Level Design Case Studies: Trainyard and Cut the Rope

[Music] okay all right thanks for coming to this presentations so the way this is gonna work is basically I'm going to talk for about 25 minutes about the level design in train yard and then Semyon from Zeptolab is going to talk for about 25 minutes about cut the rope and then afterwards we're gonna open it up for some questions because is the last session will have a bit more time for questions if there are some all right so this is a level design post-mortem and really what that what I'm gonna be doing is just kind of giving my level design philosophy I followed when creating train yard so my name is Matt Rick's and I live kind of outside of Toronto so that's me but really what I want to talk about is train yards so to give you the super brief story of the game in 2010 I was working at an advertising agency as a flash developer and I was getting really sick of my job and I've always loved games but I'd never actually made

one and so I decided to make a game in my spare time so every day on the train to work and whenever I could find time I would work on this game and over the course of the year I did all the parts and finish it all up and released it on the App Store and so it came out and at first it didn't really do that well so feel free for a few months everyone seems to love it but the sales weren't very high and then I released Trainyard Express which is the free version and the game got a whole ton of buzz and through a variety of different things that climbed all the way up the App Store to number two actually behind only cut the rope and it stayed there for a few months in October of 2010 and since then it's kind of dropped down I don't know if anyone was in Phil Hasse's talk a couple days ago but I really liked how he gave out some some real sales numbers so Trainyard is priced at $0.99 and to date it's been bought over 800,000 times and and the free version trainer Express

has been downloaded over 4.2 million times so for something I made my spare time and for pretty much the first game I've ever made it's been pretty amazing really so that's the story of the game and now let's talk about the actual gameplay Trainyard is a logical puzzle game the the only thing you do is draw tracks for the trains to follow that's it so the way it works on the left there's a blue station on the right there's a yellow station the trains are going to come out and follow those tracks and combine to form a green train so color theory plays a huge role in the game and the timing and mixing of trains whether they either crash or don't crash plays into it now again the only thing you do is draw tracks there are no points there are no scores literally the only thing that matters is solving puzzles so that's the basics of how the game works but again this is about level design so level zone to me is all about teaching it you can almost look at each level as a lesson and the overall flow of the game as a

course you're teaching the player how your mechanics work you're trying to get to a point where you've taught the player how to become a master at your game how'd it be an expert at your game and so that was what I was always trying to do throughout the course of the level design and train yard and as a teacher your most important consideration is really knowing who your students are so in the case of train yard despite the fact that it's kind of like a hardcore logical puzzle game I was aiming it at the the very casual app store audience and so that that had a number of implications for my level design I think there's this sort of perception out there that casual players are somehow like less intelligent or dumber than than average players and I don't think that's actually the case at all I simply think it's the fact that they haven't been playing games their whole lives and so they're missing out on all this experience in this whole vocabulary of games that most of us just take for granted and so because of the fact that they don't have that vocabulary I needed to take the level design and train yard

and start it from a much more base level to make sure that they really understood all the core concepts if you're making a game that's targeted at actual like hardcore gamers you can skip a lot of the steps that I made sure to include in train yard all right so as a Starcraft player I'd like to break everything down into micro and macro so in the case of train yard I see my cursor in the case of level design I see micro as being the specifics the individual design of actual single puzzles and then macro being the overall flow of the game the way though all the levels fit together and kind of the difficulty curve of the game so first of all I'll talk about macro one of the most important things I was trying to do was to add elements progressively making sure that I was only teaching one single thing at any given time so an element could be something like a gameplay mechanic it could be some sort of technique some sort of way of using the current elements in the in the game or could be almost some new form of practice to teach the player how to use the elements that they already know and along with

that I also wanted to treat combinations of elements as if they were a new element and I'll have these diagrams that hopefully make that more clear so in in traditional gamer focus games you have this sort of approach to teaching elements a lot of the time so you teach element a the next level you have element a and then you introduce element B and the next level you have a B and then you introduce element C and there you go you've got your levels but in Trainyard because of the casual audience I wanted to break it really down to a more granular level so I would teach a and then teach B in isolation and then teach a and B together and then teach C in isolation and then give it an example with a and C and B and C and finally come to the end when there's a B and C now obviously mathematically looking at this a couple more elements and you're ending up with with thousands of puzzles so it's not practical to do this but at least taking this sort of approach to level design really breaking it down to the granular parts and making sure the players actually understand all the

