piece of fusilli. Yeah exactly yeah.

Speaker A: 00:09:59

Peter we've not heard from you yet. What are your thoughts on all this? Well a good example but

Speaker B: 00:10:04

even within a genre of game they can get very different kind of results. So I'm a bit of a first person shooter junkie. Always have been. I know it's a dirty secret. But I've recently been playing a couple of games. One is Arma 3. It's been out a few years now. I like the sound of that. It seeks to be hyper realistic. It tries to it gives you very you know there's extremely good physics. There's extremely good lighting effects. The sort of weather effects. The AI the enemies have is pretty good and the way they react and the way they sort of try to outflank you and move around and stuff. Very good. But it's really difficult. When you play on standards standard difficulty settings it's really hard. To the point where it's not fun. In fact I've been playing it with reduced settings because I just can't be bothered because I want to make you progress. Contrast that to things like Tom Clancy's Ghost Recon. Cool sounding name. It's very similar on the face of it. It's a first person shooter but it's more entertaining. It's more like junk food. You have to be less invested in it and it kind of suits the interface that's available because Arma would be great if you were fully suited up in a haptic feedback VR suit with headset and everything. Sounds like you might as well just join the army. You could engage the world in the same way as you might in the army but being restricted to a keyboard and mouse and a sort of 24-inch monitor in front of you it's really too difficult for

Speaker D: 00:11:39

that. But I dare say I mean it's sort of it kind of may be the more realistic simulation because I dare say actually trying to you know be involved in sort of close quarters combat is probably really difficult. So you know your actual ability to kill 20-25 people every two minutes or something you know if we could do that then we'd have the British Empire again.

Speaker C: 00:12:10

I mean the problem with you know simulating lots of detail is it doesn't it's not more realistic. We need to get away from this idea that having detail is the same as realism as far as the decision maker is concerned and the I mean the really good example is a game I think it's called QASD or something and it's an online game where you control a runner but you basically control his thighs and his calves and you see you have one button for it to make his thighs go up and down and another one to make his calves kick forwards and backwards and the aim is to try and run and if you can get more than about a meter without falling over you're doing really well. It's incredibly hard right but it is a very detailed simulation. It's limited in what it offers but the point is that it's a very detailed simulation right of you know moving someone's legs. But it's actually more realistic for me if I just point the thing to the right and the man

Speaker A: 00:13:11

runs. I want to talk about fun I don't know if we want to talk about fun but I want to talk about fun. We love fun. Good so I want to talk about Grand Theft Auto yeah and I think what's fun about that is on one level it has a certain amount of realism in it and it's based on a kind of a real world that we sort of recognize. There's a car there's music you're driving around certainly the original driving around downtown Miami but what's great about it is who amongst us hasn't dreamt of just one day just driving off onto the pavement or sidewalk and just mowing over some people and getting your gun or even just a hammer and go walking up to a policeman and smacking on the head and seeing how and then you know going and shooting a stripper dead or something. I mean it's just it's that sort of it's the possibilities. I don't dream about

Speaker C: 00:13:58

doing that very often. You keep it light Fraser. The issue here is this sort of reminds me of this of these things phenomena in which we have come to get used to kind of tropes in in games and tropes in films that are there because we expect them to be like that not because they're actually realistic. So the classic example is the horse's hooves being with coconuts which a horse's hooves don't sound like coconuts at all but now if you see a horse you expect the coconut sound and if you don't hear coconuts it doesn't feel like the horse is walking properly. Likewise with I mean the whole art of Foley is putting noises in where you expect them even if they're not there in real life and I think computer games are full of this kind of thing where if Grand Theft also was realistic it would be really boring you know because you die immediately as soon as you got shot instead of having a health bar you know. You could say that for a lot of yeah yeah sure so games always have things like you know you get these health packs where you're instantly restored and stuff and and it's because it's boring if you don't have that you know it's actually bore realism is boring in a game um so you know what it's I think this idea that the aim is to pursue more realism doesn't really stand up to how we actually experience fun. Sure um actually there's a there's

