System-Centric Puzzle Design in 'Patrick's Parabox'

[Music] hi everyone this is welcome to system Centric puzzle design in Patrick's pair boox um hi I'm Patrick I mainly do programming and design I've been making games since around 2011 mostly freeware games as a hobbyist and in March of 2022 I released my first commercial game which is called Patrick's par boox this is a thanks um this is a block pushing puzzle game with a recursive twist it had about a four-year development time and halfway through development it won the igf Excellence in design award um could I get a quick show of hands of who here has played uh the game oh a lot of people that's very cool to see thanks um so I'll still explain the core mechanics so here's a gift of the gameplay you can see you push boxes

around but there's also this blue box which is a container for other boxes and if you squish it against the wall you can enter it and and push other boxes into it and you can push them out this is a really cool system I discovered while prototyping and it ends up giving rise to lots of compelling mechanics and puzzles here's another gift showing a situation where this blue box actually contains itself so if you exit from the blue box you will end up back in the blue box um this gives rise to even more puzzles with the recursion um so these two gifts show the core of this system that I discovered but there's also more recursive mechanics that end up gives rise to as well such as what if a box contains multiple copies of itself or what happens if you push a box outside of itself causing a paradox or what happens if the player character is itself a box that can contain other boxes it's like wild stuff can happen um so what I mean by System Centric in this talks title is that this

system that I have happened upon discovered is something I find legitimately super interesting and compelling and I want to tell all my friends about it um so my goal in making this game is to Showcase this recursive box system and on the flip side of the coin players motivation when playing the game is that they want to see more of this system it's exciting to see what new mechanic is around the next corner and this great system that I luckily discovered happens to be very generous and that there's a lot of things it has to to show um so I'll give a um selection of sort of a sorry a design sort of postmortem a bunch of Snippets that applied to this game but maybe you can apply to your game whether or not it's a puzzle game um so one thing I did thinking about showcasing the system was I Tred to keep an approachable level of difficulty um so relatively easy compared to other games in the hardcore puzzle genre um because my goal was to Showcase and communicate things about this cool system not necessarily to challenge players or to stump them or or other

goals that I might have um I did this because it's my personal preference but also this really unexpectedly broaden the audience of the game more people could get into it who weren't otherwise really good puzz puzzle players but they could get into it and maybe even make it a lot of lot of the way through um yeah um so to show some examples of making an approachable difficulty here is a puzzle on the left where this is a puzzle about you start in this little left pocket where you put your yourself and the blue box into the green clone box with the entering on the correct side such that when you exit on the right side of the level out of the green clone box things are positioned correctly and you can solve the little maze on the right here's a simpler easier version of this puzzle with it's about the same basic idea but now the constraint on the right is simplified and easier once you get the the cool recursive idea you can then move on without any unnecessary roadblocks

and the right puzzle is the one I use in the game here's the puzzle on the left where there's four blue boxes and your goal is to separate them so that there's one space in between each and there's a you have to utilize the two green recursive boxes on the right and do a cool recursive trick in order to separate them on the right here's a version of this puzzle with only three boxes and more empty space which gives you more leeway so now all you have to do is recognize this recursive trick and there's no tight box finagling or finessing that the left puzzle has so there's less less friction I ended up using this puzzle on the right but I did still like the cool complications that the left one has so I left it in as an optional puzzle on the left here's a puzzle where your goal is to enter into this green box but you can't do so directly because there's walls on all four sides of it you have to utilize the other three boxes in the level to do so on the right here's a version of this

puzzle where this green box is constrained in this little cranny inside the wall so you can't move it so this constrains the number of possibilities in this level so there's less things to do that aren't the solution or related to the solution so you can get there quicker um so to codify some of these things I did I tried to simplify puzzles as much as possible I tried to make puzzles as easy as possible while still showing their core idea while still communicating what they're about a I removed some solution steps that were unnecessary or too hard I when appropriate I trimmed the possibility space so less things I do that weren't the solution or weren't related or I didn't want to be possible and in general try to keep the puzzles B sized because keep they're more like hey did you know you can do this cool thing Isn't that cool and oh here's another one I'd rather keep people keep showing you things so there came a point there's a question though of like well how easy is too easy because for some of these easy

