Rogue Legacy Design Postmortem: Budget Development

[Music] hey everyone thanks for coming to our talk today my name is Teddy Lee and I'm the game designer for cellar door games and I'm Kenny Lee the programmer if you guys don't know we're brothers so we're here today to give a post-mortem on rogue legacy was an emphasis on budget design for those who don't know rogue legacy is a casual man's roguelike even though many people still find it very difficult it's far more accessible because we added permanency to the game so in this game every time you die all the money you found you get to spend to improve your family and I say family because with each death you never play the same hero twice instead your children succeed you to start we consider ourselves to be a budget oriented team that means we focus more on gameplay rather than trying to make expensive highly detailed worlds that

also means we're always on the lookout for solutions that are cheap fast and reusable so we often implement solutions that aren't as good if the cost savings are justified that also means we'll never make the perfect game but we're okay with that so we've been doing this for a long time now a lot of people think rogue legacy was our first foray into game development but we've actually made eight games before this on average each of these games cost us about a thousand dollars to make with development time ranging from two days to three and a half months and it's through this experience that we are really able to solidify how we go about making games so you might recognize some of the games that are available in flash and mobile so this is our development timeline for rogue legacy as you can see it starts out with me working on the project solo and as time went by more and more people joined the project for rogue legacy it took us

a year and a half to make the game and I was the only person who worked on it full-time from start to finish Teddy actually came on full-time five months before release and everyone else was contracted so this is rogue legacy by the numbers as I mentioned in the previous slide it took us a year and a half to develop the game and in total our auto park costs for the project was less than $15,000 so that's still a lot of money compared to the thousand dollars for our previous projects but if you consider the scope of the game I think we did a pretty good job of keeping the costs down when the game went up for sale we were able to cover all of our development costs in less than an hour and within the first week we were able to sell over a hundred thousand copies of the game which is well beyond anything we had ever anticipated to show how seriously we take budget we need to talk about rogue legacy's beginnings rogue legacy was actually never meant to be a roguelike instead it was a matchup of three games you might be familiar

with it had the death system of dark souls the platforming of Castlevania and the cartography system of Eitri and Odyssey and it had virtually no roguelike elements in it it was a game of crazy scope that emphasized exploration and puzzle solving so this is Dark Souls 2 d the first build of our original non roguelike game and it took at this point about three months to make I'm sorry so as you can see in this game it's none of its procedural every every single rooms been handcrafted and all the artists it's just Castlevania like we were yeah we just were our inspirations on our sleeve but what you can't tell is that we only spent a week and a half actually making content

everything else was like building up the engine and everything and it took over five minutes to go from one end of the game to the other so it's pretty massive and we actually had already had eight enemies in the game and five different mini bosses so you can't see that very much in this video because it's all separated but there was quite a bit of content for just three months of work so what you can see here is actually the original fairy chess rooms which were rooms that you'd enter and have a very highly difficult challenge that you had to overcome to get a special item and for those who actually played rogue legacy this is one of the bosses which actually went from prototype to final with very little change in it this is just a platforming test room I don't I don't know why I added this yeah it was just a test like jump Heights and stuff in any case at this point we were feeling pretty good because it was already starting to feel like a game in fact it was even kind of fun to play so we did what any developer would do in

our position we killed the project at this point we were able to better evaluate the costs for the game and came to the failed conclusion that we simply could not afford to make it it was a tough pill to swallow but our experience working on all of our previous games made us fairly certain about those projections so from here we had to go into Salvage mode and tried to build a game from the scraps we had a basic 2d platforming engine an AI logic system and a level editor that could create rooms and that's when we decided to make rogue legacy procedurally generated rooms resolved our biggest bottleneck which was art while opening up the strength of our tools which is rapid level development it was also a natural extension from Dark Souls 2d since a lot of the appeal to Dark Souls is its semi roguelike concept of consequence upon death where when you die you drop your souls and you have an opportunity to pick them up again it's it's similar to roguelikes but it's not as punishing so

here's another video this is rogue legacy sorry this is rogue legacy after we cancelling the first project this version was created two and a half weeks later okay so to really signify new beginnings we removed all of the placeholder art and replaced it with my amazing programmer art so what you can't see here is in the previous video we said that the the world was stitched together in this one it was procedurally stitched together by the system so it's pretty much the same logic as the retail version of rogue legacy but you can't tell in the videos and for some reason we also spent the time implementing a lot of new enemy types I guess we didn't really have our priorities set straight and we also had things like new zones running in the game so I think yeah you

