Controlling Carnage: Effective Side-Scroller Controls on Mobile

[Music] hello how's everyone doing it's Friday last day GDC I know so say goodbye to everyone today say goodbye to your friends we'll see each other again next year uh thanks for coming to this super nerdy talk about mobile control schemes uh you're all super nerds for being here and I love you all so thank you for coming uh I'm Kayla I'm a game director at Roadhouse interactive uh we are a studio up in Vancouver Canada we make mobile games uh we were founded in 2009 and we have just over a hundred really fancy talented individuals working on whole bunch of games uh one of those games and the one we're going to talk about today is called Warhammer 40K Carnage uh it is a action RPG sidescrolling platformer try to say that 10 times fast uh and it was released on mobile and tablet devices around May last year uh we worked with our good friends at Games Workshop on this game uh and it was pretty cool like like the title must uh

have said it's a platforming game and so this talk is about the controls for that uh really back in the day it used to be so easy we had these controllers that we could develop for and they were a known good uh as developers all we could really do was figure out what buttons would do uh when you press them in the game we focused on the gameplay actions uh and just map them to different buttons on the controller uh you know we don't have to think about what that controller layout was like uh or anything like that we had to work within that constraint uh and thankfully the developers were were thinking about us and uh with different console Generations had more and more Innovation uh more buttons more Innovation I actually took the screenshot of the Jaguar that's downstairs in the North Hall it's really fantastic um so you know Le was good more buttons we just kind of mapped stuff to that and then this happened uh we had a you know a wonderful Apple demonstration showing the new future of uh devices with no

buttons whatsoever so now we're in in a world where the most popular uh gaming devices have no buttons uh there's no pre-made controller Hardware uh and we now have to worry about things like grip input schemes and how all this impacts uh the visual representation of the game that we're making so I'm going to try to tell a story in this in this talk um our controller scheme really came out of the evolution of the game um so we'll kind of break this into three parts kind of a pre-pro stage uh the different stuff we did in the game to iterate on what the game was and how that affected the controls uh then we're going to focus on the game uh during the standard development cycle and all the different iterations we did during that period and then our kind of fining Mo mode on the game where you know theoretically we would have just been uh uh you know polishing everything but like all games we had some fun new new changes that

happened at that point in time so let's have this great story so let's start at the beginning back in 2012 uh Auto Runners were all the rage um and so uh temp temple run was get garnering some really incredible download numbers and our pitch was hey let's take uh what these Auto Runners are doing and let's put a 40K paint job on them uh was we were super inspired by games like shellrazer uh and the girl on fire um and we were really taking taking a look at what the Auto Runner scene was doing and thinking let's add some controls uh let's put a 40K franchise on there and should be great so it was going to be a auto running Lane based sides scroller uh with elements of ranged and melee combat and so the controller scheme that we really thought about for that uh was this uh we would hold the the device with both both hands uh a thumb down on the left would do a melee attack and a thumb down on the right would do a ranged attack and you could swipe up with either thumb uh to do a lane change

uh between different uh platforming segments um so yeah it was Auto Runner so we have to worry about uh controls for starting and stopping running uh all the weapons would Auto Target enemies we didn't have to worry about any of that um so yeah this was good it was a nice simple control scheme uh we built a prototype and iterated on it for a few months and and all was great but really we weren't feeling it so uh we took some time away from the project uh and when we when we returned we really looked critically add what we had uh we realized we didn't really like it uh we didn't like the movement we didn't like the combat it just wasn't feeling fun for us so we looked into our hearts and we said you know what we don't really want to make an auto Runner we want to make some classic platformer uh platformer game and so we went through a massive reboot process and these are some of the the games that we were inspired by at that time so our next big iteration of our control scheme was uh what I considered our our gesture controls and so uh at at this point our our control scheme was

being driven by a massive increase in uh what kind of new features we're going to need to to to have a platformer control so we now have a discret run button so the player can actually hold down uh their left thumb in order to run forward or release it to stop uh there is no Left Right Movement at all it's just run or not run um we have swipe up to jump uh we have swipe down to aerial attack if you're in the middle of the air you have swipe back to parryan attack swipe forward to melee attack uh and you can touch uh with your right thumb to range attack or hold and release to do a special attack like a Mega Man's kind of a supercharge thing so this was what we thought would be an awesome control scheme uh for this game and I actually still believe this we could have gotten some pretty good mileage uh off of this control scheme it was pretty complicated and pretty convoluted uh but I think there might be some Merit here of of what could be an interesting control scheme uh for platforms but really we still weren't feeling it and um although I would have

