[2018 TGDF] Justin Ma ─ Design Lessons from FTL and Into the Breach transcript

now Mortensen launching your chin touch adjusting my weight I don't eat just a potato you easily interpret she got so cheating in time now oh my god Joe so high huh hi everybody so my name is Justin MA I created the company subset games with Matthew Davis we've done two games so far as she said FTL the top one and into the breech the lower I did the majority of the game art Matt did all the coding and we work together on design everything else was done by our super talented contractors so FTL started out as a fun little side project and it eventually gained tons of notoriety worldwide something that we weren't really expecting so it's a captain it's a spaceship simulation game we tried to make you feel like you're the captain giving orders to your crew it took us about two years to complete and we continued to work on it after we released it for about two years after the success of FTL we took a little we took our time with into the breach and it was a very long process which I'll go into shortly so this talk is mostly about into the breach what is it it is a tactics game so picture of Final Fantasy Tactics or even chess but the big difference is that you know exactly what the enemy is going to do before they do it so in this gift for example you can see that the top enemy oh I have this that guy he's going to attack this building here with his red line showing you doing one damage and he's going to attack this building there so this is what happens in this gift is is if you do nothing on your turn they are going to execute their actions and attack the buildings so what do you do you get to act before they act so in this example my punchy mech here guy he punches this monster which moves him over one tile and now he's attacking his own ally so in one action I was able to prevent two attacks going to the same building and the enemies the enemies are not smart there's a really terrible AI and your goal is basically to use their stupidness against themselves and so what I mean by no outcome randomness is that there's no hit chances there's no critical chances there's no missing if you know if you think it's what's shown is going to happen 100% and so that's what no outcome randomness means so other than the combat it is a very simplistic series of short battles one after another with very little downtime in it you can between battles you can upgrade your mechs improve your pilots over time and get new equipment that sort of thing and then each game that you play ends up being very different there is a sort of inspiration from rogue likes to have procedural generation and some permadeath also the story revolves around time travel so at the end of the game you send one of your pilots back in time to start over in a new version of that world whether or not you win or lose someone goes back in time hopefully so to understand some of the choices that we made during development its kind of useful to know our mindset going in we had just finished a very successful game we had a ton of fans and had enough money to fund our next game without any worry we're a very small team we don't own a studio it's very cheap but we're not very business minded so we didn't want to just grow and hire people we didn't want to work on FTL - we were kind of just exhausted and we wanted to remember why making games is fun why are we doing this so when we started this project we had a few rules in mind we didn't want to deal with the pressure of making something bigger and better than FTL so from the start I just tried to accept that this game would not be as big it's just that what's going to happen we tried to focus on what Matt and I were good at so this is why we didn't use unity this is why it's in 2d pixel art it was largely just us trying to work with our skills and we were also committed to find a game that we really liked if we hit a dead end and we couldn't find something fun we would just throw out the project this is why we didn't announce the game or tell anyone about it for three years or so it was because we just didn't want to have expectations or promises that people had and we wanted to feel like we could throw it out if it doesn't work and of course we also did not want to kill ourselves working on the game we crunched very hard with FTO and we wanted to avoid that we wanted to take weekends off to see our family eat food you know so into the breach we started with a very simple basic design we knew that was going to be a tactics game that had giant mechs fighting giant monsters after seeing movies like Superman we were tired of media where the whole city would be destroyed and no one really cared it was just fluff it to make it seem exciting we wanted a game where you felt like it was important to protect people and very basically we wanted to have tough decisions sometimes it was worth letting your people die to save a building and sometimes it may be worth destroying a building to save the world so these were the core principles that stuck throughout the entirety of development but you may notice there isn't really a game here there's just a couple ideas so what is the rest of the game is it a huge campaign is it really short does is it a ton of mechs or is it just one mech we spent over two years about two and a half years trying to figure out what the game would be so introductions into the breach this talk is going to be two sections the first section is going to be focusing on how into the breach evolved over the time and then the second part will be about sort of lessons that we learned and maybe tips that you guys can take from that it's mostly going to focus about design so this is a very simple timeline of how the design of the game evolved throughout time the details they're not really important it's just something to give you a sense of how we tried something it didn't work we tried something else it didn't work and as you can see there was maybe there's multiple times that we would spend six months on something find out it doesn't work and then just cut it entirely so this part of the talk I'm mostly focusing on the first two and a half years of the four-year development where Matt and I with some few key testers we sort of figured out what we wanted the game to be and it was the after this period you know we had to actually finish the game which was also hard but there was a lot less sort of dramatic changes in the design during that last year so so I'm gonna go over each of the main design shifts but of course each one could probably be a talk on them all night on its own so it's just gonna be like a brief summary so the earliest version of the game was a simple turn-based battle on a grid it had sort of fancy ten like turn-based system each unit would go in there order the actually actions just happened immediately it wasn't very interesting or unique the first big turning point this is what the actual game looked like the left one is what like a concept art I guess so the big turning point was when we realized that telegraphed attacks were fun and what I'm Stella graphed attacks I mean is being able to see what the enemy intends to do or plans to do before that they get to do it so this was the first enemy that had