can see a new zone here with new enemies that have amazing AI stuck in stuck in a corner because buzzard drives us we try to salvage as much as we could and in the end we surprisingly managed to salvage everything from our previous project and this was about a year before the game launched we were able to launch so quickly despite the setbacks that we encountered because we embraced the new direction immediately and pulled the scope down with no regrets but the challenges didn't end there we like to put twist on the genres you work on so we decided to make a more casual roguelike experience this meant changing some of the most common tenants found in roguelike games so the first thing we wanted to fix was the harsh punishment

of deaths but for us we didn't want to just make death less painful we wanted to try and make it fun so that's when we decided to add permanency to the game by using the manor skill tree and the equipment system that way you never had to start from scratch it was a contentious decision because we knew we were dilating the spirit of roguelikes where usually when you die you kind of have to start all the way from the beginning even if your run was ten minutes or 50 minutes long we also wanted our game to be more skill-based and less chance so in a lot of roguelike games there's a big emphasis on the roll of the dice so you could have critical misses or instant death events and we wanted to remove as much of that as possible so we did a lot of tweaks to the game especially inside the castle generation system to make sure everything was fun and fair we still do have some random roles but for example incredible strikes but whenever it's just beneficial to the player we're okay with that so the other thing we wanted

to do was reduce obscurity roblox often go out of their way to obscure information so that players can kind of discover the world while some people enjoy that we want to make our game more transparent and easier to memorize so a lot of the pointless obfuscation stuff like 50 potions but your own labeled by color we wanted to remove all of those instances this is why we have descriptions for virtually everything in the game some of them are flat-out explained while others just provide hints so that the second time around you more or less know what it does fortunately we wanted to make the game faster and shorter so in most roguelikes the first few minutes of play are usually the same you need to start to power up before you begin to take on unique characteristics we wanted to fix that so we designed the lineage system to affront all change to the very start of the game this was important because we want to make our games to be very

easy to pick up and play and that meant we wanted to make the average life last around one minute so if you dial up playing this game don't worry it tells by design but these short lives also mean a lot of streamlining because we didn't want people to play for 40 seconds and then dare to spend like the next two minutes navigating menus again we had to just like streamline absolutely everything so an example of that can be seen as a trade names they're all purposely very distinct so you can recognize them at a glance and I'll have to read them every time it's kind of why we flip-flop between scientific terminology and slang so an example of that is a hyper gonad ism which if you have that trade I mean do you hit enemies really far back and we have the inverse straight which was hypogonadism so enemies wouldn't get knocked back at all but if you're playing the game and you'd see these trades you'd actually have to like stop and then read the word out entirely because they're so similar so in this example we turned hypogonadism into muscle weakness just so that it's visually distinct

the next point is that we wanted to remove floor cleaning from the game so a lot of roguelikes segments their games into floors where the further you go the harder it gets but if you bypass the floor without clearing it out entirely and these will start to out scale the player putting you at a disadvantage so this floor planning really slows the game down and limits players choice on how they want to progress his game since there's really only one true route that you can take and so we were able to resolve this by removing all forms of character development during a run inside the castle so this is our last point and it was actually the first decision we ever made for rogue legacy and the primary reason we were able to stay on budget so we call it the game loop and explains how players go through one life as this term might actually mean something in game design but I don't know we've never like read any

books or anything on this type of stuff um so it looks like a fairly standard loop except the RPG mechanics happen after death and before you explore the castle instead of during play like most games it seems like a minor change however almost all the previous design decisions that we spoke of were built to support this so by putting RPG mechanics after death you had something to look forward to which made it fun and because no character development happened during gameplay we could lessen downtime and resolve the issues related to floor clearing and best of all it gives you that one more time feel this is what we had eight months later so some of you might recognize it it's actually very similar to our final build so with the title screen we had like a start menu and a load menu and we actually took that out for the final version because like we spoke earlier we're so focused on streamlining that we took out any sort of step that we thought might impede players into getting to the game

as fast as possible this was actually our IGF 2013 build we're not particularly happy with it because it's missing a ton of the features that the final game does have but I think you can see generally it's it's much closer to what rogue legacy is now than what we had in the previous build with my amazing art so some of the art you might see is a little bit strange because this is not final art and we actually touched upon that in a later slide and we do have all the different areas working in this build we actually had six areas we in the end decided to cut it down to four because it made more sense with the top bottom and right areas and along with the center and the left being the entrance I love this part this is our previous death system so whenever you die is really cool a whole bunch of enemies and coins would start raining down from

depending on your performance the reason why we couldn't keep it though is because people who had some really really like epic quests that the enemy would just fall forever and if we added more enemies dropping to the screen we found that it would start to affect performance so we had to simplify things to what it is today which is just enemies appearing on the screen after you kill them so this is pretty much the retail version you see now with a bit of older art and as I mentioned earlier a lot of the features missing even though things like cherán traits and even the lineage system weren't in we were fairly confident that we would be able to complete the game on time honestly a video the final version probably would have made more sense for this slide but we thought this would be more interesting because you can just see the final game on like YouTube or something we've been pretty strict with the budget

