The Spelunky HD Postmortem

[Music] [Applause] hi everyone oops hi everyone on it'll it'll turn onow oh okay there we go he's got you hi thanks for coming I am Derek and I made a little freeware PC game called spunky and it turned into a much bigger XA game by the same name um and when I think about that how that happened how I started with this tiny little game how it became something so much bigger uh I'm still pretty amazed uh it still feels pretty incredible um but in this talk we're going to try to make some sense of it so I started working on spalony pretty soon after I finished Aquaria with alak coloka um and that was a game that we released B on December

2007 and it was uh the project took us over 2 years uh and it was a really intense and difficult development as I think Alec would attest to also um it was about as stressful as it was wonderful uh so after that I really want to work on something a lot lighter something small and and uh kind of fluffy and it's kind of ironic because spun is now turned into a bit of a monster itself um but when I think about it uh with all of my big projects and those are Eternal daughter Aquarian spun X those games all took at least two years to make between all of those projects I I always made a lot of small games and released them and when I think about it those small games always led to the bigger ones uh whether indirectly or directly so for example with I'm okay working on that is how I met Alec and that eventually led to Aquaria and all the small games that I

made before Eternal daughter um you know they got progressively bigger and they gave me the confidence to do a project of that magnitude and Spelunky the freeware version of course led to Spelunky XA um so small games led to bigger games and spun itself I considered the freeware version I consider it to be really like a refined prototype and uh and that refined prototype came from number of even smaller prototypes that I never finished so when I approach prototyping I like to treat it like drawing um you know with drawing if I want to come up with an idea for something to draw I'll just pick up my Sketchbook and a pen and I'll just start doodling all over the page uh and I'm not concerned about trying to create something that's going to win igf or create something that's going to make a million dollars or even create something that is going to make people like me or that people are going to approve of uh the key I think is really just to Zone

in on what I'm enjoying at the time the things that I'm interested in uh and when it comes to making games tools game making programs like game maker for some people it's Unity for some people it's Flash and uh libraries and Frameworks like flash Punk and flixel those are like the equivalent of a sketchbook for me because it lets me get my ideas out very easily so this is uh these are screenshots from uh a Rog likee prototype that I started working on before spalon um and the goal for this prototype was really just to have fun you know I had a lot of fun making the graphics and I was trying to boil down the essence of Rog likes into something that was uh just simple and fun to play and if you're not familiar with Rogue likes they Dungeon Crawlers that have random level generation permanent death and a lot of variation and it's a really compelling genre

um because those three things really make the game very exciting and and infinitely replayable uh and even though it's a pretty Niche genre that it it's had a lot of influence on a lot of really popular games like Diablo Minecraft and uh Tow Jam and Earl I really love that game and those games are all based on Rogue legs and I think the reason why they're still Niche is because they haven't really advanced in terms of their interface beyond the original Rogue game uh which is just was just which was just asky Graphics basically on the other hand platform games is you know they're a really well understood and well-loved uh genre um they're action-based they require reflexes and precision timing and uh memorization of the levels and I think you know everyone is pretty intimately familiar with platform games um so here are a couple screenshots from just really simple

platform games that I was working on at the same time that I was making uh those Rog likee prototypes and if you look carefully you can kind of see some little elements that actually made their way into spunky freeware oops so you know working on Platformers and working on Rog likes at the same time these working in these small prototypes really didn't get very far uh I came up with the idea of combining the aspects of Rog likes that I really liked that I thought were really compelling the random level generation the permanent death the variation the fact that you get into these unscripted situations that are really fun to get yourself out of um I wanted to combine that with uh a platform game you know and I'm a big fan of that genre and as I said before it's a very accessible and well known genre um I want to make something that was uh new and exciting but also accessible and the Big Challenge I think between reconciling

those two genres was uh the level design because uh platform games really depend on their levels on their platforms and Rog likes are exciting because they're randomly generated so I thought I'd quickly just go through the algorithm that I use to generate levels in spuni so each level in spuni really starts with a grid of uh 16 rooms 4x4 uh just like this and then we draw a path from the entrance at the top of the level to an exit at the bottom of the level you're always going down in Spanky for the most part so we generate this path and we Mark each room as being part of the path and for each room uh we have a set of template rooms that we can pick from um and what set of templates we choose depends on the area of the game that uh that the level is in and whether the room falls on a path so for example you know if we want to have if we need a corridor that connects two other rooms then draw from the set of uh Corridor