it the two's there are what it's going to be attacking and once we had him in the game and we started realizing was fun we added it to every enemy and the game got way more interesting from there we switched to a sort of basic isometric view and then after that when we were realizing this could be an actual game it became clear how hard the user interface is going to be we had a game where you have to know everything that's going to happen at all times you have to know what the players movement ranges what your tax due which enemy is doing which attack everything had to be perfectly clear just at a glance and so that necessity of a clear UI shaped a lot of the gameplay and and force the gameplay to be small and simple so the the fact that every more most every weapon can only fire in like a straight line that's because of the UI it would be too hard to explain things and I could probably give a whole talk about the UI because I was just miserable but the rest but I'll just ignore that for the rest of this talk it's fine so at this point we knew that we'd like the combat system enough to actually make a game out of it but big question still remained what is the rest of the game how do you progress how do you win or lose head what do you do in between battles that was a very long gradual process of trial and error the first pass at the meta game which is the things in between battles the first pass was this very large map that had a whole bunch of small nodes these are just mock-ups this isn't the real game but so they're all these little nodes and stuff each one of those would represent a different small city so in this version you would manage many different squads trying to hold back the enemy as they sort of spread around the large map each of these circles like I said or is a different small map and the enemies could take them over and you could fight them in it or take it back or repair and stuff like that we wanted to have a interesting strategy layer sort of like XCOM where you are managing your income from the cities and researching new technology and all that sort of fun stuff so when we tried try doing it we ran into a whole ton of problems because the map was so big and the cities you see them so infrequently there was too many cities to care about you actually stopped the the destruction of one area didn't feel impactful at all even you stopped caring basically it also just simply wasn't fun to chase enemies around the map that just was boring there weren't there weren't really interesting choices you just take a step they would take a step is boring we knew that for the player to get invested and actually care about the upgrades that you would actually have to be a pretty long game in general and we knew that we want it to be shorter than FTL instead of you know a 10-hour campaign or something so we knew we had to change something so from this point we switched we tried to keep the same idea but go on a smaller scale threats so in this in this version threats would appear in different regions and you basically would decide whether or not you wanted to go protect a region or save your energy and that sort of thing also there was no walking you would just instantly teleport to a new area to fight it so we wanted like I said with the previous version we wanted you to have a stronger connection with the city enemies we'd take over an area but because there's fewer of them you'd sort of recognize it and and realize you know how it's changing over time we also knew that we wanted to have FTL sort of FTL like events where there be the texts environmental descriptions of things that are happening and you could send your units on missions and stuff like that and similar we still had this whole system where you were managing mmm ton of different systems income population the building's repairing Tech Research all the sort of things and different squads so the left image here was the first pass we switched to a smaller eight region area but even that we realized that you just wouldn't recognize the places because there's still too many so then we shifted to the right image where we basically forced you to see the whole map at all times and we actually worked on that version for a very long time until we decided it wasn't very fun so this was getting better it was more interesting but there was still a number of problems it was overly complicated but there was actually very few interesting decisions it honestly it wasn't certain if what you were doing was a good distraction when your actions actually made a benefit the the result of your action could take an hour before you see what happens and that caused a problem that you could see in a game like XCOM the original hives come where the enemy gets harder and harder and if you don't keep up with the enemy you'll actually gets in a position where you've already lost the game you just don't know it and you'll discover that you lost the game five hours later and that is just did not a fun situation so we tried to avoid that as much as possible so we ended up sticking with this design for a while but eventually we decided it just wasn't working we were basically making two boring games instead of one interesting game so this was the big turning point in the game's design we spent basically a year or so unsuccessfully trying to make the meta game strategy fun but we we knew that the combat was interesting so we basically decided to turn the game entirely into combat just throw out anything that is not the combat so at this point we knew that we wanted to focus entirely on battles so we tried just making it a series of fights in a row like I said anything that didn't fit with this new idea of of what the game should be we just got rid of it no matter how much we cared but I wanted to be simpler but we still wanted there to be interesting choices that have costs and benefits and so we wanted there to be something that you still can feel like you're making decisions in this strategy later so we were on the right direction but there were still a lot of question how do we make the choices between missions interesting that's it was a hard tenth hard thing to figure out and how do you make it balanced so that these choices there wasn't obviously one that was better than another and then the general have you paste the game how many battles do you have when do you win how often can you repair and that sort of thing so in the left example we have you basically pick one mission out of it they you gives you two random missions and you picked one of the two and then it's just a series of that in a row later we decided that we would try and bring it back to a map even if we kept it simple and so we'd give you those two options from a list of a you know ten or whatever and in those each one of these locations ends up being useful assets the building so for example if you rescue this area it would give all of your mechs plus one movement and if you were able to successfully protect it you would lose the other one so there is this sort of decision-making of what skills do I want to have going forward based on which areas that I pick so um around this point for other reasons we switch the game to occurring on these small islands this for thematic and story reasons each island