Into the Breach Design Postmortem

I am Matthew Davis I'm one half of the team subset games I'm going to talk to you about how we made into the breach on subset games I am primarily the programmer well Justin ma who's sitting up here is the artist we work together pretty much full-time just the two of us and on the majority of the projects and then we end up doing and hiring out contractors and very talented people that fill in the skill gaps that we don't have to flush out the game stuff like music and sound effects especially Justin helped me on this whole talk so even though I'm the one giving it a lot of the ideas and the concepts it kind of represents our collective memory of how we made the game and how we make games in general as I am a programmer I'm not used to public speaking so forgive me if it's a little rough I spent the last six years in my pajamas programming in my room not standing in front of a lot of people our

first game was FTL it was a hobby project that took off a lot more than we ever expected it was made in just 18 months which in hindsight was spectacular at the time we thought 18 months was a totally reasonable amount of time to put into again and then we did the Advanced Edition after that and we moved onto into the breach with the security that FTL provided FTL why not I mean into the breach went on to take just about four years and this talk is going to go through why it took four years the mistakes were made and the and the good things we did but ultimately that whole process of of constant iteration cuts that led us from that original prototype on the left to the final game that hopefully you've all played on the right if you haven't played it it's a micro tactics game that relies heavily on the mechanic they're all enemy attacks at Telegraph so you see what they're going to do before they actually do it and so the game generates

these micro puzzles we have to respond to enemy actions before they've committed to them before actually talking about ended basically I'm going to introduce a concept that kind of runs through the thread of how we make games and how we made into the breach which is designing with constraints Justin and I are varied game design gameplay design sorry gameplay design focused everything else kind of has to follow the game design the narrative sound all the other elements ago in the game development end up kind of being a slave to the game design and then stepping one step further these usually these key designs that we hold constant the key mechanics and pillars that make everything else work and it's a process of following those key designs to create the game so it like the shape shorter does key designs make up that framework and everything else from sound to game design has to slot into the shapes that

are determined by those key constants the fun part about that in the end is you get to follow our design instead of creating a design Justin and I are terrible about big picture see the whole game what we're going to make and go make it we have to kind of discover it in iterate and prototype and fumble away toward the final game and the constraints are key part of being able to maintain that and and and make sure you're paying attention to what's important to the design as a really trivial example if you've decided to stick with a action RPG genre constraint and you move on and you doing the very classic mechanic of health that health mechanic is going to constrain a lot of the different mechanics that the rest of the game is going to be built around you're going to be stuck with probably somewhere to heal them health potions and healing spells in the link and you're gonna have to deal with what happens about health ones out where you die is the game over etc and there's a less trivial example if you were back to not using health and

there was no health and no damage then what would that do to the game how does it change how does that entire structure of how the game works would completely be turned on its head and so chasing down those constraints finding the interesting constraints and seeing where they lead you is how Justin I like to make games I'm going to start by talking about how we found those constraints to the beach the important ones that led to the game that we ended up making even outside of the specific game we are kind of stuck with our own personal constraints as a studio there are practical constraints and that there are just two of us for most of the project and then a few more at the end of the project it means we're not going to be making grand designs of mmo's and big 3d RPGs we're going to be making smaller experiences inherently our design solutions have to be smaller but it is a benefit in that we're just two of us we can iterate really well we can cut design without affecting hundreds of

other team members that have been working for years on something we can just say oh well there's just two of us and move on to something else we don't have to update anybody else we can just push ahead on top of our team there's just some subjective constraints that we have ways that we approach design aesthetics we like things that we like to see in our games a lot of these are gonna look really familiar in the sense of obviously game designers like these things but I think everyone has a different take on what it means to them what it means to them readability is important to us and for that for us kind of specifically refers to limited menus kind of that single screen combat UI like you see an FTL most of the game happens just like that screen shot you're not digging through really any other visuals we also like very low numbers and almost board game like design with low health low damage where one point of damage is meaningful and then making that two points of damage is a huge change being able to make those

really big changes but it's small increments it's fun or looking at the shield bubbles we don't have in FDL you don't have percentages you have just this really visceral bubble system that you can just glance and see the accounts that I like tokens and a board game we like streamlined design where you're looking to minimize wasted time and in hopefully making shorter experiences we're not looking to make at least in these first games that we made really big 40-hour games we were hoping to make something you could sit down and playing at lunchtime get a lot of design and compact gameplay in like a good punchy 20 minutes and be able to put it down as opposed to you need to commit your afternoon away to enjoy the game interesting choices is another one that it sounds ridiculously obvious we're all out there making interesting choices hopefully in our games but I think that it's something we forget quite often like we get so bogged down in the designing systems that if we don't stop to say wait have we actually made an

interesting choice is this huge mess of systems actually creating anything worth thinking about or am I just doing busy work when I'm working are you through them beyond our own personal constraints and studio constraints we start out with what the game is to begin with and the easy one that we all start with is genre that provides all of those early constraints and all those early designs you can hang everything else off of the trouble is you don't want to get too stuck on it and you want to find we want to find something that is slightly fresher in it if we can within to the bridge we knew quite early that we wanted to experiment return based grid based gameplay I was playing a lot of hop light which is what you see upon the screen there it's a brilliant elegant clean roguelike experience on mobile that if you haven't played you should Justin I have always liked tactical turn-based games things like Final Fantasy in advance or the big touch tones especially for Justin and we just thought it would be an interesting design spice to explore we wanted to