templates and put it in there and within each template there are these little randomized chunks that are specified and that kind of it just adds variability in more interesting uh things to these set templates and then at the very end we add monsters treasure traps and whatever else to kind of complete it and that's about it so working on a game like this is really fun it's very satisfying because you can create a lot of content uh with a small team you know if you look at a game like door Fortress it's essentially programmed by one guy it's this two-person team and they're making entire worlds you know like uh it's just it's really awesome and as a developer it's really fun to work this way because you're always uh surprised by your own game um you know even playing Spelunky now like having worked on now two different versions of the game still

having fun with it because I'm always running into things that excite me so it's really fun to develop uh spunky is not the first sort of refined prototype by any means either that turned into a bigger project we see examples of this in all kinds of media um Super Meat Boy for example started as the flash the free flash version called meat booy uh and Google they've got this thing called Google Labs which they describe as a playground of experimental prototypes that people can just get feedback on and um you know if anything good comes out of it they turn into a full-blown product it's a great way to to test out ideas and it extends even into other creative mediums like writing the uh author Haruki morakami in the intro to a collection of short stories that I think is called Blind Willow sleeping woman but I might have gotten this mixed up but anyways and this the intro to that collection he describes

his process and he says that right after he writes a big novel uh the first thing he does is he he writes a bunch of short stories and he keeps writing those short stories until he comes up with uh inspiration for his next big novel so um so making these small uh prototypes within refining them into something releasing them is a great way to test out ideas it's a great way to refresh yourself after a long draining pro project is a great way to you know get that uh relief creative tension you know and it's also just a great way to get into your next big project so Chris Hecker who I know loves to be quoted out of context um in an interview with Edge that he did with John Blow he described small independent games that don't fully reach their potential as m and he said that big uh you know big

games bigger games that that reach their full potential um he described those as trees and I really like that analogy but uh the way I feel about it is that you can be a mulch and a tree person and it actually goes really well together so if if you're someone who wants to release big games and I I really Define myself and my work by the by the big projects I've released but at the same time working on the small games on on the mulch has allowed me and given me the opportunity to work on the big game so um on that note I'm going to hand it over to my friend and partner Andy Hall he's going to talk to you about the XA version of [Applause] spaly so as Derek said my name is uh Andy that's true um this is a picture I drew myself I was a little bit generous with my hair but uh uh other than that basically he despite that what I mostly

work on on Spelunky xaa is the programming although I am helping with the other aspects as well um so I thought I'd start with how Spelunky XA came to be and basically it all started with a man we all know and love Jonathan Blow uh John was a big fan it took a while you guys had to pause while you read down the list um so John actually was a big fan of the original game and he contacted Dereck and said hey Dereck would you be interested in doing an Xbox live version of Spelunky and of course Derrik said yes of course um and that kind of got the ball rolling with the green light process with Microsoft so the green light process was fairly easy because Derek already had the original game to send into Microsoft they could play it see what made the game fun see what made it work and also they saw that the game already had a built-in audience there were players that were already enjoying it and loving the game a lot um I think a lot of

people myself included feel tentative about releasing smaller projects that maybe don't express the extent of what we think are abilities are but I think spony really shows that people will see the game for what it is and be able to appreciate the things in the game um where no matter what level it's at um and I think that the feedback from those games helps improve the games so they'll get even better um and you know the game and the ideas involved will will evolve as you work more and more on the project John also actually offered to let Derek use the braid engine for spun XA uh Derek took him up on this and actually had a very early version of the game running in the braid engine um and that's when I came on board we spent about uh a month and a half working with the braid engine um and we were having some problems understanding some of the code um there was unneeded features such as the time rewinding with which um we

weren't using that were complicating matters um and when we contacted John to ask about some parts of the code he had moved on to other projects and really wasn't familiar with the code anymore so it made it difficult to get information on what was going on um at that point we decided that it would be best to start again with an engine from scratch um because we had the original game to look at we knew the features and scope of the engine that would be required to make it and we felt pretty comfortable uh that the two of us could handle it and we also like the idea that we would be able to reuse the code for later projects that we work on after Spelunky so as we're working on Spelunky we're really treating it more as a sequel to the original Spelunky than trying to just do an HD remake of the game um this might be scary to hear but uh we can't be afraid to change things from the original game and this took me a little bit of time to get used to because I love the game and I think um