had its own terrain and features and once you save one island you would just go to another one before whatever the end would be so the island Islands were divided into different regions each has their own random mission and when you completed one the adjacent regions the night the region's next to it would become available with new missions so this helped the the design in a number of ways having random mission objectives mean that each time you went to the island it felt very different you can also see the potential rewards of omission of the glance that was highly important to us to make it very clear in both of these versions even though we're quite different what the rewards are you could see what it is from the island view and there was still short-term decision-making do I pick this one or this one and there was also long-term decision-making of if I choose this harder mission I could potentially get to a better mission later and further on in the island it wasn't very complex but it at least was a decision so we were generally happy with this design but there were some problems that and challenges to get it to be right similarly how do you balance the difficulty and rewards so even in the last left one you can see the sort of an asset idea gaining plus one reactor for all mechs eventually we decided that that was too hard to actually balance some things were just clearly better than others so he actually made all the rewards very very simplistic and frankly more boring but it made for a much more balanced game so in the right one which is the closer to the final game you could see it's just like you can gain power or money or upgrade and that's it there's not nothing else but there was still a decision-making in there so it was good enough eventually so he eventually found a design that we liked but this whole process of this meta game thing was very long and very hard and we had to very often give up on something that we tried very hard to make interesting and we thought was a really good idea and we just have to get rid of it we of course had that option because we our first game had enough you know made enough money that we could just keep working this way that we were very fortunate in that way but that didn't make the decision-making any easier so this is the end of the evolution of the game's meta game design but I still want to talk to about a few more features that were particularly interesting nope I forgot how to finish the island without doing all missions basically the players felt like they were missing out if they couldn't go to each of the regions and we had to figure out some way to do that so like I mentioned before we spent a lot of time trying to make FTL events work so these are a text events something pops up something happened refugees are coming that sort of thing and you have a couple different options and the options would change based on the equipment's and the pilots that you have this general design it helps add a lot of atmosphere and character by only adding text it's not too much work to do it they also add randomness in a way that you can sort of strategize around it's not like a 63% chance of something it's you know it's a 50/50 or something more obvious that you know that what the risks are they also allowed us to have pilots and equipment to interact with the game outside of combat we thought just only interacting inside combat would be kind of boring so we wanted like pilots to be able to have personality traits that would maybe you know if they're heroic or something that would change the or they're very aggressive that would change the options that you have but unlike FTL it we're in FTL you jump to a new location and there's a new event it just made sense you were discovering a new location but we ran into some problems with into the breach how do we even trigger them there's no clear time marker where when you should get a new events we would just make it happen once a week but that just felt very random and because they popped up randomly it just sort of felt like it was jarring and it didn't fit with the game we thought we would help them create more storytelling but it just never felt right and because we weren't certain what it was adding to the game it would just be a ton of work for something that we weren't certain about so similar to the text events we were also trying for a while on the right this sort of deck based system because we love board games and we love the randomness of board games it's a it's a random that you can sort of work around over time and understand and so we tried to have something like it a car a deck of events in the game but we scrapped it and we in fact we scrapped all text events entirely so another design that feature that stayed in the game for a very long time was the concept of support units these were weaker more traditional military units that could be killed and then replaced later in part they were there to make the mechs feel more powerful we wanted to convey that this was a really new technology that came from the future so how do you do that you compare it directly to something much weaker and I personally wanted them to have really interesting upgrades you can see one of the mock-ups is like a turret and how I overtime you'd improve it and it just has tons of options and you can specialize it to be more interesting so there was a couple core problems with having support units and all this technology and stuff in a very simple game balancing the abilities of the support units compared to the mechs was very hard they needed to be useful but not too useful the mechs needed to feel really powerful in comparison but what does that mean well how can how can the support units still be useful it also meant that we had way more equipment and complicated systems in general it was just a lot of stuff to have to manage and in the end we found it wasn't really necessary early on when we had the city view we would have the support units like located stationed in different areas and then they would stay there permanently but once we switched to the sort of battle system of the simple combat it the the support units didn't really have a role though their role was filled by the mechs why even keep it so besides this there was just tons of things you were managing the assets the support units we had convoy units and and defenses like a turret or something and all in all it was just way too complicated and we just cut everything just focused a hundred percent on the mechs you may be noticing a theme here of just cutting everything and smaller and smaller so the next issue that we struggled with was thematic wrapping basically what was the story justification for the gameplay it it didn't have to make sense but we needed to make the world feel like it fit like there was a thematic core of the game so we knew that you needed to buy things and upgrade your mech but we didn't know how who were you buying it from why did where did they get the technology we also knew that the destruction of the city would cause you to lose the game but we had no idea why we had to prove that in the game somehow and along this path we tried tons of things we had morale and panic and resolve the civilians of a city would evacuate and run away if it