nitty-gritty things that you're trying to teach them okay so instructions are a big thing in game and George fan had a fantastic talk this morning about this in plants vs. zombies and I kind of have a few similar takes on it here so in early versions of train yard there was kind of a wall of text teaching how to play the game and that was totally ineffective no one really reads text and so I was trying to figure out what approach to take and I ended up looking at things like this so this is the nykeya like page from an Ikea manual and I think we can all agree that most of our games are probably easier to then assembling a piece of IKEA furniture and so I figured if this works for them it'll probably work for games and I discovered that in fact games like Angry Birds there's a picture there's no text there and yet you can understand how to play this relatively complex physics game with a single picture so it's really really powerful and there's actually a couple other benefits to this or there's one big benefit is that

localization becomes less important because all of a sudden anyone from any culture can understand how to play your game just from seeing a single picture so there's one there's something I like to call the wingdings test which is where you take a game replace all the font files in the game with wing-dings and then give the game to someone who's never played it before and if they can understand it with all that nonsense text there then you know that you've got a game that is no longer dependent on text at all and that's that's actually a really fun task to do all right so with Trainyard it's much more complex than a game like Angry Birds I couldn't just simply show them a picture and have them understand how the game works and the best way to show someone how a game works is to show it actually being played on iOS file size is a premium because of kind of the 3G download limits and so I needed to do something else and what I ended up doing was this sort of system so this is the tutorial just like one screen from the tutorial system and it's it runs on top

of the actual game itself and it's kind of I spent like two months of my development time just making this tutorial system alone it's really robust xml driven very adaptable and it's running on top of the real game engine so there's there's this super ugly hand you can see on the right screen that hand fires real touch events into the game and does exactly what a player does when their finger touches the screen and you can see that I've used text there but the text is only there to add some context and some terminology it's definitely not necessary to understand the game at all and they can actually watch the game being played and see all the techniques being done only teaching the player what they need to know this is George also touched on this this morning but front-loading everything that you can do in your game all at once is not a good way to teach the player they they don't know if you're getting all these instructions right up front or all these tutorials how do they know which thing they're actually going to need to know right now there's no way of them choosing so they

just choose one thing to focus on and it could be the totally wrong element that they actually don't need to know at all something else that can happen is that you teach them everything up front and then you have this whole game for hours and hours and hours and all of a sudden they get to some super hard level in the game and you have this element that you only taught them at the beginning of the game and they've long since forgotten it so this the easiest way to solve this is just to split your tutorials out over the course of the game and in the case of Trainyard when you start it up the first tutorial is literally teaching the player to drag their finger across the screen and that's it and then they have a few puzzles that let them practice that then there's another tutorial and it teaches them how to drag their finger in a corner shape and then there's a few more puzzles that let them practice the last tutorial doesn't come to over 70 puzzles into the game like 3/4 through and and so they've been playing for five or six hours but it spreads it all out so that they know what they need to know at a given time this seems similar but it's actually kind of different so you

can teach players things through tutorials or you can teach things through gameplay but it's important that players have an opportunity to actually practice those things and try them more than once if you teach someone in a tutorial and you have a level that uses that tutorial or uses that skill but then you never come back to it for the rest of the game and then all of a sudden I see a lot of games do this they'll teach something right up front and then you get to the end of the game and all of a sudden there's this ultra hard puzzle at the end of the game and it uses that element and you haven't actually seen it for hours and hours and hours so I think it's important to come back to things over the course of the game make sure you're giving them a chance to practice things not just when you teach them but come back to them over the course of the game again and again so this is just an abstraction of this sort of difficulty curve I was trying to follow when I created Trainyard so you have this ramping up in difficulty and then every few levels you drop it off and give them kind of a break and this is just having a constant

ramp and difficulty can be incredibly overwhelming for a player you never really get chance to relax and so these dips is given the chance to kind of chill out and do something and feel good about solving something that's easier they feel kind of content that they actually have some skill and some mastery over the game and like I was talking about in that previous slide these kind of dips in the difficulty curve give you a great chance to go back to some earlier mechanics some simpler mechanics and make sure they still really understand them and have an opportunity to practice them so level unlocking is one of the things in Trainyard that's probably one of the things I'm least proud of it's so the way it works it's kind of confusing to explain which is pretty much a given fact that it's actually too confusing in general but so each puzzle is assigned a number of stars based on how difficult it is so an easy puzzle one or two a medium puzzle four or five a hard puzzle nine or ten whatever and then when they solve those puzzles they're given that number of stars and then with certain number of stars it unlocks the next