loved to have iterated on this control for several more weeks or or months or or what what have you uh we had a public demo in about two weeks and so we we just couldn't afford to spend any more time on a control scheme that had too much too much risk so we decided to scrap it all uh and Scramble for something that would be super easy super lowrisk and something that we could just kind of deliver for that public demo in two weeks so we went to this uh super simple here uh we removed a lot of the the different moves and we just focused on four simple uh moves and that was hold down for run um press to jump press to range attack press a melee attack uh we did have uh the modal support for both the ranged and melee attacks right there so that you could hold and release for super attacks although I don't believe we actually had those moves for the Prototype that it was planned at least at that point and so it was a really simple control scheme um Although our assumption was that players would initially gravitate towards gesture controls kind of a default uh when they

picked up the game what they would try to do we felt that the controls for these buttons were pretty straightforward and it wouldn't take them very long to to figure it out uh and that proved true so when we showed this uh control scheme at the public demo at games day uh players would pick up the the iPad uh and try to do some gesture controls to move the character around and a second that they found that that didn't work they would immediately find the buttons and it was very quick for them to to figure that out um it was super effective they love the game uh it had really good response from from the players there uh and it was great so basically the positive feedback on this control scheme uh cemented this direction for us going forward and really killed any lingering priority for the gesture controls uh that we had previously all right so that's kind of our pre-pro area we're now in in proper development right now uh we have a proven prototype people like it we know we're building uh let's get into active development and let's kind of refine those controls event so so one of the first things we really started thinking

about was thumb sizes and we were super scientific about this I took an iPad around and I put the paper app on it and I told people draw me circles with your thumbs by holding the edges um and through that exercise I was able to get a whole bunch of markings on on the iPad that showed a huge Variety in sizes of thumbs um it's actually quite amazing these are all adult thumb sizes but you can see there's actually quite a huge difference in reach between a small thumb and large thumb and so when you're targeting on iPad the actual reachable area that you have for controls on the edges is a lot smaller than than what you might think it is um yeah and so basically if you want to create a button control scheme on the edges right there you have a very limited amount of space in order to allow people with small thumbs to to not go through a lot of strain to get to get to those so we also Ed this uh this thumb placement to figure out you know where would our camera be in the game and this was an interesting fight between uh input uh

space so basically having uh a good amount of space dedicated for input uh but also having a good amount of screen composition for the game this was a constant battle uh you know it would been nice to have the action a little bit higher on the screen but that just kind of leads to a really bad uh visual composition and so we kind of um set up with this kind of a structure right here where this line uh denotes uh below that would be where buttons are and above that would be where the main action action exists uh button placements again would be below that line um and uh we would kind of hold in in that lower position right there um the other thing with this is that the uh jump button is the highest button on the right uh because generally if you are in an action segment your thumb is is going to be lower which will allow for more visibility of enemies coming on screen um and so that allows you to do Mele and ranged combat against enemies um when when you see them right there and if you're jumping it's less likely to be with an enemy coming coming off of the

the edge of the screen the other thing about buttons on iPad layouts is you can't just put buttons wherever feels the most comfortable for the player you actually also have to figure out how do I put the buttons on a screen so that they actually are appealing and a visually uh good way uh because again these are going to be viewable things on your screen you have to take into consideration the visual composition of the button layout um when when you're putting them down all right so we have buttons in the game but buttons are a problem buttons are a big big problem and those big problems are uh what I like to call misclicks and thumb drift and so if you ever play u a game on mobile and you're getting really into the game at some point and you're trying to press buttons your thumb is just going to miss a button uh and that happens because you're getting into the Zone uh you don't really know where that button is because you're totally entering a flow State uh and what two one of two things is going to happen either you were going to attempt to press a button and something else will

fire so you think you're hitting shoot but you jump or you think you're jumping over that pit but you shoot by mistake and you die in the pit um or you're just going to try to press a button and you realize that your thumb isn't actually on any active hit Area at that point uh this is hugely problematic because the second that this happens the player is immediately brought out of their Flow State and is now fighting with the controls of the game uh this is awful it it sucks um we found that when people were playing our initial game uh they would be hitting the Run button and very slowly their thumbs would drag into wherever's most comfortable um it was for them if you remember I don't know if anybody grew up playing um regular Nintendo but when I first played it I would always be jumping with Mario like that and it's the thumb drift on on iPads is like the the moving the controller around kind kind of thing for me um growing up with NES controllers so anyway these are these are two big problems uh and we need to kind of figure out how to fix that again uh real controllers don't have this problem uh they have these tactile buttons uh your