got too destroyed and things like that so this was just a lot of small trial and error things and it took a long time to find something that we liked so we initially had something close to X comms research where you you it takes a lot of time to get something new and that sort of thing and it just was way too complicated for our silly little game we tried like I said panic him around all these things and no matter what it just felt tacked on and arbitrary and we knew that we had to connect the mechs like function there there the fact that they work to the buildings but so that became the core focus of the thematic wrapping in the end we basically tried to find something that was as simple as possible and just so that the player would accept it and not think about it so we cut research upgrades were just all unified into one thing and the we said that the mechs were connected to the city's power grid so if the buildings were destroyed the mechs no longer had power and then you lose the game it's not really an elegant solution but it just it was good enough and have provided the sort of function that we were hoping for so this two or three year journey took us all over the place how did we survive this time basically we tried to stick to our original pitch as much as possible but we would then follow anything that seemed to be the most fun this style of development is probably only possible in very small teams you probably couldn't done as easily cut 20 peoples worth of you know a year's worth of work and half everyone wanna continue to work for you but since it was just mad and I it was easier to just just cut everything and just not worry about it so like I said we tried our best to make something and basically every feature work but in the end nothing was safe from criticism we try to keep an open mind to to be willing to change anything in the game this was very helpful to make it better but was also very stressful because you just felt like nothing was ever finished and you felt like everything could be thrown out tomorrow and of course the final game was way smaller than we pictured maybe like 10% of what we pictured originally we just cut and cut and cut and cut and sometimes it didn't feel great but sometimes it was wonderful because if you cut 30% of the game you are 30% closer to being done so this is a four cutting okay I didn't go into too much detail but this is a very general progression of the evolution of the game so now let's talk about how we made these designs decisions over time I tried to spend a while to think about how Matt and I approach game design and then I tried to take those ideas and make some sort of tips for you guys very often these things are just the way that we work naturally they're not rules that we made up and then followed it's just a reflection of how we work in the that we like so I'm not saying this is perfect for everybody but maybe you could find it useful so first I need to talk about how I personally think about design when I think about a feature I usually break it into two aspects when I'm calling low level and high level design other people may have different words for this or think of them as the same thing but I find it useful to separate them so what I'm calling low level design is the basic details yes Mario can jump but how high if you tap the button versus hold the button does he go higher or not it's essentially what are the mechanics and what are the actual numbers high level design on the other hand is more abstract why does the feature exist what role does it play Mario's jumping isn't just the numbers it's a movement mechanic attack how you activate buttons and switches it's also the reason why the game is fun to begin with so a features high-level design if it doesn't result in a better game I think it's just the feature is completely pointless if you just have if it doesn't actually add something to the players experience that's positive so the basic difference is essentially one describes what the feature is while the other describes why it exists or how it interacts with the player so I'll go through two examples first pac-man or mrs. pac-man and I assume you know it you connect you collect dots and then you avoid the ghosts and every once in a while a cherry appears so let's think about this cherry for a moment so what's the low level design it appears once you get a number of points it appears at a specific place if you get extra points if you get it and the ghosts don't care about it and it disappears after a short period of time it's fairly basic understanding so what is the high level design what is the role that it plays at least what do I think it is it creates a risk versus reward scenario the cherry disappears after a short time so is it worth trying to go for it to get the points or do you just ignore it and try and survive it also shakes you up during the middle of a stage you basically have to reconsider your strategy you may have had a plan for this whole time and now something else appears and you have to think what you know what do I do now and when that happens you can often make a mistake which is a fun thing for to happen it also increases the difference between good and bad players it increases the skill cap so good players can learn you know when they appear and they can use that knowledge to their advantage to get a better score second example is the ghost in spunky if you don't know this game it's a indie platformer it was one of the first like notable games that at least I know of that took the rogue a roguelike formula and took it to another genre the game is divided into a number of different maps and if you spend too much time in one map a ghost will appear and chase you what is the low level design it appears two and a half minutes it slowly moves around it moves through walls you can't kill it it kills you and it also has fun with gems so that's how it works but what role does the ghost play it encourages you to go fast and throw caution to the wind the game has tons of traps and if you have all the time in the world you could very easily ignore them and so just the threat of the ghost even if the ghost isn't there the fact that you know it exists makes you go faster and often make mistakes it gives you interesting choices you don't have time to go everywhere so you often have to pick do I go here here or do I go to the store or do I go to the treasure chest and that sort of thing it reduces the feeling of grinding grinding is when you just have to do everything and it's not fun so if you had no time constraint you would basically feel like you have to go get every little piece of treasure because it would help you improve your your chances of survival but that's not fun to do to feel like you have to do that so reduces that feeling and similarly it also increases the skill cap for players good players can travel really fast and get everything before the ghost arrives and you can even use the ghost to your advantage if you can you know go around and stuff like that so it's important to note how important the low-level design is to create the high level design that you want so if the ghost was a little bit slower it came at three or four minutes there would be no threat you'd finish every map and no issue if it