section of puzzles and yesterday I was actually talking to Semyon about this and cut the rope has something very similar but the difference is that in cut the rope every puzzle is worth three stars and in Trainyard they're worth variable amounts and I think some players it wasn't a huge deal but I think some players really got confused by the fact that they didn't really understand where these stars were coming from and that could actually just be an issue in my communication of it more than an actual problem in the game design itself so something that happened a lot because I had this unlocking mechanism I would get a lot of emails from players that said you know I got a new I got a new iPad I got a new iPhone and for some reason iTunes screwed up my dad and I can't I can't get it and so I have to start over from the beginning and I really don't want to have to play all these easy puzzles again and at first I was kind of like I don't know what to do you know the easy puzzles are easy just go through them I guess and then I realized there was a really easy solution and it was just to put a button on the option screen that when you press it it allows you to unlock all the puzzles it says are you sure you want to

do this and if you do you just have the whole game available to you it's really like like a cheat code from an old-school game except instead of making a cheat code some obscure thing it's just right there in options if you want to do it and this solves so many problems with like syncing and all that and it took me probably an hour to or not even that much it was like a huge amount of effort for the amount of actual results of God it was really awesome something else that worked really well was having a main game and then bonus levels so in train yard there are a hundred main puzzles in the game they kind of constitute the main game and what happens is when you finish those hundred puzzles a screen pops up and it says congratulations you beat the game and then you can post it on Twitter and Facebook and then right down below that there's a little button that says want more puzzles play the bonus puzzles and to me the real game is in the bonus puzzles there's sixty more puzzles and they take way more time and way more effort and there are much more interesting puzzles but a lot of the

reason I did this was because I knew a lot of casual players and this is exactly what I know happening a lot of casual players would get to that screen when I said you beat the game and they would kind of close Trainyard and be completely content with how far they got and so those players left the game really happy and if I ever make like a trainee or two or something there'll be big fans of it and then the hardcore players who are the ones who are better at the game than I am anyway they would go into the bonus levels and have a blast playing all these super hard levels so this this worked pretty well except there was one issue that I didn't really consider and it was this kind of mystery demographic so there's the casual players and the hardcore players but what I didn't realize is there was this also there was also this demographic which I kind of realized afterwards that I fall into this demographic of myself as the players who just want to complete everything they're not necessarily the best players at the game but they want to complete everything you throw at them anyway and so what was happening is that the casual players were happy the hardcore players were happy but I made one kind of fatal

mistake and that was that every puzzle I made I put not every puzzle but every kind of hard puzzle I made I put it in the game so I had some puzzles where I had spent like five six seven eight hours solving them and finally solved them and they're these like insanely hard puzzles that I put at the very end of the bonus puzzles and so these completionist players would go through all the bonus puzzles and then get to these insanely hard puzzles and just get frustrated and quit because there literally puzzles that I myself struggled with and and the hardcore players would still solve them anyway so I came to us an easy solution is basically just don't put those puzzles in the game at all they casual player will be happy completionist player happy and the hardcore player they won't be happy no matter what you do so so what's the point really no but I actually found another solution to the hardcore player and that is creating a level editor and so I did that much later on and that was actually a huge success and all the

hardcore players are happy so now everyone's happy and it's all good times all right so now moving on to micro the specifics of individual puzzles each level needs to have a purpose so when I was anytime I would kind of have a blank slate there in front of me I would know what I wanted to accomplish with that puzzle before I put any pieces on the board so kind of like what I was talking about earlier a purpose could be like teaching some sort of element some sort of mechanic giving them practice or it could be some sort of theme some sort of idea I wanted to evoke through the design of the puzzle so there's a lot of different approaches that can that can be done there but the portent thing is that you know what you're gonna do before the actual creation of it and visually I really like to aim for symmetry and balance I want everything in the puzzles to look intentional to be will let me just show you an example so this is a puzzle called magic carpet for some reason and so you can see there's