thumb is touching these buttons at all given times and it's giving your brain all this uh backend uh bandwidth of all this information that you need to know in order to ensure that you always hitting the button correctly so if your thumb does drift off of a button there's there's feedback there to tell your thumb to get back into place and reorient itself and all this happens without you having to consciously focus on it um so so really if you just take this control scheme and put it on mobile it's not going to work because this control scheme the sizes of these buttons the layout of these buttons are entirely dependent on having that tactile feedback and allowing for your thumb to know where they are at any given time so how do we solve for the lack of tactile feedback uh on a mobile uh control scheme and so the hypothesis that we came up with was maybe we can use Muscle feedback from vastly different thumb positions uh to maybe offset the loss of tactile feedback so I want everybody to put their hands out pretend that you're holding an iPad everybody come on it's Friday Friday morning and I want you to put

into three different thumb positions vertically extended horizontally extended and tucked and if you switch between those three the thumb positions that you have right there and the kind of feedback you get from your thumb are very very different and so you always know where your thumb is if you think that you have buttons down in those positions and so what we did was we created a control scheme uh that kind of utilized that and so on the right hand uh thumb you can see uh jump is on your vertical extended U melee is on your horizontal extended and your range attack is on your tucked position uh and what we did was we changed the angles of of all those different controls on there to account for if your angle is a little bit off on each of those those controls it's going to account for that and it'll be good um jump has the largest hit area on the right hand side because on a platformer game if you miss jump it is going to knock you out of uh the Flow State a lot faster than if you just maybe miss one uh uh shot um we also found that uh the vertical extended

position is the most relaxed position for players and so it's much easier for them to kind of get into that that position as opposed to the horizontal extended which takes a little bit more muscle control uh to get down to uh tucked is the most uncomfortable of positions and so we try to uh um you know use that one more sparingly so uh on the left hand side uh we maintained run as uh as just one big uh hit Area so essentially this this has a very simplistic control scheme so that as a player you really only need to have thumb dexterity on one hand and on your other hand you can just focus on either my thumb is on the screen or it's not on the screen and I'm either going to be moving forward or not moving forward uh the hit areas here extend out further than what the largest thumb area that we measured was so that there's no chance of you putting your thumb down on the screen without it doing something so you will never get into a situation where your thumb's on the screen and nothing is happening there will always be something happening um and we have the

buttons down on the screen the buttons don't really act like buttons they act more like um focus positions that tell your brain where to try to put your thumbs it's kind of an orientation thing so that when you look at the screen you know what you're trying to hit but once you're actually in that flow State they don't really matter as much anymore and in fact we fade them down in Alpha as the player is playing so they kind of get out of your way uh but they're mostly used to kind of reference where your thumbs are supposed to go when you're actually hitting the screen you're probably not hitting anywhere near those buttons you're hitting all around them uh and these large hit pads uh really account for that this became super effective uh what we ended up having was if we had a bug in the game where these hit pads were disabled and only the buttons were working people would be freaking out and coming to my desk being like Oh my God our game is completely broken I don't know why it just doesn't feel good anymore uh but then the second that we re-enabled these hit pads it worked really really well and we found that uh when players were playing this they were always able to hit the buttons that they felt like they were hitting and they just felt super

effective uh at all times so this was great uh but of course we aren't just shipping on iPad we also have to ship uh on smaller devices and so what we ended up doing was splitting our devices into two categories uh devices that are less than 7even in in size uh and devices that are larger and so the the devices that are seven inch and larger uh would be our previous control scheme uh and these smaller devices uh using this smaller control scheme and essentially uh the way that I thought about this was if I can have both of my thumbs touch each other on a device I'm going to have to use a different control scheme than if they can't touch each other and so we had to make some adjustments here because if we accounted for the same hit Area uh the actual thumb and buttons would be all over the screen and you wouldn't be able to see anything and so we had a kind of more squished uh layout that we we came up for the iPhone uh again melee is on the horizontally extended button range is still mostly tucked and jump jump has a little bit more of an angled um uh vertically