was became quicker like a minute early you would basically have to run all the time so just by changing one small aspect of the low-level design you would have completely different high level design so that's some examples and knowing what the low level design is very easy but how do you figure out what the high level design is I don't really like to ask questions so just continuing with this Blanc II example first I ask how does it force the player to act it puts pressure on you to keep moving forwards there's a large area but you only have a limited amount of time what would it be like if there was no ghost you would feel completely free to take your time you'd explore everything enemies would feel less dangerous you have that grinding feeling how does it interact with other features so it forces you to prioritize one thing over another and also there's a ton of traps in the game so having more pressure works well with making mistakes and makes the game more fun generally so on paper had just having a ghost that chases you doesn't seem like it would be that important but if you understand the high level design you can see how it impacts and changes the player experience in many ways and maybe why it's necessary so if you think a feature in another game add something important to its design and you want to do the same thing to your game I suggest figuring out the high level design and then bringing that to your game instead of the low level design so for example if you wanted the ghosts and you wanted something to make the player move faster instead of copying the ghost exactly or something figure out something that has the same result as the ghost that says causes the same feeling for the player if also if you have this high level design as your goal it's actually a very useful metric or goalpost to compare with you can keep changing the low level design and basically ask the question is it giving us the high level design that we want sorry there's a fly I thought my face also if you copy like I said earlier if you copy the low level design it can actually result in a very very different high level design in a different game so this is generally why I suggest copying the high level design instead okay so in my three examples here are the hunger meter in roguelikes where if you don't eat you die that's pretty simple and there's very little food and you basically it forces you to move quickly the second one spunky ghost and third one is the fleet in FTL basically it was the threat of the bad guys who are chasing you which forces you to move to the right so these are very different mechanics but the high level design is basically the same thing if prevents the player from seeing the whole map it forces the player to have to make choices and maybe make mistakes and it gives you a sense of foreboding the longer you stay at you start to feel nervous so these are the same high level designs created by very different mechanics so I would say that spunky doesn't need the ghost but it does need something that fulfills this high level design so it's it's useful to keep that in mind when you're when you're trying to come up with a new mechanic to work for your game and of course each game has their own challenges because the low level design is so different like I said with the timing of the ghost each like the fleet it has this the timing or the size or the movement everything is has their own unique problems so as I as you saw in the first half of the talk we went through tons of design iterations figuring out what the high level design of a feature is is one of the primary ways that we would Madden me would decide whether or not we should keep a feature or not we would basically analyze every feature in the game and figure out what it was adding and then cut everything that wasn't helpful so for example the huge map in the left of the big city it seemed like it was a great idea but in actuality it was taking away from what we wanted the game to be similarly if we find that if we analyze every feature and we find that multiple designs have the same purpose or function then we would combine them into one just to make the game small easier to understand why have complexity when you don't need it so the middle example like we had all these different types of units but they basically perform the same role so we just simplify that into just equipment and also analyzing high-level design helps you be objective and it helps you have distance from something that you care about it's very hard to cut something that you spend six months on but if you can very logically and very cold and very coldly decide this is not working for your game you know it's helpful to sort of let go and move on so basically thinking about high level design is one way that we try and be objective along the same lines if you want your game to stand out it's also quite very useful to question everything but is in the game that they just assume is in this type of game so if you start figuring out the high level design of every small feature in the game sometimes you can come up with a new interesting way to do something that is normally done in the same way in every other game so ask yourself does your RPG really strength intelligence dexterity or whatever do may just have to wear cloth armor do you you know in shmups do you have to have lives and continues and that sort of thing if you analyze everything with the high level design lens you can actually try and figure out is this helping the game design or not and of course if you can use this to come up with a new way to do something similar it just makes your game more unique in braid for example I don't know how Jonathan Blow came up with the idea but he basically questioned the assumption that time moves forwards and if you start questioning even basic stuff like that you can come up with crazy stuff and interesting new mechanics when you take nothing for granted and anything could change but that being said most standard design is there for a reason there's nothing wrong with sticking with something that works you don't have to come up with new something new for everything usually there's a balance between giving the play or something that's new and giving them something that's unfamiliar if there's too much of one the game is not very fun so I just ask that you at least consider these features to see are these necessary are these helpful and if you in the end decide yes these are necessary and helpful it's fine don't worry about so this is very obvious perhaps but is super important try and spend your development time efficiency it just it's so much effort to make a game so you need to you need to save your time when you can it's very critical to know your own strengths and weaknesses if you're not familiar with network code and you can't hire someone else don't make a network multiplayer game just stick it's great to learn new programming languages but if your goal is to release a project on time it's better to just stick with things that you're strong with there are so many ways that you can improve games and it's really important to figure out what are the best ways to improve it how do I use my time effectively so with FTL we have this jump animation that we did in like minutes like