reduce random chance in the game with pretty good coming off of FTL FTL was complimented and criticized for the high degrees of randomness in the game as it responds to that not so much to fix it but as just a what if we did something different and we wanted to play with the game that would have fewer dive holes if you were just random slot machine it ish mechanics other than John were a lot of the thematic trappings are quite important to us we approach games thinking how do we want the player to feel what are the basic experience you're gonna have and building the key game designs around that concept friend of the breach collateral damage was kind of that core beginning thing there were a lot of movies at the time that had a lot of damage and didn't really seem to care about it and we wanted to have that feeling that when you're committing to these massive explosions and massive damage they also committing to whatever you're doing to the land around it and so the mechs vs. monster theme kind of fell into that quite naturally because when you have giant

monsters and giant robots fighting you're naturally going to have that collateral damage so we start out with a prototype I mean in that last screen in the terms of the genre and the theme was basically all we have to go on when we start we don't really dig into huge design documents as I had said earlier so we made something that is a relatively standard tactics formula it's not really shaking up the Johno too much it's just playing within but people have already done before we you can see that this here is a telegraphed attack which we did have in the game from the beginning though a lot of the other units did not have telegraphed attacks they just did standard on their turn they moved and hey you we were already playing the building damage and that we had this casual encounter but it wasn't an end-state or fail state it wasn't really anything that important it was just a score or something to remind the player that they're hurting people when they went buildings near damaged and we stuck

with relatively familiar weapon designs even from the prototype you could recognize here the laser beam weapon that does end up in the final game I find it shocking when look at the prototype that their ideas like that we stuck with because it seems like at this point hey that's most of the game why did this take 5 years but from this point and this kind of standard tactics point we explored a bunch of different ideas before we ended up settling on that telegraphed attacks being the primary one one thing we played with was time in a turn-based game so inserting real time elements where the enemy would have a countdown and every time the timer which which zero like you've corruption were spread it's some kind of mechanic would increase their power and the the obvious attempt at doing like a chest timer in a game like this they constantly created that pressure it wasn't bad it just wasn't what we're going for we wanted to make a turn-based game and have that kind of really careful calculated play and blue weighing your

decisions and that man contrary to that we played a lot would turn order not just turn order in terms of how and when people acted but also what you acted and how you're acted on your turn the traditional tactics seem usually ones would be be moving your attack and that's what a unit gets to do and so we had a lot of stuff is even if you can see it with this is what we're talking about and this is a functioning prototype a lot of these were functioning games we spent a year doing functioning games like this that we would throw out because we didn't like so here you can see that there's this strange concept where you can reposition or move your entire squad for an action point you could attack with your match for an action point or you could move some tanks and TAC with them for an action point even explaining it is hard and that's a really good example of why it was terrible and on top of that we had this complex time line system where like when enemies would act and when players attacked and

get a free movement it was just it's a mess it's that didn't work we added a time line system where the more you acted the further back your units would go on a time line so like if you shot and you moved then you'd end up at the bottom of the time line you wouldn't get act again for a long time or if you just moved the new public attacked quite quickly the ultimate result of that wasn't that interesting it was still essentially you had to move and shoot because in order to shoot you're gonna have to position and so mostly you move the same amount down the time line and it ended up with just a turn-based system wrapped up in a really complicated veneer and that was something that you've only run into a lot when it comes to experimenting with these designs is you get a lot of really complex rules but not particularly interesting or deep rules it goes back to that wait am i doing anything interesting here am i creating interesting choices or am I just making systems for the sake of it and that's what most of this stuff was but the fun part about doing this and the in the quote wasted time on these

prototypes within the genre was that we kind of get to prove to herself why the genres standards exist there's a lot of times that you might want to put money into her game / turn-based systems into a game you just throw stuff in a game because we've all been doing it and we love the games to do it and so we're pretty sure we want to in our game too but when you experiment with changing them you get to learn why they're there you get a better understanding for why they work and you get a better understanding for how we can change them how we can turn them on the side to do more interesting things with them so from this point we focused on the telegraphed attacks and by telegraphed attacks I mean that all enemy attacks have shown that there's no miss chance and there's no hit chance so those telegraphed attacks are reliable the information is always going to be accurate and the end result hopefully was that the player turn was completely deterministic there would have to be random chance in the sense

that when an enemy acts it randomly chooses its position and its target but the hope that we had was once the player was doing their thing they could completely rely on the information and what their units can do to come up with the solution and enact that solution reliably and so from there the question is how does that guy at the design so if we're gonna run with this concept of telegraphed attacks it's going to and we did not realize that actually at the time end up shaping the entire game from that point we thought it would be maybe a minor shift in the system but it kind of breaks every other tactic same system that we were used to and we had to find through mostly trial and error how it broke all those systems and so I'm gonna run down kind of some of the more interesting in the major points and how it changed to come back from that point first as a as a generality if you look at a tactics game a lot of it deals with managing threat you have these attack zone x' that the enemy in gate that the enemy has and that the player units have