extended type thing it was still super effective it wasn't as um as effective as an iPad control there was still a little bit more chance of you misclicking between jump and and melee uh but it was pretty good uh and it actually feels pretty solid on iPhone so we Al also can see the kind of line uh where we try to maintain buttons below it and action above it and this is what it looks like um in gameplay obviously with the the buttons super bright right there for this demo the buttons again Fade Out uh so when you're actually in gameplay you don't really notice them as much uh we keep gameplay above those buttons so that was pretty cool that was working well uh then we were getting into uh our fining so we had a solid control scheme in place and now we're ready for a nice calm fining period we're good we're comfortable we're happy it's going to go fantastic except it never does that in games um so uh we were at Alpha and people we're like you know what our gameplay just isn't robust

enough we're just we're we're going to need more controls uh we're feeling like the game is just Mindless hack and slash it's lacking in depth it doesn't have what we need in order to to really have player engagement over a long period of time and really this complexity of controls was a really hotly debated topic throughout the entire project if you remember at the very beginning we we were really targeting this Auto Runner casual audience with this game and as we were moving through development we were getting more and more complicated control scheme by adding buttons and more moves and all the kind of crazy stuff and so we really we were trying to figure out you know who is our target audience what are we trying to get to what kind of a game are we making uh and the controls were were feedbacking in into uh what that target audience would be so uh in the end we we added two two new uh gameplay elements and that was a defensive uh capability uh basically block uh as well as adding uh full Left Right controls the actual Left Right controls were added in the last month of of development um they were added this time last year during GDC um and it it

completely changed uh the feel of the game uh instead of actually requiring uh dexterity and only one thumb we were now moving to a situation where we required dexterity and two thumbs it was a much more complicated control scheme um that was that that players need to need to work with so this changed our target market um it moved us from a casual player to more experienced gamer um but we felt that it was it was worth that that that compromise so our final control scheme uh looked like this um again all we did to accommodate the the the new block moves and the left moves uh was replicate what we felt was working really well on the right hand side over to the left hand side of the screen um Right Movement is again um quite large uh block is very large it's very easy to to get up there uh like I said before the tucked command is the more difficult one uh thankfully our game is designed such that there's a lot less left movement than there is Right movement and so we were able to design around that uh and this was good uh the full Left Right Movement allowed for a lot of

new gameplay elements we have Arena combat uh we have enemies that can Ru Dodge attack from behind uh and we were able to create some better interesting uh platforming segments with this new control scheme so that was good the last tweak we did uh before we actually shipped the game uh was to allow for sliding between hit pads so originally what we had was you had to release your thumb and then touch another pad in order to to activate it uh one of the last things we did was allow the player to slide between them this was uh completely necessary for Left Right Movement so the player can just kind of move their thumb between those two positions uh and it also worked really well for doing things like jump and shoot uh allowing the player to shoot well in the middle middle of a jump command uh and that felt really good so you can kind of slide your thumb between those three hit pads so some some closing thoughts here um so this control scheme uh it's been really successful uh it's allowed us to have six hit pads with full modal capability uh this allows us to do things like charge shots for all the

different weapons it allows us to do Mario Style jumping uh to allow different height variants with that uh it's very easy to learn uh nobody that picks up the game uh has a hard time understanding the control scheme it's it's right there in front of them uh and it's really responsive and predictable uh nobody ever says things like when I'm I'm hitting a button I'm hitting the wrong button it always seems to work really well uh but we are continuing to refine and iterate on the control scheme I don't think this is ever be a done thing uh but we're working on it there's areas that we can improve on um I'm still not happy with some thumb sizes have a hard time reaching some areas and so we're looking at ways to customize the control schemes and customize the scale right there to to allow people to fine-tune it a little bit more uh and we're also looking at ways to have better graphical uh feedback that better conveys the different hit areas as opposed to just using the buttons uh to convey that information really this stuff is is super hard um you know controls for mobile games you don't have this predefined user experience to kind of work with you have to build all the this yourself there is no physical input and