early on in the game and I always wanted to change it and I just never and similarly there was like five planets and I we never went back and changed changed it because people just didn't seem to care they want what they cared about was more content and more interesting experiences with the game so I basically put all of our effort into things that we thought were wouldn't have the most impact in the game and of course you still have to ship the game in the end so at the end of the day sometimes it's better to have a smaller game so this tip is a little bit more situational and it's a little bit esoteric so when you have a choice between two game designs and both would work fine be pick the one that is nicest to the player if the if the game feels unfair the player because they think this other design we work just as well the game is just instantly unfair and you'll never change that person's mind and even if you give them all the reasons in the world it doesn't matter if the game is unfair it's a bad design thanks Sarah I think it's way better to get the player to blame themselves whenever they make a mistake if the player can blame the game for making a mistake it feels unfair but if they can only blame themselves then then I have succeeded I guess so in X come if you walk in front of an enemy and get shot it's your fault but if you fire and they have a 95% chance of hitting and you miss the game it's the game's fault it's not your fault and so you leave feeling worse it feels like the game made a mistake even if it didn't make a mistake but that being said the player is not always right very often things that are unfair are necessary to the game so these this is really just four things that it could go either way and then you're like I don't know which we should pick just pick the one that's nicest to the player so examples in the left there this is the last boss of FTL you could destroy the boss at the same time as it destroys you and if the boss shoots first for example we could just say that you lose the game but because it doesn't hurt the game at all it's fine yeah it's atmospherically Fitz if it's a suicide mission so just be nice hey you won the game they sacrificed themselves in the middle example there there was a mechanic where when you kill an enemy you can heal and there was some very weak enemies but that we didn't count as real enemies and we didn't let you heal but then people complained a ton about it and we realize just just let them heal it's fine it's the same thing and I mentioned this earlier but this is a bit of a different angle but so the 95% chance to hit and you miss that feels awful but if you have a 5% chance of happen if something good happens and it happens you feel great so in general if you can have randomness always be in the favor of the player have the expectation to be something that's bad is gonna happen and then when good things are happened they love randomness so that's something we try to do in our games that's why we have this this grid defense mechanic that that you see on the right there why the number is always so low you have to assume that it's never gonna happen and then when it happens you're related so these last two things are just my personal preferences first if you have like a strategy game and you want the player to understand what's going on and use strategy use low numbers whenever possible as humans we can only have like seven numbers in our head at once or something like that so if you want people to feel like they're making informed decisions like they really have control keep everything low it's just it's easier and I think that hiding the having high numbers is actually just hiding the strategy you're trying to make it seem interesting but the actual strategy could be simple so for example your attack may do 3,000 damage he has 10,000 health but that's essentially the same thing as three hit damage and ten health because all that matters is how many hits it could even be one damage in four hit points and that it's effectively the same thing it's Mario versus rabbits did something pretty elegant where they had a any major random event was 50% chance of happening which is a very simple way and that being said if you want your game to be flashy and you don't care about hi strategy or whatever then go nuts have billions of damages done there so my second suggestion is only if you don't have it like a narrative game or like a very story driven game I personally feel like your time is better spent crafting a believable universe than it is to add more text or have have more dialogue so for example the atmosphere in hyper light drifter or journey if you know those games were way more compelling to me than most games that have text even and to me that's entirely because they have a very interesting and compelling world even if they say nothing and you understand nothing just the fact that it feels real is more impactful and similarly Metroid Fusion if you remember it it has tons of text but I don't think it improved the game that much maybe like 5% in general it just like slow the game down a ton so obviously this isn't good if you are making a very narrative game in general but I also feel like if you could tell your story in 50% less words do 50% less words there's no need if you look at RPGs like Baldur's Gate compared to new games like pillars of eternity there's like 20 times as much text and I personally just just get me the game but that this is just my opinion so these different tips were all based on how we approach and discuss design and I wanted to give you one more example of design that we struggled with in into the breach and so I'll trying to show you how we would use these principles to learn from our own mistakes and try and make something better so in FTL we had a final boss it was okay some people loved it some people hated it if you don't know what it was it was a very challenging multi-part fight each of the phases there's three they would have different skills one had drones and when I boarding and such we wanted the boss to be really well-rounded and have no particular weaknesses it even had a few unique skills that no other enemy in the game had like you would just gain shields randomly and and so is something new you had to deal with that's the low level design so what is the high level design note this is this is my evaluation looking back years later this is not where we were aiming for this is just what happened it was a dramatic conclusion which is good it ended up having a huge difficulties back spike you could have a very easy game and then suddenly just get completely dumped on at the end good players would often have to upgrade their ship specifically trying to counter the boss you'd have to play throughout the whole game thinking about the final boss that's kind of awkward since only certain strategies would could beat every phase of the flagship instead of having more interesting strategies work we actually reduced the amount of strategies that worked because the boss was so well rounded you could not for example win with a boarding party only you know sending your crew onto their ship however it did cause huge positive moments for people who won but that's usually because they spent hundreds of hours in misery