and it's about staying out of their attack zones and putting them in your attack zones they tend to be in this shot from Advance Wars quite large that you're not going to actually be able to avoid all of the threat you're going to end up having to deal with hiding in trees for cover to reduce damage dealing with random dice fools that buries the damage a lot of ways for that threat to get mitigated in some way XCOM uses cover very large and big you Isleta zones but a robust cover resistant to reduce the chance of getting hit telegraphed attacks kind of change all of that because the telegraph system means that the that threatens own shrinks to something incredibly small it becomes trivial to move a Mac out of the way managing the threat no longer becomes part of the game and since 90% of the tactics aim was about managing the threat then that kind of breaks the game and so you kind of you you could give up at that point and I think there were times that we did because we didn't

see immediate solutions but you can kind of dig it down and see how can you make sure they can still be functioning threat in a game with telegraphed attacks the solution that if you've played the game you know we've ended up on is that the buildings are where the primary problem lies and that the enemies have to destroy the buildings to win and the building's if they get destroyed is how you lose and in hindsight that seems ridiculously obvious those buildings can't run away so just targeted buildings it seems extra obvious because we wanted to target buildings anyway we wanted that to be part of the game so it seems like I should've just slid into that point but it's another good example of how we get stuck in shanwa and expectations for how game design works on that we still wanted the mechs to be the primary thing in the game to be the thing you wanted to protect and so natural even now when watch players play the game it's so natural to see that you want to defend your units the whole every other game before is don't let your things get hurt

and kill the enemy units well this needed us to be no no your units aren't that important it's the buildings that are important but to start with we did these designs where if the buildings got damaged it would reduce a shield at the new mexico vulnerable and so even the day of the game would still get pushed back to the mechs being the meaningful element in the game but that never really worked it always fell apart for the same reason and that once you hit that point it was trivial again it was only interesting when you're worrying about the buildings getting damaged so you follow the design down from Telegraph to tax you end up with this defensive gameplay that's defending these immovable mod objects which again has these knock-on effects in strange ways for how that affects what I'm going to go into here which is the fail state and the wind state from the wind state point of view we saw quite quickly that killing enemies just fundamentally wasn't as fun as manipulating them when you have a system with these telegraphed attacks it's really satisfying to move their telegraphed attacks the teleport spells

and the grappling teleport skills and the grappling weapons and all the various racing push the enemies were so much more fun and they made the player feel really smart when they're engaging in that and so instead of running with the traditional just kill all the enemies we wanted a way that we could enhance that that nice part of the game that worked better than others we arrived at the turn limit concept where it was just X turns until the battle ended it's quite simple we've all done it in a thousand different ways in terms of defend the hack 20 minutes or defend this zone it's it's the defensive gameplay kinda lends itself to that concept it can be miserable if it's done and quite frustrating so we wanted to try to avoid that if we could but the immediate initial end result was great and that we could make interesting non-lethal weapons if it no longer became kill the enemy it just speaking survived a couple turns then you could do anything you wanted with the enemy and have lots of fun with it and it didn't matter if anybody died the bonus benefits is the streamlining

streamlining came in kind of naturally a lot of turn-based games can have dead turns a dead space like when you're chasing down the final enemies or when you're just finding the enemies to begin with while with a limited turn count we found that we could really tighten it up to be impactful in every turn that there was always interesting decisions to be made you never really had to burn time just looking for things and then as a result we discovered that micro battles are quite fun the end battle lengths friend of the breach is only four or five turns per per battle that was not something we set out to do if you told me make a game with four turn battles I would say that sounds really hard and I have no idea how to do that but with the telegraphed attacks and just kind of following it naturally we discovered that for term battles was really fun and it was a cool thing that I I mean I'm sure it exists elsewhere but I couldn't think of an immediate example for it and so it was a cool thing to just stumble across one ourselves other than the Wednesday to go back to

the failed state I've already covered that the mechs getting killed couldn't really be the fail state and so naturally fell on the building damage being the fail statement the trouble with that is that you can't say because ten buildings fell down the whole world is lost from a thematic point of view that just sounds ridiculous and so the big challenge comes and comes up a lot for us and how do you like the matically wrap that game design we followed the constraints until we get to a design that we think is the clear solution and since it's the most important thing the gameplay design is the most important thing to us we need to then come up retro actively with something to explain why and we explored a lot of stuff that's very traditional like morale and panic or energy levels none of them felt great even the last one that we ended up with I still don't think feels great I think it's the least bad option the narrative explanation of the buildings of powering the mechs and if the building to get destroyed your mechs don't work anymore it functions I don't even think we really explained it very well in the game we mostly run with

the idea of oh look the player sees it says power grid that sounds important and it works it's not brilliant you wouldn't write a book about it or a movie about it but it functions enough that the you can ignore it for the purpose of the game design functioning following that further that end state creates another problem on that you have these checkmate scenarios where the enemy is attacking too many buildings that you can't stop this example here these two buildings are going to get killed and that one man can't do anything about it so you've know you've lost in chess that would just be checkmate stop playing it would be very frustrating to throw up like a the the enemy vac saying checkmate you've lost like that's not how you want a big epic battle to win you want something a lot more thematic than that but having the player player a dead turn then hit in turn which might as well be a I lose button is really mean and that's a really unsatisfying way to lose the game as well and so we had to do