kind of a diagonal symmetry there with the rainbow effect and really the reason I do this and the reason I have that kind of symmetry and balance there is that I want everything every puzzle I make I wanted to look handcrafted I wanted to look like someone spent a lot of time and effort actually making this puzzle I don't want it to look procedurally generated like some sort of algorithm just spit out another puzzle for the player to play I really want it to be just like human and so that's that's why I go with that now a Yas for another example for this so so this is one of the player made puzzles kind of all my examples of bad puzzles are from play your puzzles so but uh so this doesn't really have that symmetry it doesn't really have that balance to it but the thing that really bugs me about it is that rock in the middle of the screen rocks and train yards serve one basic purpose and that's to limit where you can draw tracks and it limits your

solution pace space for the puzzle you can be incredibly powerful incredibly powerful way of making the player think in certain ways and use certain techniques but in this case it just kind of gets in their way it doesn't make the puzzle any more interesting it doesn't make the puzzle require any more cleverness it just makes a puzzle take longer to solve and one of the benefits of train yard or one of the greatest things about the the gameplay mechanics is the fact that there are many possible solutions for every single puzzle but all of a sudden this rock just kind of gets in there and restricts their solution pace but doesn't make the puzzle any more fun or anything so really not a big fan of just arbitrary elements thrown everywhere using themes so this is a puzzle called cooks Ville Creek and it is meant to evoke the feeling of a river now unfortunate I don't have a video of it in action but if you can imagine a blue train will flow from the top down to the bottom and then split into four blue trains and then flow to the left side and to the right side and across

the bottom and finally up to the top right now you don't really have to visualize in your mind but the important thing is that it has this kind of blue flow throughout and I was trying to evoke the feeling of a river and there's no real perfect reason for that the only thing is I just wanted to look handcrafted I wanted to look like someone took the time and effort to make something interesting so this don't overwhelm the player so this is another player made puzzle and anyone who's playing Trainyard instantly is going like what is that because it's there's like 60 trains there and 60 goals and it's just an overwhelming amount of information and the way I kind of like to look at this is that each player I almost think of it like a computer or something every single player of a game has a certain amount of memory a certain amount of state that they can hold in their brain at any given time and if you're if you're starting with too many elements on the board it could be the easiest pause the world to actually do the solution of but if you're starting with too many elements on the board they're just gonna

get overwhelmed and confused with the amount of information and so whenever possible I'd like to use the absolute minimum number of elements I can to get whatever purpose I'm trying to achieve across there's no point in throwing more things in just because I can I really want to keep everything to the absolute minimum necessary when when beta testing or or when watching someone else play the game if you ever hear someone say I don't know how I just solve that puzzle or if you're watching someone and they start devolve into trial and error to me like try a bit of trial and error like educated trial and error is fine but when they it just straight up trial and error that's like the ultimate bad thing for a puzzle game because they might as well not be able to solve the puzzle at all and so that either means there's some problem with your core gameplay mechanics or much more likely that your difficulty curve up to that point and your levels up to that point aren't teaching the player the skills they actually need to know to solve that puzzle so that's a huge problem at least in my mind that's a huge problem that needs to be solved somehow so having a

level editor so this is the trainer level is like this super ugly Adobe flex thing but it works for creating puzzles and it's visual which is great but the most important thing you can have a level editor at least in my mind is some form of versioning you need a way that you can quickly clone and duplicate puzzles and try out experiments and if they don't work you can go back to whatever you were doing before and another thing that's really important is being able to switch between solving the puzzle and creating the puzzle as quickly as possible or a level if you're making level based game whatever you need some way that you can quickly swap between them if you're waiting for compile times or something like that it's going to drastically amount it's going to drastically impact the amount of experiment experimentation and iteration that you're actually willing to do and finally I want to talk about making players happy so when you're creating a puzzle game players love or human beings in general love solving

puzzles we love figuring things out so there's no reason to go overkill with like experience points and all kinds of blinky things on screen all the time you don't need to do that because players love the fact that they're solving things and figuring things out the only thing I really tried to do is when the player is already feeling that that kind of internal joy from solving something I like to amplify it so in train yard when you solve a puzzle that's when I go crazy with particle effects and sound effects and all that sort of stuff it's really kind of lock in the feeling the player is having and finally discovery of on top mechanics so a lot of this has just been about teaching things kind of explicitly but the most powerful thing you can do as a level designer is give the opportunity to the player to figure things out themselves so this is that puzzle color theory from the start of the game and color theory is interesting because it comes twenty levels into the game and up to this point I haven't ever told the player that colors can mix together I haven't showed them that that's possible