it felt awkward at first feeling free to change things in the game um but that's the only way we can really uh evolve it and make it feel different and special um and I this was made easier by the fact that right the the original game will still exist we're not deleting it by making the Xbox live version um people can still go back and play that and appreciate it for what it is um that said though we don't want to break what worked about the game we want to improve it so we want to keep the complexity and actions and things that made it really fun to begin with so obviously we're trying to improve the game in all sorts of different ways and I think the most obvious and I think most of you have seen a lot of the screenshots that we've put out so far is the art upgrade to HD less obvious is that we're really working on streamlining the controls and the interface to make it easier for players to get into it when I first played Spelunky I didn't stick with it very long on the PC see I had a little

bit of trouble with the controls um and it wasn't until I kept reading on the tig Force Forum uh forums that people were raving about the game then I said okay I need to syn some more time into this and really uncover what is so great about the game um we wanted to eliminate that hurdle of the controls so as an example in the original game if you want to go into a shop and buy something you have to use two different buttons just to pick up the item just to check the price to see if you want to can purchase it and want to purchase it um if you do want to purchase it you have to hit a third button um if you don't want to purchase it you have to hit another two button combination to drop the item and you risk hitting the shopkeeper who will become enraged and shoot you in the face um so to make the shopping experience a little more pleasant now when you're in the shops if you're overlapping an item that's on sale it'll immediately pop up with the price um and if you can buy it uh there will not be that little red mark over the RB button and you tap it and you'll immediately purchase and

equip the item so Rog likes are founded on this idea of having tons of different monsters and items in the game and all the different interactions that come up between them is what makes the game interesting and fun to play so we're going to be adding more of that stuff to make the complexity go up keep the variation of the games up so that not only is the game fun for new players but existing players will find new and exciting things in the game as well um and then there was one feature that was probably more requested than any other feature from the freeware game um and we decided we had to do it for the Xbox live version and that is uh multiplayer [Applause] [Music] so so the game will support four player local play um and it really fits into kind of the chaotic nature of the game um if you're playing with someone you're able to toss them up to a ledge they wouldn't be able to reach um but if you

mess up right they might Ricochet off the wall and fall into the spikes um but maybe that's what you intended all along cuz you want their jetpack um so basically um we'll have more information coming out about the different multiplayer modes uh as we get closer to release so in conclusion the the freeware game really acted as our design document in making the Xbox live version and it made it easy for me as the programmer because I didn't always have to contact Derek about small details with the game I would just refer to the original design um and make it function in the same way um I also use this funky wikii a bunch uh for everything from finding out how much the different gems cost uh to the different interactions with monsters um and I think swy really shows that there's more benefits to releasing these games than I think people realize and that even though you have these small polished games that uh you think kind of are these small project they really can go hand in hand

with the larger commercial projects um and grow into something really cool so um I think if we can take some questions now that would be [Applause] [Applause] great I have a very simple question actually uh I was just wondering how long did did the original spun take to make from like once you got like oh I I know what I'm doing to and compare that to how long is uh the current spun taking to um I started working on the first version of Spelunky like the summer of 2008 and I think I released version 1.0 um like I don't even remember actually I when I first released the game it wasn't

it was it wasn't it was like version 0.6 or something and I kept working on it for a while and I think I I think I ended up releasing a version 1.0 it probably took me like a like a year and a half for the original freeware game to to get it to 1.0 and this version's probably going to take around two years okay so it's it's actually pretty similar for the first version of the game and the XA version um and I think that's because the first version involved a lot of design work which is which can be really difficult and you have to kind of like go back and and forth a lot um but the XA version is just uh it's just a lot of like um you know there's a lot more technical work and stuff like that so it ends up evening evening out a bit actually all right thanks hey I was going to bug you about this later this week but I thought it' be make more sense to Buggy about in front of 500 people um so when you talk about spun and your development process it all sounds very loose right I mean