something strange in hindsight of inserting randomness into an otherwise deterministic design so we followed this design down that was guided by deterministic design but the kind of one of the final key solutions was oh wait we need a mischance we need this random chances still be in the game this way when the player hits and turn there's this grid defense that's a very low percentage chance that whenever the building gets damaged they won't actually get hurt and so there's always a chance no matter how dire the situation is the player can cross the fingers and hit and turn and maybe keep playing and that was an incredibly important part of the game we like to build out systems that are interconnected we don't like them just have them hanging there so we built a lot of things like pilot skills to increase good defense or ways to overpower the good defense and build a lot of structure around it but ultimately I feel like that's if you're building like a building and you need this gross ugly pillar in the middle of the room to keep you ceiling up and then

you hang some pictures on it so it's not that bad it doesn't help and ultimately be just kind of annoyed players they really hated the pilot good defense skill it had to be very low because we couldn't have them relying on it and it was just kind of drawing attention to an otherwise just necessary but not super great mechanic um in hindsight I probably would make it almost hidden something you wouldn't manipulate something that just always had a chance to do but you don't talk too much about the player you just they know it exists so telegraphed attacks led to the defensive gameplay which leads down to the turn timer Wednesday and the power grid fails stay and all the stuff that kind of spun out from that but I'm gonna sidestep and talk a little bit about the UI challenges that came along with the rest of it a lot of this is going to focus on you I guided design because you'll see that it's amazing how much the UI ended up influencing the design instead of the other way around we found

that we were going for a chest like interface which as you're looking at that board no assuming you play chess how the game works you know what the pieces are you know who can attack who where the threat lies you don't need to mouse over the pawn to double check what it does we wanted to have a feeling like that with into the breach we wanted you to see each puzzle and have it act like a little micro chess puzzle that you could plan out your moves without double-checking everything on the board the early iterations of that didn't have that it was a mess even looking at the screen now myself I don't really know what's attacking what this is an example of the game today without the icons that we ended up adding and it spent years to add the icons and figure out which ones we needed some of them were added as late as just like three months before release it was a case of they looked so painfully obvious in hindsight but it was putting it front playtesters have them be confused to get it back but in final play testers have them be confused

put it back and it's just at nauseam for years to sort it out these are the ones we ended up running with they're highlighted in yellow to make it a little bit clearer and the initial goal with these was just to show what was attacking what so you could see that this guy quickly was attacking this and this guy quickly was attacking next we thought that was all that really mattered is that the target for each enemy was what you needed to show but we learned from iteration experimenting more was that the attack type was also important to show that it was really meaningful if this firefight unit was punching this Mac or if that unit was shooting this match because when you move him it's going to dramatically affect the end result if this unit will punch this space or if his if he will shoot this space and damage the building having the UI need to show both target and type hugely constrained the possibilities because you've just made it exponentially more difficult to show the information you need and so that's

what we ended up saying we'd had to go back and instead of building a UI they could describe all of our weapons we had to start building weapons that could be described by the UI the new constraints that we started with or just three attack types so the same ones you saw on the last one that there are a killer attacks Malay attacks in projectile attacks they provided enough variety to have interesting ways to interact with them but will also restrained enough that you could quickly recognize them they use not the games they're not doing anything that crazy we did not originally have particularly limited attack zones you had this situation where you have no idea what this is doing where this is coming from it is these units here but there's no way of knowing that their threat range I think was the traditional like a radius around an artillery unit which was hard to predict and see on the screen we also experimented even with higher weapons these strange attack ranges you can see this is almost like a four-leaf clover

and the hope was that you would have almost like a tetris light game where you could layer these shapes and kind of figure out the optimal placement against the enemy shapes in a very traditional tactics method of you know managing those threads owns but it meant always checking threat ranges and then every time you moved your unit you weren't sure where that unit was going to be a ball attack it meant looking at an enemy unit you couldn't tell which enemy was attacking where and so we shred all that down to just being orthogonal attacks just shooting up down left or right this had the benefit of four player attacks as long as you can move to this base in if you knew what places you could move to then you knew where you could shoot because it was very easy to tell you could shoot up down left or right from where your destination and similarly all of the enemies lined up with their attacks so it made it much easier at a glance to see what enemy is attacking where and the last major constraint that there were many many more was these

static enemy designs this is kind of running with the almost chest like interface in that once the player has learned the units the units are reliably the same Firefly unit is always a fire fighting unit it has the same health the same weapon in the same movement that created a huge design challenge later on when we started worrying about difficulty in balance when it comes to balancing the difficulty it's super helpful if we could have made micro changes if we could have said this one extra unit does one actor point of damage and that way you could shift the game slightly without breaking it because in other which really runs into a problem that if it's slightly harder it could be nearly unsolvable but since we needed everything to be static we didn't have that option we did run with this concept of purple units which are significant power upgrades to the existing ones they have they across the board more health more damage and more movement but consistently so and they behave very similarly so this Firefly unit still shoots out a

projectile just like this one we explored more changes to that like the original design was this unit would also push its target while the basic version of it did not but that shift was really hard to remember and keep track of you had you thought you had the firefight down you understood how to stop a fire fire unit and then when they get upgraded suddenly blocking a pushing attack is hugely different than blocking a static attack because they might push him back into a building and that kind of swapped to how you deal with the units was too difficult so we didn't cut that it had a very strange knock-on effect that you'll notice that there's very few game enemies in the game with push attacks there's only I think just the beetle maybe one other and it was kind of a direct result of this in that we didn't want push attacks on the early units because we wanted the game to be more accessible and functional in those early islands and easier and push attacks made it really complex but then