at all and so I put them in this situation where they can do nothing else but discover this mechanic and I think that's the key thing is that if you're making game or focus games it's pretty easy to to kind of assume they have some knowledge and can figure things out but for casual players they can get frustrated really quickly and think that there's just some knowledge that they're missing to solve a puzzle so you need to kind of give them the minimum amount of information not information but a minimum amount of room to work in it's almost like you don't want to give them too much rope and so that's what I was trying to do here and this this specific puzzle I've heard so much feedback from people about how this kind of this was the moment where the game changed for them when they discovered this for themselves alright so wrapping up that's basically all I have to say now I hope some people can find this useful not just for puzzle games but for any kind of level games yeah so that's about it thanks for listening and that's my stuff now Samuel

is going to come up [Applause] all right so whoever run my name is Semyon I'm a creative director and co-founder of Zeptolab and I will continue my talk on the level design from the perspective of our game cut rope and I will also share some sources of information that we rely on when we're trying to analyze what are the things that we did right and what what could be better and it will take another 25 minutes I mean my part overall and then Matt and I will answer your questions so in case you haven't heard about cut the rope it's it's pretty popular game which was initially released for mobile platforms and it happened around one half year ago and it's it's built around really simple principle of delivering object from point A to point B by cutting the ropes

in some specific order and with right timing and you can see how the game evolved from the very early draft to the final look and you can see that point B ended up being this green green creature code unknown and because the game was originally intended for mobile devices it uses a really short session approach with the with the levels so many levels are even most of them can be really completed in several seconds so game can be really considered a snack of gaming and those snacks are the levels and our job as designers as the out designers to make sure that those snacks are tasty so in order to do that we set three goals from the very beginning and first of first goal is that we want players to feel smart it doesn't mean that they aren't but we want to remind them about it that's the kind of purpose of our

game overall the second goal is that we want to we want to appear to really wide audience of the player from hardcore to the very casual like the one you can see on the picture and the third goal is that we want players to feel hungry for more and that's essential for us to keep the game hind the chance all right so we define the goals now we needed some set of principles we would use two while creating the levels and there are seven of them and I will now go through seven so the first one is the positive reinforcement and it's not a big surprise that it works much better than the negative one and in this example if you play the game you know that you can move that blue holder on this blue rail and if you on the first picture if you move it really fast to the right side the Rope will pull the candy up and it will do this kind of jump and end up in

the mouth of the little green monster so in the first example we kind of forcing player T to do this trick by placing the spike Center however the better way of doing it would be placing the star there so first it would be a positive motivation for the player it would encourage him or she her to do this trick rather than forcing and what what what's even better that the employer would have an option of not collecting collecting the star at all and this that really helps us with the goal of appealing to the white audience because you know many kids don't really care about the stars the all they want to do is to see unknown happy so it is good to have several of options for different kind of audience and what what even greater about the the stars that it's really a self manageable difficulty was in your game it's up to player to decide what like how hard the game should be

for him whether he wants to collect all the stars or just complete the levels and it's and and you know also all of those kind of players will feel satisfaction because they set the goal for themselves and they achieved it and to give even more value to the stars will introduce this this functionality of the level books as being unlocked by some specific amount of stars and we might choose that of course those values are not too high so unless well unless player is keeping too many levels he'll be able to collect necessary amount of the stars and by the way skipping the levels is another cool thing that we got out of the system we basically let players keep any any amount of the levels he wants of course unless he hits the the next the next level box which helps to minimize the frustration of you know being stuck at some particular level so the next principle is that one player launches

the the level he should already be able to plan the solution and what's over all the beauty of the physic base like games is that players already knows the biggest bunch of the rules from the real physical world and that's the reason why even small kids can play this kind of games because they they don't know the mass I don't know the physics but they know how these things work around them right so the only thing the only additional thing that way we basically need to explain to the player is cutting the ropes and they already get it from the title of the game so that's pretty cool and but but here's the tricky part even though our physics engine is pretty cool and we're proud of it of course it's not as complex has not detailed not as detailed as a real life so it is possible to trigger some events within the levels which would wouldn't look natural or would look unexpected to the player so our job is to make sure that such things don't happen and as a