puzzles players would stumble into the solution without understanding what happened or they would not understand mechanics at a very deep level which is not great but I decided Well I'm okay with this happening sometimes as long as I'm making this trade-off in favor of or in service of making the game more approachable and streamlined and this is sort of maybe a controversial thing or it's there some subtlety involved here about sometimes it is bad but just sharing my perspective that it worked well in this cont text okay so puzzles don't exist in a vacuum they're surrounded by puzzles before them after them and on all sides of them here's a sequence of three puzzles where a introduces a new mechanic and B and C expand on it in more complex and difficult scenarios a play tester was having trouble with c and I realized oh yeah that's kind of an unfair leap from B to C so I added this puzzle in which Bridges the gap

and this was an improvement end was a nice puzzle but another player was still having trouble on C and I thought well I i' like C but it's I'd rather not have people get stuck on it rather than move on so I did just that and put made C an optional puzzle so now after you solve n you can just proceed onwards this is the kind of thing I did a lot in the game with puzzle ordering um here's another example so here's seven puzzles the first of which introduces a brand new mechanic it's the infinite exit mechanic and then there's six other puzzles that give you gradually give you more practice and more complex scenarios when I play tested this it just didn't work that well the introduction was just confusing and the constraints were clear there were just weird weird abstract puzzles that you couldn't quite grasp so I cut the ended up cutting these five puzzles and the two I didn't cut they were actually okay and I kept them with some changes and here's what I ended up with next I added a better introduction to this mechanic that's more clear and I added two better

reinforcement puzzles to reinforce your knowledge um and not doing what I knew didn't work in the first time and I added two optional puzzles which are the ones in the lower right there that were also good so this was an improvement but then another play tester had trouble with this mechanic at the very first puzzle which was they thought well they thought this mechanic put them into a lose state but it actually doesn't so I was I thought well let's just add a level that shows you that and actually I was able to add four levels that clarify this doesn't put you in a loose state and this constrained environment actually ends up being kind of having nice puzzles so now I have an even smoother introduction to this mechanic um which is great I tried to apply this kind of iteration across every single puzzle sequence in the game um so these each of these boxes is an area in the game and if you squint or look closely each of those tiny little boxes in there is like a sequence of puzzles um just showing this to show the scope of the work I triy to be diligent and do this for every sequence

um so to codify some of these things um the general verbs I did was inserting a puzzle modifying a puzzle deleting a puzzle reordering a puzzle or optioning a puzzle and I did these in order to patch up conceptual leaps improve tutorialz or clarify misunderstandings um insert puzzles to reinforce knowledge or give someone a second chance to learn something if for example they stumbled into the solution in the first instance and also to manage playtime because I want people to be moving through the game and not get hung up on a puzzle even if it's good I have so many other things to show I learned to treat this kind of puzzle management or puzzle ordering as its own distinct skill when making this game and I found this skill to be just as important as the skill of making puzzle in the first

place here's a GI showing a the tools I use to for example insert a new puzzle into this sequence of three so now it's a sequence of four all I do is drag the puzzle into the ordering and then I can recompile the game and all the connections work no additional configuration is required on my end having this tool be quick and joyful was very nice because I was making so many of these changes and I would rather encourage myself to do lots of these things as opposed to if the tools were slow or miserable to use related I also did not have puzzles in the hub area even though that'll be cool it's like a recursive game so why not have puzzles and puzzles but I was iterating on this order so much that I would have been breaking them all the time so I made a trade-off there and it was it was good I iteration on the port the puzzle order a lot um helped um related to this unlocking leniency in puzzle games so on the left here's an older version of the game where you can see this player has gotten