we couldn't change those units in the later portions because we had to keep that static enemy design and so by accident we ended up with that mechanic almost being cut from the game and we didn't really even realize that into quite late in the process of that that's what had happened which is kind of sad but it was worth it in the sense of that consistency a game played really made the enemies manageable you can see all these combined resulted in massive cuts to weapon design Justin had made these amazing upgrade trees where this is a unit and never actually existed in the final version but the huge spread of how weapons might change and get managed in the shapes they might make and when we actually realized all these UI constraints we had to ditch all this stuff which was a shame but it was for the best and that that tighter more elegant weapon design really benefited all the other constraints that were in the game so you can see how we're growing this web off of telegraphed

attacks and while it's also influenced by a lot of those core personal constraints that we had they talked about the beginning it is ultimately back to following telegraphed attacks it got us to everywhere that we were and this was a multi-year process taking all the way up to the end of development in terms of sorting this stuff out I wish it was as easy as just oh there's obvious solutions once you pay attention to this but for us it wasn't quite that simple and as a better example if I wasn't quite that simple I'm going to move and talk about the strategy layer design which is a big shift away from the combat for a minute as I've been rambling a lot about the combat here and for the strategy design what I mean by that because you think the whole game is strategy so what am I talking about this all the stuff that happens in between the combat the that big XCOM map that you're used to seeing that gives you an excuse to tie the missions together and has lots of other stuff on top of that we love Dax come and let the games of its ilk and we wanted to have

that mechanic in the game we wanted to have you care about your pilots and manage squads and repair them and general resources and deal with the upgrade trees and research and all that stuff because all that stuff is super duper fun and we thought of course you don't want this it's worth noting that we were simultaneously designing the strategy layer and the combat in the way this talk is structured it might seem like we're like oh great combats complete let's figure out the rest of the game but it was a lot more in parallel to that and so a lot of the constraints weren't quite figured out and we probably and we suffered a lot more under strategy side for that reason we started with a very large-scale world with lots of nodes and you'd have your little mechs quads multiple squads were managing running around the world and chasing down enemy units it was very bored gaming it was bad abstract we even had a cart system where he draw cards to get events and an insert technology into it a lot of deck

building rooms were all the rage and so we were trying to put that into the game or you can see here that this is a corruption mechanic that like the enemy could could like explode into neighboring territories that was me playing far too much pandemic during development we've I fall into that all the time of wait what did I enjoy yesterday okay I'm gonna put this in the game it was dull to chase the enemies around chasing squads and pawns on the board just wasn't fun you didn't actually care about these abstract nodes it didn't satisfy what we're looking for at all and ultimately and this wasn't necessarily negative inherently it was going to be a huge game this had to end up feeling very ex-commies Civ in that it was going to take you 20 hours something to do that was we could tell a head of time that that's what the design constraints would lead us and that wasn't really what we're looking to make and so we despite spending I don't remember exactly how long we're probably six months at least

maybe a year we cut all of this and tried something else we went for something smaller something that you could maybe care about where you'd have these four different maps that you kept returning to and you'd be repairing the buildings and managing and kind of learning to love your little home here the end result wasn't great as well oh it's worth noting as well that there are two games here it looks like one in the sense that we've got these form spires but this site has a lot more of this energy management where you're upgrading squads and repairing squads and having a research tree and it's a lot more juggling with the resources well this side ends up a lot more X commie with events that come up in certain time scales with the whole time system and you have a units in and out of battle but it's just the core units with a few extras and big text events kind of like FTL ultimately both of these were very

complex but not very it goes back to the I'm building out this huge system but I'm gonna stop and say I match making interesting decisions in this huge system the research system was completely failing it was failing for the reasons that I've talked about and that the weapon design really couldn't support that research system but we didn't know that at the time we were blindly charging head because we thought we liked to operate systems be suffered as did the other one as do most strategy layers from ambiguous fail state it's not necessarily a problem a lot of games do have it and it functions civilization I think has quite an ambiguous failed to state and that you don't I don't normally play all the way up until I've lost my last city like I see the writing on the wall and I say oh I'm just gonna scrap this and go I'm just a new game and we would rather the player played all the way up to the end that they always thought they had a chance and there was really clear like a health bar

which is what we ran with that can show you where your fail state lies and what's gonna happen and text events didn't work in this game I think this is more our narrative failings as writers and it is that the inherent mechanics didn't work since just didn't I work in isolation to prototype the stuff we didn't work yet with Chris Avalon who did the brilliant writing in the game at the end and we had a lot of troubles writing the text events for this game FTL was so easy to build out a world that you go to new systems and me aliens and there's politics and you talk to people and those pirates and Space Station's and just an incredible diverse set of crap to pull from well this was a single city with a couple mechs and it means that you couldn't talk to because they're just mindless beasts and the narrative potential for us as amateur writers was not there and so we really couldn't make anything interesting out of it and so at this point we're years in the development and we're incredibly despondent and frustrated that none of