proof you can see this picture taken make sure from the YouTube video where apparently one of the game fans have replicated several levels from the game like in real world and it seems to work you can check the video on YouTube and just to give you a bit more details on this principle here's another example where it basically it's not only about the physical behavior of the object it's also about how we put the object within the level and in this example they kinda float up in a bubble and on the first picture it's not very clear whether it will squeeze between those spies or not so the better way of doing it would be either moving the spikes further away so that it would be absolutely clear there the candy will pass or moving them really close to each other or maybe even uniting them so to clearly communicates as the candy won't pass we can go as far as you know placing a little candle there but it might be a bit of an overkill and here's the principle that

might also talked about basically it's not only about that the solution should feel illogical it's also that the wrong way of solving the level should feel wrong and we've seen many times when we were doing play testing when some player was I was playing some level and trying some wrong way of solving it over and over and over again and wasn't even considering the option of the alternative ways of solving it and that happened because that the wrong way of solving wasn't communicating clear enough that you know it's wrong way it was inheriting that I almost did it last time so I'll give it a try so it should communicate something like you try that good job didn't work try something else and in a way it limited about it a bit limited our game with the amount of solutions every level has so it's not a

really a sandbox approach they're usually just one or few ways of solving the level levels and there are some good and bad things about it and the good is that we can really control the experience of the player all right and here's another thing is that no game is perfect including cut rope of course and it is possible to trigger some visual glitches some technical bugs some control controlling issues by placing some elements in some not really smart way so we need to make sure that it doesn't happen and in this example again with the blue rope holder if the rope was really short and player wants to cut it he can accidentally move the blue holder instead of cutting it and that's a bad experience so at some point we agreed that we would move it sets a minimal amount of the rope attached to such holder and we won't use any shorter ropes and it already solved these kind of problems in in many levels

another thing is that as you have pressed the button on the top right corner of the screen we don't want player to do any actions like putting the ball bubble or pressing some mine only like ear cushions within the level because he can occasionally press the reset button and that's again not but not a good game experience so we're trying to make sure it doesn't happen another thing that might talk about is that well of course everyone knows that if a tutorial that is a very important part and I agree with Matt that images working even better than text but another important thing here is that we realize that if we're just like trying to explain something to the player but there's still a space for completing the level without using that principle or using that gameplay element that were showing in sometimes sometimes players will just skip it and and then they'll get

frustrated later when they'll need this knowledge so basically people need a reason to learn and in our case reason is usually that they can't complete level until they try what what they what we want them to try and in most cases I don't really need to read the text they can just play around with the with the objects and they'll find out how it works and here is the most important principle and you'll see there is a proof for it principle in our game in our lair was quite a lot it's okay to ignore any of the principles and the the only requirement here that there should be a reason for that and in our case a reason is usually some cool combinations or combination of the objects or some cool idea and the benefits of having is a coolness factor of it is much higher than the harm that it causes by violating some of those principles so

it's okay we think it's okay to break the rules and in many cases it works for good so that's it for the principles I also wanted to share some some things that we do with the level ordering and this this spreadsheet might look a bit complicated but we kind of got used to it so it's really good visual tool for us to see how the levels are going the basically has the list of the levels the list of the gameplay object like elements and we can see which objects which object is used in which level how the the the how the unlocking rate looks by this gray area you can see the tutorial levels you can evaluate the arcade and puzzle kind of levels so it's pretty cool and we're also easily trying to make the last level not the hardest one I mean the last in the level box which is basically like each 25 each one of those have 25

levels the hardest level is usually the one before the last one and the purpose of the last one is it sort of directed to the goal of keeping players hungry for more it's usually not very hard not to avoid causing any frustrations but it's really inventive for trying to keep some really cool idea to the last level within the box and it should basically communicate to the player that the the best part is still waiting ahead and that we haven't run out of ideas and another thing that proves that we haven't run out of ideas are the new gameplay elements that we add and from the very beginning with every update who set quite high bar of expectations by adding new gameplay elements with every update and we wasn't sure how many of those who would be able to come out but so far we've been quite cool inventing those and our process of coming out with those gameplay elements

that really changed the mechanics by the way in many cases it's well thought is quite quite traditional we brainstorm who pick the ideas who like most we try dropping the levels on the paper to understand whether you know we will be able to create some cool set of the levels based on that then we prototype it and know tweak the parameters start ratings level levels and put the final art and from the very beginning the logic for adding new elements for us was to add more directional freedom to the candy so there was a gravity affecting obviously pulling Zeke and it down so we needed something that would move it up so here goes the bubble and then we need some horizontal freedom as well and we add the air cushions and what what's also cool about them and what we're trying to keep with all elements that we add that those are intuitive and fun to interact so it's fun to pop the bubbles it's fun