stuck on the third puzzle in this area and in this version if you get stuck on a puzzle in the main line there's no other option you are stuck you got to solve it in order to move on um which is bad um on the right I made this change in the Laton development where it looks more like on the right where you can see this player has now made it through a lot of puzzles because at any given point I made it so that you have access to multiple possible puzzles to make progress on this is very common ad acknowledged in puzzle games to always have multiple paths available a but I'm just mentioning it here to stress that this is very important in making the game more approachable and I want to be generous in showing people cool stuff about this system not putting roadblocks in front of people so I've talked about puzzles and puzzle ordering and I got information on how I want to change those things based on play testing and I want to endorse a specific format of play testing which is

video recordings with the play tester narrating their thoughts out loud this is more work for the play tester because they have to set up a recording software and also takes effort to narrate your thoughts all the time but I found it worth it with a video you can see all the little things and attempts players do in Puzzles with the N voice narration all the time you can hear their theories of what they're trying and thinking about the system and what their emotions are at different times you can measure time taken on different parts so if a part of a puzzle takes 10 seconds versus 20 seconds versus 60 seconds videos are rewatchable at later dates so I can see what a play tester 6 months ago thought about this puzzle and what was the state of the puzzle at that time um and to get so those that's the video format and here's a to get more specific about opinions here's past Patrick's Patrick's parabox play testing preferences um so um I'll list off a bunch here I tried to get fresh play I tried to get fresh play testers I'm tripping over my I'm trying to get fresh

play every time because you can only experience a new mechanic for the first time once and these mechanic introductions are exactly what I want to get smooth non-expert players were very helpful because players who had lesser skill would stumble would encounter more stumbling blocks and a lot of those stumbling blocks are the kinds of things I want to fix to get people streamlined through the game I tried to pay very close attention to players mental models so I knew where to for example clarify misunderstandings I tried to treat small bumps that players occurred encountered as potentially very important because that could be a big hurdle for another player again thinking of streamlining the game I like to do batches of one or two play testers at a time in order to give each play test video um due attention and respect that play tester's time because I would watch these video in detail and take copious notes um I tried to get play testers who would play the entire game so as to not

of course neglect the whole second half of the game um personal learning was to what for me was to watch the whole um video and take notes before making changes as opposed to pausing the video every time I wanted to make a change because that would give me more context by the end of the video and it would just help me work more efficiently and I did these play tests every so often whenever I felt like the game was ready for another batch of feedback so it was nice to have a sort more or less constant in the background thinking about iterating on this puzzle ordering as opposed to like backloading it or something um side note it was difficult to find players who I knew who were nonexperts would were willing to record themselves and narrate their thoughts and were willing to play the full game but those players are also like by far the most helpful play testers I did about 15 of these full game video play tests throughout

development uh not counting other formats of play testing but here's 15 videos by the end of development I thought oh it's sort of I'm reaching diminishing returns with doing these play tests but actually after releasing the game and watching people play I realized no actually I didn't reach diminishing returns all those little changes and puzzles I was editing and inserting at the very end were actually doing great work and I'm glad I did them and I could have done more so that's a learning for me is I'm going to keep doing lots of play testing okay now I want to switch gears a little bit and talk about um making puzzles uh to me the purpose of this recursive box system I thought of it as it's not like that the purpose of the system is that you can make it can make cool puzzles I like to think of it as the purpose of the puzzles is to Showcase things about this cool system okay now here's a list of a bunch of techniques I've accumulated from

other people and along from my travels about brainstorming techniques for making puzzles um by far the most common technique I used was finding an interaction in this system and then making a puzzle that forces you to do this interaction as the key move in the solution of that puzzle so for example on the left here's a situation where you can enter a box by pressing it against a stack of boxes against a wall which which is kind of cool and on the right here's a puzzle that forces you to do that to construct that stack and then press against it in order to free the box from the wall so one thing is like well how do we how do we find these interactions that we can then force you to do in a puzzle well in general just playing around with the system and exploring the mechanics and you know normal stuff just playing around um but to get to more specific some things that might be helpful are making a big playground level with lots and lots of objects you can experiment and making test levels that test edge cases or extreme