this is working that we honestly don't think that there's going to be a game here as we're developing this in simultaneous with the combat even the combat while it was looking interesting it looked like it had more potential it wasn't great and the strategy layer was terrible and so that was really between the two you weren't looking at a great game and so we really did consider quitting and dropping it and starting something new multiple times in development we took the breaks to make all the prototypes just to kind of get away from it and clear our brains because it was miserable in hindsight I think that all of the strategy layer stuff was failing because we were ignoring the constraints that were imposed by the combat I've introduced most of them already and that the telegraphed attacks with the limited weapon types meant that you couldn't really upgrade anything all those awesome research trees couldn't be a part of the game and that was an inherent part of the combat that we were refusing to acknowledge because we desperately wanted upgrade trees to be in the game that building defense focus meant the unit's were not as important anymore and so building out the squads and

caring and what we did have that pilot system in the final version we wanted to be a lot more complex and having to devote resources to repairing mechs and devote time to repairing mechs but that and making you care about the buildings more didn't mesh and we didn't really acknowledge that in these early versions either so instead of actually completely giving up on the game we completely gave up on the scheduler and we moved on to this version which was fundamentally cut everything except for what were time missions together so you just would pick from a list of missions and that was the whole strategy layer you had we just wanted you basically to go from a complex system with tons of choices to a very simple system with one choice and that one choice would have clear rewards and you would make one interesting choice and then start your next battle and that could work and the amazing part was it did work um the the combat at

this point was functioning enough that this could sustain the game in a way that although the strategy letters weren't sustaining it so it was fun for the first time we sent it to people and they're like oh this is interesting this actually maybe you have another fun game here and this was years down the line we haven't told anyone we're making it and we still not gonna wash out and tell anyone like publicly we were making it but it was really encouraging to finally have this it wasn't it what we're necessarily making looking to make it wasn't what we set out to make but it was an acceptance again of that least bad option the acceptance that this is what benefits the design and it also put us back into a position that Justin are most comfortable as a headset at the beginning we're terrible at that huge big picture design those huge strategy layers you had to kind of build all the pieces at the same time put them all at the game at the same time and hope it worked this meant something incredibly small that we could then add them top of that we could iterate on and add small pieces and kind of figure it out as we went

which we were much better at doing and so from there we could build it up into something you'd recognize today the first thing that we wanted to do was to keep that map visual I think having a actual physical place is very important it helps you kind of survive in the world and exist in the world and care about the people who are trying to help and so we quickly put a map back in the game though it is ultimately just a thin veneer for a mission list you can like on the final version or close to the final version it's just three missions here that you could select from and that's no different than the list is just on a map now instead of one a list we wanted to make sure we could keep the choices interesting because as I had said we basically wanted to cut the game down to just that one interesting choice between missions and so the question became how do you make that one choice interesting we experimented a lot with different rewards we had this asset system which both of these versions of the game had where battles would give

you these buildings and those buildings would affect the combat in different ways like increasing your mech health increasing your reactor cores that sort of thing it didn't work for a number of reasons one of which was it was really hard to balance to make it so one of them wasn't the obvious choice we hate it when you're presented with three things and one of them is ten times better than the rest and so you just always pick that because that's not an interesting choice and creating assets that would blend with the combat which at this point we understood was very tightly balanced and couldn't really take a lot of poking before it fell over and that's what these assets did these building effects would just topple the game if you poked it too much we also needed to worry about how to paste the game which is an excellent place to find ourselves because that meant we were finally thinking oh wait this is a functioning game how can we less out how long is going to be and we need to worry about how hard to make the game same idea so this stuff kind of lays down to our final touches and I'm going to talk a little bit about these final touches in the last couple

slides here before closing out the talk the resource design was we moved it down to some very simplistic that those early assets were so complex that we need is something that you could more quickly get a grasp on so we moved it down to just three resources just your power your upgrade cores and basically money the difficulty could be shown by the resource count so just three resources would mean hard and one resource would mean easy and the map design gave you an option to plan ahead and all this together combined into that brilliant single screen UI that we would love to chase down that we love to get into the game it kind of mirrors the combat in that sense that you can without checking tooltips of whatever we wore does you can see instantly what your choices are and you can make a reasonably informed decision taking into account what you need right now and what you need in the future it's not complicated it's not

fancy but it's totally functional which is all we really needed from the strategy layer at this point from a game length perspective like we did with FTL our approach tends to be just where's the upgrade curve going to lead us how quickly do we want to do all-out upgrades and what is the final point we want to get to it upgrades but with into the breach we found that there were multiple options and Justin I didn't really necessarily agree on which one was the best option I liked the two island game lengths it meant that mechs were going to be upgraded a little bit you'd get one or two secondary weapons you weren't going to be able to cap out your upgrades you're gonna play the game in just like an hour hour and a half so it was a lot more of a single sitting game and I really enjoyed that tighter smaller experience Justin really enjoyed v4 Island more epic experience where every Mac gets a secondary weapon which ups the complexity a ton when you have

a lot more options per turn of what you could do and it really adds a lot of overhead to the game which is cool and Justin loves it and chill on players level as well but it wasn't necessary what I was going for and so with the sense that we didn't really have a constraint where it led us to a specific answer we tried something weird which was it to do this dynamic difficulty length dynamic not difficult to life that length of the game so the idea being that once you've finished two islands you could choose to end the game at any point I'm Justin and I intended this decision to be purely player choice it was supposed to be oh crap I've got to go eat dinner I'm gonna end the game or oh I'm bored of this I'm not having gonna sleep a lot of fun so I'm gonna end the game or wow I love this squad my weapons are so much fun I never want this to end so I'm going to do four islands but when you bury that choice in the middle of a strategy game like this it gets a few skated and the player doesn't expect it to be just a personal choice you got I see a lot on the forums on the subreddit and stuff of people talking about Wednesday ideal time to end the game when should you