and unhappiness so when we approached the final boss design of into the breach we tried to keep in mind the things that worked and try and fix the things that we thought didn't so climactic challenge great the difficulty was way too hard so we just tried to make it reasonable at least and not just overwhelm you we didn't like forcing you to think about the final boss at the whole throughout the whole game we wanted you to basically be able to win with any type of strategy that works in normal game and we also wanted to avoid being surprised by like unique you mechanics right at the end of the game we wanted to not have to feel like you have to play one whole game just to learn something new to lose and try again later or have to look at the Wikipedia or something like that and this last change it's not intentional but it happened inevitably if we remove the challenge the crazy difficulty of the boss it's no longer super satisfying to defeat which is sad but oh well so it actually turned out to be quite hard to to find a design that fit those high-level goals we consider tons of different things a big scary monster that had tons of abilities a really long battle where eggs were hatching and there's tons of monsters or like a escape sequence stuff like that but if we had a single big boss a kill this unit and you win type of mission it kind of clashed with the core principle of the game you spend the entire game defending people and locations and suddenly we just ask you to kill something that's kind of awkward and doesn't fit with the with the whole rest of the game we wanted the boss that matched that feeling of protection since into the breach is solvable it's just inherently an easier game you have to be able to to win and because of that it is it for good players it's easier to do well and there's also a very thin line between something being hard and fun and just unpossible for a game like this if we just add one more enemy suddenly you can't prevent damage and you all 100% lose and so there's very little room to play with and in getting this balance of difficulty right similarly scaling the difficulty was very challenging we had easy normal hard we also let you get to the final boss at a different point three different points in the game so that would mean nine different bosses that we have to balance the difficulty of and so all of that made it difficult to find something that would work so in the end we ended up focusing on something that was similar to normal fights but a bit more challenging and higher tension so there was way more enemies than normal there was no giant new boss but there was also environmental like dangers like lava and stuff to help make it easier to kill all of the enemies your primary goal was to protect a bomb which would go off if you defended it for long enough this sort of defense equals victory just live long enough type of victory match the rest of the game and we also use the normal enemies a lot and this this means that the player could use their knowledge of the rest of the game to be able to defeat the final boss and it also meant that we could use the difficulty scaling of the enemies without any changes it just made remove tons and tons of work so this wasn't a perfect solution but it at least gave us the sort of high level design that we were hoping for so this process of coming up with something that would work and fit our high-level design took a long time to figure out but basically we tried to figure out what we wanted the boss to feel like and then and what the player to feel like while fighting it and then we tried to figure out mechanics that would fit that feeling so first we analyzed the high level design decided what we wanted to keep and what to change and then we would just try things trial and error does this work does this not work until we found something that fit our high-level design so that's all the tips I have for you very often we don't follow these rules they're not they're not set in stone these are just things that we've done in the past and every team may approach design differently but hopefully this gives you some insight into how we think about design how we solve problems thank you and thanks to Johnston for translating the slides and I'm sure it made it a lot easier to understand thank you for having me [Applause] no manifestations Justino - mother - spider challenge cheering - sound now shadow Yamin told her when she had her juice out normal microphone tikani now along it just a joke you a decision sound now yolo hello auntie ma first of all thank you very much for the talk and the question I have for you today doesn't really concern the talk directly but I mainly wanted to know that when you were walking on fto with your partner you walk in the team primarily of - but how did you manage to deal with situations that usually have a dedicated staff member to attend to with stuff like finances law and maybe PR and marketing yeah we just we just have to do those things basically at all times when you're working on something some problem can happen for example Windows 10 updated and it broke FTL and then we said okay I guess we're gonna make no progress on into the breach for two weeks you know so that stuff happens and it would be better if we hired someone but in general yeah we just have to do those things ourselves and we don't like that part of development I don't find that fun so that's why we sort of like do things like this and give talks and do the game and then hide in a cave and just work by ourselves and talk to nobody you know so this is Evan flow of business stuff and development stuff also you could always email me or something - if we don't run out of time if you have questions how do you know you're satisfied with the design and you can move on to the completion of the content so I get really invested like when I come up with an idea and I really like it I just think about it all day honestly most of the time it would be Matt who is just like one day in the morning he says I don't think this is working and then we just throw out everything and start over and so when he doesn't say that we just keep going and eventually though I think and when we talked about all the time and eventually you were just like well it isn't broken yet so I guess it works and so we just keep trying and they're like yeah it's still not broken so I guess this is what we're doing there was no it there's no no this is it guys we solved it never never does it feel like that all right I fought off another question I felt like asking which is when you were making into the breach you may take took extra careful care to make sure that the game was kept secret until you are comfortable before was announced but during development when you are trying to figure out what worked and what didn't was it just you and Matt who just tried and play it over and over again by ourselves and try to figure out something works or not all we were able to get second opinions from someone else so for some background with FTL we we worked on our own and then we did a Kickstarter and then suddenly we had 5,000 people and we had to give them updates and tell them things and that was kind of dramatic in traumatizing so with into the breach we we just we didn't want to do