cases and one technique which I found helpful toek out a few more puzzles which I've learned from Brett Taylor is to enumerate um list out all the mechanics in in the game and then list out all possible pairs of mechanics and then for each pair consider what might a puzzle with that pair of mechanics be like so for example the Clone mechanic plus the empty mechanic or the Clone mechanic plus the eat mechanic or what about clone and transfer what about clone and possess clone an infinite exit so on another technique was to try interesting or extreme geometry so I'll list these out from top left a level with no up easily accessible upwards facing walls interesting geometry or what about a level with no walls at all and it turns out this is actually a cool puzzle um what about a box with that's entirely filled with other boxes um what about a 3X3 very small recursive area that's fun to navigate in

or what about a box that contains eight copies of itself which creates a really trippy and interesting interaction um these These are what I like to call things almost happen moments uh I really like these so when watching a play play testers sometimes they would try things that wouldn't work out for whatever reason they weren't the solution but what if those things they tried actually were the solution um so and make a puzzle about that so here on the left here's an example of a play tester who would want to free this blue box from the left wall and they try to push this green clone box into it but it doesn't end up working but on the right here then I made this puzzle where you actually do use these clone boxes to free boxes from the wall um so I like to think of this as sort of like crowdsourcing puzzle ideas to play testers um which is pretty cool highly recommend this technique and it's also another reason why I like to pay very very close attention to each and every play tester and every minute of their actions to really get into their mental models and make a lot of observations

these These are what I like to call show levels so on the left here's a level very early in the game which just shows you you can enter and exit boxes and there's actually no way to fail this level on the middle here's a level where all you do is enter four levels deep into boxes and then just go out four levels it's just to show you you can go lots of levels deep and on the right here's a level where there's a blue box that hasn't contains an orange box and all you have to do is flip them so that the orange box contains the blue box all three of these are absolutely trivial levels to solve you might not even call them puzzles but they still showcase cool things and players find these levels delightful so you don't always need a challenge um when making puzzles I tried to be diligent about investigating variations so for example here I made this puzzle which I liked and then I thought oh what what would an easier version look like and oh actually that that's end ended up being pretty good and oh what would a harder version look like oh actually that one's good enough

to keep too and what kind of additional constraints could I put on oh here's another additional constraint like a weird line constraint and so now I and now I have four good puzzles um so I tried to investigate all these possible variations with every puzzle I made um I got used to doing lots of digging when making puzzles so this game has 364 puzzles that it shipped with but also over 600 unused draft levels and these were levels that were ideas that didn't quite materialize into a real puzzle or there were older versions of puzzles that did ship or there were puzzles that just weren't that great um so at the beginning of development I got kind of discouraged by this but by the end I learned not to get discouraged by this and that it's just part of the process and there's a lot of trial and error involved um related to this I found that one kind of simple heris of what makes a good puzzle system is simply what's the

number of how many puzzles can you make in it um when making pair box I early on it just ended up being that oh I can make a lot of puzzles in it and that encouraged me to continue on with this prototype um nowadays as I make more prototypes I find it helpful to keep this heris in mind when thinking about which prototypes I want to pursue more it's sort of just an obvious heris of well the proof is in the pudding how many puzzles can you actually make in it now I want to switch gears again and lastly talk about a few mechanics a little bit so early on in the game I faced this one of many dilemas I faced about the mechanics was what if you try to enter a box but there's like multiple openings on that side of the box which opening do you enter I was concerned about this and you can see in this image right now the player enters no opening because the player only checks the center spot on that side and here you can see it's filled by a wall so you can't enter it I then tried tried this system where you can actually enter any spot and it just prior IES ones that are closer to the center so now it's like a you can