when you're making a mobile game you have to fight with grip device size screen real estate all this crazy hard stuff what you know what we were doing here is we didn't try to uh apply this tactile uh overlay virtual d-pad type thing to the game we knew this was just wasn't going to work but I feel like we only really scratch the surface of what we can do with this you know there's there's a lot of problems in our industry and there's a there's one big problem that is our our huge Reliance on pass control SCH te if you had a controller to someone who's never played a game before the learning curve on that is just super high and I think there's a lot of opportunities for us here especially in the mobile space uh to create game experiences that allow us to change the way that people play our game change the way that they interact with our game you know with this platform for the first time we as game developers can actually take back the control of controls um yeah and so here's my last inspiring ins slide uh let's you know let's try to move Beyond this tap tap tap and virtual d-pad stuff that we're doing on mobile

on let's try to innovate and create new ways for our Gamers to play our mobile games and that's that got lots of time for questions now um hey there hey um did you guys try any sorry any AB testing like with alternate control schemes uh so the question was did we try any AB testing with alternate control schemes uh we didn't really have a lot of time to do a lot of uh user testing so we had some small internal groups uh of varying people with varying skill sets and what we'd actually do at points where we had about three or four control schemes in the game that you could switch between the options menal menu and have people play through all the different ones and give us feedback on that and that allow us to kind of fine tune stuff near the end when we were doing that left right control stuff

we had about four control schemes in the game at that point in time that we were toggling between and kind of getting some internal feedback at that point okay thanks cool uh so were you satisfied with the auto making your game auto run uh and or did you how far did you get with the Left Right Stuff and have you like had more thoughts on how you would move forward if you tried to do another one um so we shipped with the left right so we uh we so the question was did we uh were we happy with the auto run stuff uh we shipped with the Left Right controls um I was pretty happy with it um I was super against Left Right controls I was fighting tooth and nail to keep it uh just moving right um I didn't like the additional complexity of the control scheme right there um but once we actually got it in there and we started looking at the different gameplay elements that it opened up it was super awesome uh one thing that I'm still not happy about is um with the uh super extended thumb positions while you get that uh confidence that you're hitting the right thing it's a lot slower than having small smaller uh button controls and so you lose a lot of Fidelity in

moving left to right and so those really tricky platform elements are still quite hard to um to put into the game because the controls aren't really precise enough so one thing I'd like to to work on is uh what that right balance is between game design of level design for those platforming elements and uh where we can kind of take the controls to be even more responsive uh while still maintaining the um predictability of it cool hi hi I wonder if you you discussed making um the user able to adjust the position of the buttons as for their own thumb size yeah so the question was if we've uh figured if we've considered allowing the player to adjust the positions of the buttons um one thing that's brought up quite a bit is moving buttons between those different places quite a bit um I'm not against that we just haven't really done it yet uh I think one thing that's more interesting is probably scaling the size of those hit pad areas more than anything I think that's one area where I think players can get into to something that's more fine-tuned for their thumbs um I'm not

sure if it'll be as effective for players to change the angles between those two different between all the different hit areas um and there's also going to be some some problems if the players are trying to uh move buttons to different areas like so one one comment that people brought up was that they want to jump on the left hand side but the problem with that is then they can't move and jump at the same time and it creates kind of weird Edge case type stuff so we're looking into it right now and just trying to figure out what what we can do that makes the most sense hello hi uh you alluded to precision as being still kind of one of the problems you have with the control scheme is there anything else besides that that uh you're still kind of see as a hurdle something you want to get past so the question was uh uh other areas that we feel like is a hurdle to get past um so I don't think Precision is is a problem I think the um quickness of the of the responsiveness is is a problem I think it is very precise if you if you're going to hit something you are going to hit it it's going to be good um the one area that I want to get Beyond is some

people still find it UNC uncomfortable um and trying to find a way to get that grip um comfortable for more people is I think a big area that I want to work on next um to kind of see what I can do um you mentioned about having sort of Mario Mario Style jumping where uh which to me sounds like you can control the height of the jump is that correct corre um did you have any challenges like keeping that sliding sort of um uh input and still having players be able to kind of control that height accurately since their their thumb would still be sort of resting on the button when they were trying to slide yeah so so what we ended up doing was um um while there is Mario cell jumping in the game it's a lot more subtle than a lot of other Platformers um we didn't we don't have it so that it's it's uh super required it's more of like a subtle feel thing in our game than anything like that and so there's no gameplay that requires you to do a really precise prec Mario Style jump it's more of it's there you can you can

kind of feel it if you release the the button early you'll see that your your jump height isn't as high uh it's just one of the small little polish things that we we did as opposed to uh a big gameplay type thing anything else cool thank you so much everyone and have a great [Applause] Friday