that so most of the time it was just Matt and I are partners and some close friends would try it most of our friends used up all of their goodwill on FTL and didn't want to help us anymore but there's a couple guys who helped we uh I also just show it to a couple people basically once a year or twice a year at GDC in a pax there's like a couple people that be like here play this and tell me what you think but other than that we it was only in the last year that we like how to discord group and had like 30 testers who would just try it every day I don't suggest people to do it this way we're very much like the way we like to design is I want to make a game that I like and the reason why that is because that's what that's how FTL was developed and we we sit figured that worked so let's just try that again let's just try and make something that we like and if we like it then somebody else will like it and we are very hard on ourselves we're very critical like Mike I said Matt like it's just it's very hard to impress Matt so so we don't rely that much on testing honestly we don't have any I don't even know the word telemetrics we don't want like get feedback from the game from people online and stuff like that yeah it's pretty simple you said you had difficulty levels in into the bridge well the first question is why and the second question is does it does it affect the AI or only like it points in damage and those kind of things so it it's very the second question is easier it's very simplistic honestly it doesn't affect AI it doesn't affect damage every unit does the exact same thing basically the only thing it affects is how many enemies there are because in this game if there's two enemies and you have three units it's very easy to prevent the damage if you have two you're three you know mechs and there's five enemies it's very hard to prevent the damage and just shifting one or two numbers it makes a huge differences why do we have the difficulty levels it's it's because that people want something different out of the game different people want different things some people just want to fly through it and feel powerful and some people want to bang their head against the wall and figure out you know like prove out beat you game I'll show you game and so and obviously with this game more than FTL like playing it on easy I think it's still fun I think even if you're better than the game it still feels it's just a very different experience it just feels less like um trying to show off and more like I just enjoy feeling you know a certain way in this game also in general the with this game in particular there is a huge breadth of skill levels in this game there are some people who just see it understand it play it on hard and they're fine and there's then those people who say I've spent 15 hours 20 hours playing on easy and I cannot get past the first battle and we generally don't like to force-feed you like tutorials and stuff and so we tried to make the the gap between the difficulties wide enough so that you can at least get something out of the game no matter where you are you know you can have your your kid play it on easy and like you know they do okay maybe I don't know and I you know I personally thought that's why we called the normal mode normal or whatever instead of medium is because it's it's like this especially with FTL it's like this is what I think the game is but I at least respect that other people have different opinions what they want to get out of it it's also a very low effort way if you have good difficulty scaling it's a very low effort way to appeal to more people it's like I just have to change a few numbers and suddenly all these other types of people like the game which is very convenient anything else okay back to just a question about cutting features attitude deal actually with fostering it was it always for the greater good like it's always cutting some things for making the games actually better or sometimes did you have to decide to cut things just because of the timing or anything like this if you wanted to really is it quicker and how did you deal actually with cutting things it was it always a good thing right away or was it on the long term that it felt a Terry right so the with the first question like do we always cut the game for the greater good no we cut for every reason possible like because we feel like it because it's any anything could convince us that cutting is a good idea very often it was because we didn't think it would add to the game more often it was because it would just take too much effort and we're like this is not worth the time sometimes it's because we want to ship earlier and other times it's because we're just sick of something which doesn't go away the second part was it was like hard and difficult and sometimes it was very emotionally challenging my wife tells me like you this whole talk is basically about non-attachment and if you know like Buddhist the theory essentially and how you basically don't get attached to your game design I think we are naturally more in that direction like I am so attached to things in real life but game design I'm like doesn't work throw it out I don't care so sometimes it's challenging to let go of something that I really like but usually I just file it away as I'll try it again in some other game which I know will never happen but it makes me feel a little better oh no but yeah it's it's always painful but usually in an hour I feel like oh thank god we did that like thank God I don't have to think about this anymore it's gone yeah believe there was one in the back one thing you mentioned is that you should follow the fun and would that be more difficult for you to follow the fun if FTL wasn't such a huge success oh absolutely yeah the only reason why we were able to do any of this was because we didn't have to worry about money for our next game we could take all the time we wanted we could you know the flow in between the development we basically spent a year just sort of messing around with little prototypes and just having fun as much as possible in between doing porting or whatever but um but ya know it that that was a luxury definitely it also is a big function of how small our team is like I don't think it would be possible to get a team of 200 people to be okay with changes every day and we was it was mostly because that Matt and I just by chance we are very similar thinkers we when we see a problem we generally come to the same solution so it just it makes it easy to just both agree yeah okay we should do something different and very rarely do we get in a situation where one of us says no we have to keep this and the other says no we have to get rid of it and that is just pure luck and everybody else who's involved in the game doesn't really get to decide it's just pretty much us to making those tough decisions obviously we care about things that other people invested their time in but in general we we don't even bring on other people until we're certain we will won't cut their stuff okay yeah so the last question I guess hi are you a fan of a vaca dia Evangelion yeah yeah thank you all right thanks guys I'll be outside for a while if you want to chat and come say hi no that's okay thank you [Applause]