attack the boss when is it harder when is it easier when where's this strategy to this decision because every other decision involve strategy but since we scale the boss to always be the same difficulty regardless of when you end the game that was never part of it so just ended up kind of confusing a lot of people for some people they really enjoyed it but there was enough confusion and enough oddity to the design that I'm not 100% sure I would chase it down again as a great way to do it but it was a fun experiment and it was something that was cool to put in the game when it comes to difficulty I could probably ramble for like a half hour about into the Reacher difficulty probably longer but I don't have a ton of time so I can I have to condense this down into what I think the most interesting parts about it are inherently into the breach is a strategy game or a puzzle game that's masquerading as a strategy game it creates these little micro puzzles that you have to solve and the result of that is that too hard becomes unsolvable

instead of just hard you end up with a situation that as soon as you have six monsters attacking six locations and you've only got three mechs then it's not that it's hard as that you just can't stop and if every time that was happening you do not feel like it's hard you're just feeling like this game is unfair and it doesn't really leave a lot of wiggle room for what you can do and it really was a threshold that was just very just tight that as soon as we spawned one too many enemy the whole game broke and so we really had to build up a lot of does the balanced designs around that and how we spawned enemies and how we design maps to try to reduce that and to make it as functional as it always good so the game would always be solvable and the difficulty ended up being very much constrained by the design we did not say this is how hard we went into the bridge to be into the breach said this is how hard I can be and we found the limits on what we felt were the limits that the game could do in terms of pushing it to

an upper limit of difficulty and finding a lower limit that still kept the game being interesting with interesting choices even for lower skill levels on top of this in difficulty and the breach was weird because players disagreed hugely on if the game was hard or not we had early steam reviews that were back to bat this game was way too hard this game was way too easy it just clicked for people in a way they didn't click father's when you're asking is like when we were doing the early play tests and people didn't want to get their units hurt they didn't want to defend the buildings they're so used to their units being the most important thing that a lot of the decisions were asking players to make I'm not necessarily intuitive and what they're used to doing and maybe we did a bad job teaching that and explaining it ahead of time so they would be able to click with them faster but for some players it took a lot longer to click and so the game was a lot harder for them I think the end result friend of the breach is a good chunk easier than FTL but the really cool thing is that I think that it's still super fun and if

you told me 10 years ago that an easy game could be super fun my naive early designer self woulda said that's silly I love heart games but I think the puzzle game design inherently benefits kind of easy play it's really satisfying to do a crossword puzzle or any sort of puzzle that you're good at and to find solutions and they're just kind of mellow out and solve the game and kind of enjoy yourself in a way that you couldn't do it FTL and I think FTL as a contrast when it was too easy it was dull because you didn't let say you just felt like you were being pander to or not making interesting choices anymore and so it's just another great example for how these games define themselves more than it's anything that we're actually doing and it's just a case that we need to chase that down so on a whole hand we've been better at a lot of the ideas that I'm espousing here when it comes to like following designs and letting the designs constrain the game we probably wouldn't have taken four years to make this game I think that that was not an

inherent necessity it was just us being naive in our approaches some of the the key things in could've done better is we could have been faster at cutting bad design and realizing faster the cutting design was really fun I love the feeling of getting rid of bad design now I love the the freedom it gives you when you finally get rid of that horrible thing that you've been stressing about for the last year it's frustrating to lose a year of work but it's so freeing to move on to the next thing you think maybe it'll be better this time we need to realize sooner that we're stuck in a genre it happened with the focusing on buildings instead of mechs it happened in the strategy layer constantly that we're saying we want this in the game we'd like this mechanics if we're gonna keep you here gosh darn it and this is exactly what we need but the game was saying no no no that's terrible this doesn't work and we weren't seeing that fast enough we needed to follow those natural design constraints we needed to see what the game was imposing on us instead of us trying to impose nonsense on the game and ultimately the

importance of that there can be mediocre design in the game we went into the breach was so much time and the luxury to really explore design after FTL that we were overly perfectionist in our approach that we needed every mechanic and every element to be that perfectly balanced elegant thing and that doesn't work or from our experience maybe better designers could do it but we found that as I've covered a couple of them here there are cases where the designer needs to this ugly wart that sits on the side and that's okay this tragedy layer was not really one in I still don't think the strategy layer is brilliant but it does a completely sufficient job tying together the combat which is interesting and letting the combat thrive and find its place and really dominate the play experience is fine but in strategy lar just get out of its way do the necessary job but mostly not annoying you is all it needed to do and that's what we ended up with thank you

very much for listening [Applause] I hope some of it is helpful I know that not everyone has the same approach to things but I think that our journey maybe could I would hope provide something useful to everyone if you have questions you're good to go I've played about 80 hours of this game it's awesome thank you I had to take it off my computer I have to do work sometimes one question when you are at that last bar of grid power and you're about to lose the game it does seem like the building's resist a lot more than the alleged percentage it does not no it's not no if we have we have personally from a design point of view I don't like changing the game I don't