to click on the air cushions because they make farting sounds and stuff like this it's really cool it's important to keep the game entertaining entertaining alright so another so Eddie new stuff is cool but we also have a luxury to learn from like each time we add at new stuff and and and do new update we can and we shoot learn and analyze what we did right what we did wrong and not only we can use this knowledge to you know while adding new stuff we can also look at the existing stuff and maybe tweak it a bit maybe improve it maybe change it there's some controversy about changing the levels in the game after they being released we feel it's okay to do it if it you know keeps the if it doesn't harm the scores of the players and if it's it will not go into wild with that so to understand what needs to be

changed we're using several sources of information and one of them of course being our user reviews this is a great channel to receive the feedback overall and you can quite often see also the feedback about the levels like this one's saying that the level 9 is 1 then both is impossible why would you do this to me and if we see like several of those there's obviously something wrong with that cell right another great thing is to have players as fans on your Facebook page not only it brings some marketing possibilities but also it's a great way to to communicate to like that talk to your players and we're also from time to time asking some things that helped us tremendous is on how we'll move forward and how we may be improved existing stuff surprisingly Game Center can be also used as statistics tool so we keep leaderboards for each one of the level

boxes in the game and we can see in this case the amount of quickly submitted scores for each of the level box so this way we can see that obviously not all players leave reach the very latest levels of the game we have plenty of also been updating for one 1/2 year but still we want you know as many players as possible to get to the end so at some point we rearrange the level boxes a bit with some levels and we were able to to improve the percentage of the player that we see and the end of the game by 5% it might find fun it might sound not like a big number but considering the audience of the game it's huge and there's of course like more complicated more complex and advanced ways of getting data getting statistics like flurry and four level designers the most important metrics are trailing rate and skipping rate which we track for every

level in the game and basically failing rate is is the amount when players sees the sad face of unknown when the candy gets broken when it falls out of the screen this kind of things so it if this parameter is high usually it means that players get the they understand how to solve level but they can't execute it well so it means that probably we need to simplify the arcade part of the level if the skipping rate is too high then probably players like can manage the difficulty is a timing challenge of the level but they just don't understand the puzzle component of it so then we maybe should simplify it a bit or add some tip to the level that we'll start with that would keep how player to understand how to solve it so and it's really tough balance really on the one side we want as many players as possible reach to them reach the very latest

levels of the game and then like impatiently waiting for updates at the same time we of course want to keep some challenge in the game so it's a really hard balance and we're trying to keep it and we're trying to learn and that's that's one of our biggest challenge challenges so the results that we got so far as as well designers we created more than 900 levels and there's a half of them ended up in our games obviously because we're trying to create more levels that we need and that would pick the very best of them and we put them in our games and we're proud to constantly invent new things new gameplay elements and there have been already more than 22 M added to the game but the most important achievement of course and it directly serves the goal that I've highlighted in the beginning is that players once more and that's pretty much feeds the expectations of the little green guy alright thank you very much

Matan I will get the last request there are microphones I have one question so in the updates of Calabro you said you it's terror treatments so did you change any like existing that levels and how did you change that like so like so that it doesn't affect the balance of the game or like the existing state data of the player well we will making sure that on the technical side there is no problems happening sometimes we like there are some bucks of course coming up and we're like trying to fix them as fast as possible but basically it's more of a technical challenge it can be solved pretty easily our from the design perspective the most

important thing is that players who wouldn't wouldn't really get upset about this kind of change and from our experience they don't they they notice that maybe something I've changed but they're okay with it yeah Howie just came to the ideas of your games I mean this trains this role and actually about rope I mostly who is this guy I'm sorry how did we come up with the ideas but again so yeah that's a funny thing I I came up for the idea for train yard while I was riding a train to work and so it's just like wait a second this would death yeah I really wanted to do something where there was something to with paint but I mean there's no real paint in the game but the idea of colors mixing together and then it really felt like I was almost discovering the play mechanics rather than inventing them it