to every game we've made rogue legacy is no different but we did come into it with a fair bit more experience here are a few key budget tips that we try to use on every game that we make autonomy work with people who can be independent this is important for a small indie team because everyone wears multiple hats and it's tough to micromanage people and fix their mistakes you got to be able to do that on your own this happens sorry this has implications though that not every developer would agree with this means our development process is not very democratic opinions matter but when it comes down to it the most trust is placed on the experts and nothing is decided by vote a good example of this happened with our character artist glabra Pataki we originally had a separate character and background artist and I was stuck in the middle trying to get their styles to work together you can see from this picture how much the background on floor ground clashed with the character art things were clearly

not working and thankfully glabra decided to step up and fix things for us so that was the previous slide was what we had before and this is the final product what's amazing is that glabra had very limited experience on backgrounds before he'd worked on them but he'd never worked on something of this scale budget-wise it's crucial to work with these kinds of autonomous people for robe legacy we didn't catch on fast enough and it cost us dearly before glabra fixed things 20% of the entire projects costs were spent on this problem and we didn't even get it resolved we went through many many different artists the second one we follow is to make custom editor for each game so try to make the editor for the game not the company this isn't for everyone but for us in the size of our projects it always turns out cheaper and faster so this is a picture of the editor that

we use to make the rooms in robe legacy it seems intimidating but it's actually quite simple the map editor can only do one thing make rooms because it's so rigid it only took us two weeks to make there were tweaks that came afterward but the bulk of the work was within that time frame what's really great though is that we could create in bug test the room and as little as five minutes there will be much that we can salvage for this for our next project but that's fine because the time that we saved creating contents more than makes up for the time spent on the code our last point is to always have design alternatives this ties in with what we mentioned at the beginning of the talk where we always look out for cheap fast and reusable solutions it's usually necessary to come up with multiple ideas to find the one that best suits the resources you have when everything is categorized based off cost it's a lot easier to gauge how far you can go with

an idea for example the original lineage screen was going to show your entire ancestral family tree so you could track spouses children follow branch offshoots and so on it was really expensive and we ditched it early for the lineage line that only shows your preferred children and we lost some of the ornamental features with this change but the idea of an expanding legacy was retained and we were able to do it for a fraction of the cost so another quick example of design alternatives is the skill tree economy system the choice we made for this with very contentious and Hillis traits how a budget can play a big role in the decisions we make in this case we actually chose a bad solution because it's so much cheaper so the original skill tree economy is fairly straightforward every time you spend money to upgrade a skill that particular skill price would raise it seems simple but there's a hidden problem to it there's no way to save money in the game because cheiron's always taking it at the beginning of

every run so optimal skill bills always meant putting an equal number of points into each skill so you'd put like ten points into house and then it get too expensive and you start putting ten points into damage and so forth in other words there'd be no skill diversity we want to encourage players that personalize their bills which meant there had to be consequences of player choice but harsh consequences were counterintuitive to the casual scene we were going for so the repercussions had to be very light so in the end we decide to use a universal modifier buying a skill increase the cost of all other skills by a very small amount as well as the skill you purchased by a large amount what this did is it deterred players from evenly leveling up their characters throughout the game resulting in more unique skill builds this is just one example even though the solution worked it was terrible for a lot of reasons first it feels bad hiring having everything go up in price when you purchase something and salary items get feel more expensive

it's not a good thing to do and secondly it punishes casual players who poop on blind lee upgrade their skills because they'll start reaching that point where things become really expensive really quickly and they don't have an optimized character for it and thirdly and probably the most the biggest problem was the negative effects of the fix aren't noticeable so people would see the solution that we did and did but they wouldn't know why we were doing it so they'd go on to forums or they go on to send us email asking us just to remove it and it's it's such a like a knob tooth problem that explaining it isn't very easy but on the other hand the pros were too much to pass up so this is what we like to call a zero dollar solution there was no art or audio that was needed and implementing it took literally less than 10 minutes and because it's so self-contained almost no time was spent balancing and there was no QA required either so our game is full of these kind of

workarounds some are better than others but they're a big reason why rogue legacy had such a small budget this example sort of encompasses our design mantra of good not perfect so where do we go from here we're now financially stable and a lot of these budget workarounds might not be necessary anymore so the question is would we still do things this way and the answer that we most likely will there's more to it than just saving money finding simple solutions to big problems is useful in any scenario and we often surprise ourselves as those things we come up with for us it's also the most entertaining part of development because that's when it's the most challenging anyways that's it thanks to everyone for attending this talk and we hope it was a insightful [Applause]