you're you're sort of creating these systems that are going to procedurally generate all sorts of stuff that you can't even imagine what it's going to you know what what the final layout is going to be like and so on right and you're surprising yourself every time you play but if you go and like study Spiel Lun like I have there's a lot of very tight things that happen in terms of the design the reason you want to play just one more time and you end up playing a 100 times in a given evening um is because it see it seems like there what were decisions made about how rare certain things are How likely you are to see a damsel How likely you are to find a shop with a shotgun just waiting there to blast the owner with and and it it doesn't just seem like like just uh anything goes like just throw all these random things together it seems like there's so many tight game design things that come in there how did you come up with that stuff I mean were you conscious of it or did it just kind of roll out of the procedurality and that kind of thing um you know I got to say it it wasn't too strictly planned land a lot of that stuff like for example how often does a

certain thing appear every time um I kind of just went with what felt right you know one out of 10 or something like that and it it it tended to to work you know I just tried it and and it would seem to work um did you make decisions at a certain point where you tested it and found like oh God uh you know having a damel come 50% of the time is too much and I'm going to tweak it down because it feels like it's I want it to be more rare than that I mean did you make decisions like that at certain points when you're testing uh I did but I can't say that I remember anything that was so significant that it you know I I added it and then it was just completely broken the first time around uh I think part one of the things that helps is just having um so many things going on and not having any one thing kind of dominate everything else so that um you know even if the the the form foras aren't completely uh Focus test tested to a te

or anything like that it still feels really compelling right right so I guess we just have to chalk it up to genius I I wouldn't say that I didn't say that so um in fact in a lot of ways I feel like um it's it's maybe a lack of Genius that that helps you come up with solutions for some of these things like for the for the level generation I I really just tried to do what was simplest and what what what seemed like the most just Brute Force way to to make a level that worked so I'm no genius quote Derek you uh could you talk a little bit about the uh production requirements for the XA version is it just the two of you or yeah it's just two of us um and we have devits and and Andy actually lives in Connecticut so we

usually you know he'll come out here and we'll we'll Jam together for like a couple weeks at a time but usually we just work over Google video chat um and pass code back and forth I think uh I'd be I'd be pretty embarrassed to actually go through our entire you know our entire production process because it's pretty ghetto fabulous so definitely got our Indie cred going on and do you guys have to worry about um you know like marketing and PR of this or you kind of relying on the community that is something I worry about and it's something I worry about because it's it's honestly something that I'm not that interested in when it comes to um you know I can't say that I'm really passionate about marketing at all I I like talking about my my game and and I'm really enthusiastic about working on it but I'm just my personality you know I don't know my personality I'm just not um I kind of I when I'm working on games

I kind of really like to put my nose to the grindstone and actually kind of uh my inclination I think is to really just Hideout so I am worried about about marketing and I'm we're trying to think of things to do that fit our uh personality and our fun um but it's it it's it's something that we really have to work on I think yeah I mean I come from a more traditional Product Company I was a toy designer for a while um five years for 5 years five long years um but I when we're talking about how we're going to promote the game I'm very much in line with traditional ways of marketing it um and you know suggesting things I I mean I see Derek's side of it that you know we do want it to fit the aspect of the game you know kind of the feel of our team and you don't want to kind of do something that's not in line with the game um but at the same time uh we don't want to be too understated as well so I think we're still trying to find that balance I do want to mention that um there's a third person on our team and that's our

musician Eric who is in Norway right now uh fiston um he runs the pause Indie music indie game music label so anyways want to put that out there thanks and I think we got one more question um in the last few years there seems like there's been a Resurgence and the more difficult games like spunky and and like Demon Souls that sort of thing what do you think think has caused this Resurgence and how do you balance like difficulty versus frustration in your game right I I mean I think it goes back and forth so you know it feels like games are getting too casual you hear people really um complaining about about not having games that they can really enjoy and I think the the key to making um a challenging game that's not frustrating is just to to give the player all the tools that they need to overcome the challenge and never make

it uh never make it the game's fault that they lost essentially so um you know with with games like Rog likes one of the things that's so fun about them is that you have all these tools at your disposal you never you never know when you could get into a situation where you know I don't know you're you're you're surrounded by by monsters and all of a sudden you have to use a tin of goblin meat to get yourself out of it and you you feel really smart about it but you know Rog likes give you lots of those tools so even though they're really tough uh it's always fun thank you and I think that's it so thank you very much thank you [Applause]