was just like once I started looking I was like wow there's like a whole lot of really interesting possibilities here and for COD rope we basically had the physics engines the rope engine already in place it was done for our previous game but it never was used there so then we started brainstorming of like about what can be easy the gameplay we can we can build around that that rope idiot and it was really a good starting point we've made several prototypes around the rope and one of them was the best and it ended up being the game this greasing oh the green thing it just product of many many doodles you know and experimenting and he he really plays a big role in success of the game I believe because he's he's kind of like a baby monster so he in a way he kind of encourages the parental feelings so it's not the each level is not about getting some amount of score but rather it's you know satisfying the

needs of this little bring cute cute creature so this is the emotional relation the emotional goal is much much bigger than any you know competitive goal thank you thanks hi I have a pretty picky question for Matt oh boy I love train yard but I notice the the color mixing trains don't follow the primary colors of light and I just wondered if that was a deliberate decision on your part well the yeah I've had a few people complain about whether I should be using like subtractive or additive color systems I mean the basic thing is just that people seem to understand subtractive color systems more than additives so the people seem to really understand the fact that blue + yellow makes green rather than using like magenta and whatever so it's yeah I mean that was that was really a reason people just seem to really associate with it and I think part of it was just because of the fact that the game was really originally was supposed to be about like these these cars of paint and there was

like almost like a pink crash on the middle of the railroad where these colors were mixing together yeah great thanks so Matt you sorry I might yeah there we go you mentioned that you'd created some levels that took you like up to eight hours - so I'm wondering how do you go about making a puzzle that's hard for you yourself to solve but not impossible like how do you know that you've created something solvable so challenging yourself so the problem with train yard is that there's no way for me to really know whether something solvable at all so I would just literally just sit there and have kind of basically the my first step is just to work it out mathematically whether the colors add up and the number of trains to add up and then just make this puzzle and try and solve it and try and solve and try and solve it and I have a few like a pile of puzzles that I've never been able to solve that are in my pile of like moby-dick puzzles that one day I will conquer but yeah it's just really a process where I try to solve them and sometimes I can't sometimes I can how much time a single

puzzle sake for a designer to make and how many level designers were working on the later levels of control usually takes I would say several hours like maybe two hours or for the for the basic like for the first iteration of the puzzle and then of course when we have the full set of the of the puddle foot level box then we all play and we do play testing and then more iteration set so it's like iteration process and it's it's it's not that much related to some particular level but it's rather looking at the levels as a whole so we have I think three level designers at the moment working the levels for our game right thanks thank you

some young when I saw when you showed your spreadsheet so you had a difficulty rating on a scale of one to ten in a pre-release tape before you have like high volumes of playtesting how do you get a reliable measure of difficulty there is no reliable measure if it's a gut feeling really and we were asking different people to play these levels and then put the difficulty numbers that I felt how difficult is and then we were trying to calculate the average but then you know you can't really the trickiest part is then you can measure the difficulty of the arcade game play but with the puzzle you have no idea like one person can come out with some solution right away and some will spend like several hours you know completing it thank you this is for a train yard how do you determine the grid size and how does that affect designing your levels it's a good question so the story of trainer is actually that it started out as a flash game and originally it

had an 8x8 grid and then I never released the flash game it was just like a prototype basically and then when I put it on iPhone 8x8 wasn't working and I really liked the the the kind of not quite balanced of having that middle row with seven like I really liked that for some reason and yeah so basically seven worked really well with being something that your finger can actually go through the puzzles really it was there a second part of the question or that was there well actually I guess wouldn't you make it bigger on the iPad like maybe more grids but a bigger puzzle or in theory yeah like the way the game works adding even one row would would exponentially increase the amount of possibilities which is a great thing and I'm at some point I may consider making a version that has more rows but yeah it completely like the amount of possibilities with the 7x7 grid right now there are over 40,000 player made puzzles and there's still like things

that surprise me every day so that it's pretty crazy how much possibility there is there really I I was wondering what your total sort of development time was and if you're comfortable was what is cost it's a bit tricky to say what the development I for control because the physics engine and the rope engine were developed as a part of a previous project but approximately it's around maybe four or five months I would say but that was only the only the initial game you know since it was released like one a half year ago we've been constantly working on it so it you can say that like four months and one half year and as for the budget working on cut the rope was we pretty much didn't have any budget it was just a bunch of guys working from home but then I'm just doing what they like I think that's all right all right thank you

[Applause]