like doing those the mechanics I like the game to be a static experience a lot of the design and inspiration for me comes from like board games and the like which are inherently very static experiences that don't accommodate to the player and I like making design just from a personal standpoint of we don't accommodate to the player we don't try to artificially create those exciting saves not that there's anything wrong with that it's just from our approach that's what we do the very few times we do anything that accommodates the player and into the breach is spawn mates and enemies if there are too few enemies and we will bump the spawn and if there are too many enemies we drop this bun because I talked about difficulty it makes a huge difference to how functional the game is and how many enemies on the board and so that's one of the only moments where we check the player progress or check the board state and change the game based on that so is that based on there are going to be a certain number of opponents great dog an amazing game thank you so there in the

game you have this chance where you there's a debris that drops and you have a random number of turns where you can come and go and pick it up what was the reason to add that feature in the game did the time pod that shoots down and you collect that yes um I'd have troubles remember the exact set of things that led us there I'm part of it was too manic we liked this idea that we're running with this time-travel mechanic but people were going back in time and you're doing this loop and all the pilots were supposed to be time travelers so we need to some way to automatically introduce time travelers into the game and it just felt really epic and cool to have this pod flying down and crashing you know the earth and then on top of that it fit the other constraints quite nicely in that it follows the core design of defendant things it added something else to the board to defend and so it fit the mechanics along with the theme in a way that would felt quite pleasant I think we had the time talk pretty early and we liked it from the beginning and we didn't really change all the way through hey thanks for the game thanks for the

talk thank you so you mentioned that things like the grappling hook and that in the teleportation spell were very fun for players and make you feel smart and all that kind of stuff I'm wondering if you ever experimented with having a starting mechs be more like the steel judoka where where it's like they're their primary attack is more of a movement and enemy manipulation attack than being because the initial set is more like direct damage and the enemy movement is kind of secondary based on the early experiments the if anything that first starting squad was still too complex for a lot of players we especially their artillery weapon which would you need to shoot next to an enemy to push it instead of shoot at the enemy and all games have always trained us to shoot at the enemy and so putting people in front of that and saying do something like indirect fire do something like manipulating them was a lot to ask and so we were focused more on how do we just introduce something like a push mechanic in a very natural way so they

could get comfortable with it and even then I'll see steel judo Dooku toku have a huge divide and how much players like it in the there are players perhaps at the game click more that they really enjoy because they get to screw around with his mechanics more and then they're a player so how on earth do you do this and so it was important especially for the starting squad to have something that everybody could try that makes sense I love this deal Jack thank you thank you for the talk when you early on decided how to move your units and you said you you were inspired by self did you consider hexagon Oh No I restarted with a grid from the very beginning and genuinely my only reason for that was that the first prototype that you saw in the beginning was like a city and the idea was that there were city blocks and it was a grid based system like a city and that was where I was coming from when I put that together and I know that's an arbitrary thing you

know certainly silly and we and I think I guess once you sort of programming it changing it to hacks have been a huge change the fundamentally changed all the design have you done up at that point and so even though we ran away from cities and ended up on larger fields the grid did the square grid still felt more natural I think it's also great for the pushing mechanics though is that it's a lot more intuitive to push and be aligned on a grid system but that's a happy accident not like a intentional design thank you for the talk and congratulations on winning last night thank you I was wondering if you could speak a bit about how the telegraphed attacks influences your enemy AI design and like how the AI behaved um the AI is very simple where it's just every enemy acts entirely independently there's no planning between them to make sure they're covering a certain amount of buildings or units every turn they just look at every where they can attack they take the best options and the second-best options and they randomized between them and it's as simple as that

and do you what about the spawn points of the new enemies because that's completely randomized within a spawn zone Wow we we always I always as a solar programmer because the whole game was programmed by just myself and and a small design team I always picked the absolute simplest implementation first so I do something just stupid simple to put it in the game and then if stupid simple works I say great and move on you have a few different like map types and you have spa like mapped out for them yeah all the maps are actually pre-made yeah it doesn't look like that and a lot of players don't realize that because of the simplicity of making maps it takes two seconds to make it mummy not two seconds Justin did miss the work so for me it was super easy but they are all pre-made and we have like 200 or something in the game and they had to be it pre-made made it a lot easier because we could control stuff like bad map could really influence the

difficulty in ways that were horrible and so we needed a controller tightly and again it goes back to stupid simple and that it was easier to make the maps rather than me design some algorithm to generate the maps this will be the last question but if you want to do more questions I can go to the wrap-up room is there specific room I should go to thank you go for it great game great talk that appears about your process like you have multiple iterations of mechanics multiple implementation so what how do you get to the point where you suddenly realize it needs to change is it play testing is your mindset is self-critique how do you approach that it depends on the situation and there was a lot of play testing when it came to like teaching the game and seeing what mechanics would work based on this was raw usability a lot of the earlier decisions is Justin my play testing a lot on our own it's one of the reasons we like the cedral gameplay and that we

can surprise ourselves and we test the game against ourselves quite easily and there comes down to just a lot of intuition and we can tell this mechanics work in this mechanic we try to balance against it and I'm sorry I don't have a great answer for it but it's it's with a lot goes in the right yeah thank you [Applause] [Music] [Applause]