see with these numbers for number one is in the center and then it sort of spirals outward which are ones it tries um I like this because it was more flexible I could make more levels and it you know but I ended up then realizing no actually I don't like this and I think this is bad because it's sort of a complex system that I'm introducing and it felt kind of arbitrary and like how do you break ties and stuff and also we're going to be this is something that the player has to keep in their mind the entire time they're playing every level in addition to all this recursive stuff mechanics going to be piling on top of it um so I reverted to the center only behavior on the left and that ended up being a good call here's another example I knew I wanted to try this mechanic of well I wanted to be able to transfer consciousness of your player into another box so that you can control two like multiple boxes with different shapes my first attempt was to add this little arrow object where you just push a box onto it and then you control that

box very straightforward obvious thing to try and this worked you know I can make some puzzles with it but it kind of felt weird or wrong or arbitrary it's kind of a touchy Fey thing but I didn't quite like it um I next tried this which was you can you transfer Consciousness or possess a box when you squish it against a wall and if you hit a solid wall then you transfer Consciousness otherwise if it's if it's an opening on that side you just enter it I like this a lot better because this synergized with other mechanics because different sides can have different geometries so you can synergize with other mechanics and it just felt it just felt better so I went with this version on the right and I'm glad I did not go with the first one I'm glad I did investigate another variation here's a few mechanics I tried very early on in development um so I had this like a weight object where you could like push a weight inside of a box and then now that box becomes weighted which means you cannot push it it's like heavy um and then I had also had an object where you push the shield inside

of a box and now that box is shielded which means you cannot enter or exit it there like a shield you bump into around the box which is it's okay and then um then it had sticky boxes which stick to each other when you push them and you can unstick The Boxes by putting them on a block Edge and like using the recursion to unstick them um these are all these all felt kind of cool and I was like oh yeah I kind of like them but I want to I didn't like them that as much and I want to contrast these with other mechanics that did make it in the game which are for example a box containing multiple copies of itself or a box containing a horizontally flipped copy of itself or of course the player itself being a box that can contain other boxes these are sort of more natural extensions of the system that sort of almost don't quite feel like mechanics I mean I don't know what terms we're using here but these felt a lot better to me they gave rise to lots of puzzles and they were just organic um so I didn't use any of the ones on the left um so my T learnings from this are always investigate more mechanic don't just go with the first things that come to my

mind and try to always think of what's natural what natural extensions can I do here are three mechanics that did feel natural but didn't make it into the main game so on the left here is what happens if you if entering a box takes priority over pushing a box whereas in the main game it's actually the other way around the priority in the middle what happens if you exit a box against a wall if you could like squish out and like free that box from the wall like that kind of makes sense and on the right is what happens if you could push a box around from inside of it kind of like piloting a submarine or something um these are all cool but they just didn't didn't work out for the main game they didn't play well with other mechanics or they didn't feel right or they like didn't make a lot of puzzles combining with other mechanics they just didn't work out but I still these were still in my mind and I liked them and I thought they were worse showcasing when I'm thinking about showcasing things about this system so I came up with this idea of actually include these three

areas in the post game which do have these mechanics but they're confined to that postgame so they don't contaminate the rest of the game and the other mechanics um so I was happy with the solution these are cool puzzles and mechanics I'm glad I found a way to Showcase more things about this system that otherwise I wouldn't have a been able to Showcase without this format so you might want be interested in using some kind of appendix format in your game or maybe you can think of another creative format thing to Showcase more things you otherwise wouldn't have been able to all right so to recap everything um so my goal in making this game was to Showcase this recursive box system that I had luckily discovered and that I found really cool um to that end I made the puzzles an approachable difficulty and style because my goal was to Showcase and also this ended up really broadening the game's audience I put a lot of iteration into smthing the learning curve which I

talked about inserting modifying deleting puzzles um I really end this play testing with video and narration format um I talked about a few techniques for finding puzzle ideas like finding cool things that are worth showcasing and a few instances where I iterated on the the core mechanics of the game um here are five online puzzle design resources which I highlighted recommend these have been very influential to me um would recommend checking them out um and here's my contact information uh thank you [Applause] [Music]