Speaker A: 00:15:19

a game called I think it's called Checkpoint or Passport or something like that which you can play on your ipad or iphone and um where you're a cold war era yeah papers please and you're a you're a border guard um in eastern europe and you have to check people's papers and they're sort of different documents coming through and you need to work out which ones are false and you've got

Speaker C: 00:15:40

to decide whether to accept a bribe because you're not getting it is this outsourced does

Speaker B: 00:15:44

this crowdsource work by the uk border forces they found a way of and the thing is I played it

Speaker A: 00:15:50

it was got great reviews I played it I thought it was a load of rubbish it was just so boring I again I don't want to sit there sort of checking papers that's no fun um anyway look we need to wrap up um so what do we what do you want to finish off on how do you how do you want to wrap all this

Speaker D: 00:16:06

up well I suppose I suppose the um the the you know one of the um applications of this this technology now which is uh even more relevant to you know analytical decision making is that um these games have these particularly these open world games where you've got vast kind of uh complex simulations of of real worlds maybe cities or you know environments um are now increasingly being applied to real world decision making uh and um trying to determine you know for example what happens if there's a particular type of disaster in a city what will happen in that that world and building up uh like agent-based models where every person is an individual and acts in a particular way in in a in a in a riot or you know some other civil civil disturbance so the utilization of this technology for to support real world decision making uh is something that's expanding and where this this question of how much detail should you put into that that world because you know now the the the amount of processing power that can be applied to that means you can detail you can provide detail on perhaps every individual within a city and their flat and the sewage system and you know how traffic works in that city but does that necessarily provide a uh more accurate or more useful simulation of of the world and that's and

Speaker B: 00:17:36

I've seen lots of concrete examples where a more simple model is actually generally much more useful not only because it's probably easier and quicker to run uh but also it just gives you a better decision so dsdl does lots of modeling for to support decision making and one of the one of the key areas of modeling and simulation comes in is deciding whether or not to buy whether when when they're buying a new tank say whether to buy tank a from ba systems or tank b from Lockheed Martin um and the the one of the tests that we'll go through is a sort of modeling and a simulation test to see given all other factors kept set kept the same how would a versus b perform um and lots of different models that they use for doing this sort of thing and generally the the best models are more the more abstract ones where large formations are modeled as entities and there's a stochastic relationship between the an action and a given event rather than trying to model everything down to a single bullet with every

Speaker C: 00:18:40

single entity modeled individually quite quite apart from the fact that that's it obscures the audit trail so it's harder to unpick what assumptions are driving outcomes when you do that um it uh it it also means that you know small errors in your assumptions about those very detailed things can can can be magnified in in when you when you simulate the whole thing i've got a good real life example of um from uh again the same friend i mentioned earlier uh who used to design military simulations uh who apparently was at one point trying to simulate what impact uh it would have if we if our early warning aircraft had a had a wider detection radius so if we effectively made it more effective uh and uh the in the simulation uh this aircraft when when they sort of simulated having this wider detection radius so they could detect threats earlier uh it it turned out that everything performed more poorly and um it took some time uh you know in the face by the way of of sort of various uh senior military people saying oh well this just goes to show what a load of rubbish computer simulations are uh to actually identify which bit was was driving that and it turned out it's because there's this one of the doctrines one of the one of the items of doctrine says that when you detect a threat if you're an early warning aircraft you should land immediately now that's fine if you have a short detection radius it means you're you're immediately under threat as soon as as soon as an enemy uh fighter appears but of course if if you don't change that bit of doctrine uh actually does make you less effective because it means that the the more threats you can detect the longer you'll be on the ground and and i mean that's in a sense not realistic because actually in real life you know that the doctrine would have changed and it took some time to identify that um because it was quite a detailed simulation that's the kind of problem you wouldn't necessarily have if you did if you dealt with these things at

Speaker A: 00:20:34

a higher level okay um well look on that note we're gonna have to wrap up um it's time for us to jump into our convertible corvette and uh cruise around north london knocking over police and small children and and going on a shooting spree right yeah average average monday afternoon so thank you very much for listening to the cognitive engineering podcast i'm fraser mcgruer we've been here with chris ragnick hare and peter coghill of aleph insights and until next time thanks and goodbye bye