This Problem Changes Your Perspective On Game Dev

this will completely change how you think about game development it's the best theory I've ever found for why my games have succeeded so far and in the context of game development I've never seen this discussed before making a game is like running a search algorithm you're searching for the best game you can possibly make this is your boat you're on a massive Lake and your goal is to find the deepest spot as quickly as possible you can move measure or teleport to an already explored position for example we could move four times make a measurement

move nine more times measure again and then lock in our best result as our final answer have we found this deepest spot in the lake at this point I would say fairly unlikely right the next Captain comes along says I'll do better I'll start with a measurement they keep making regular measurements and as soon as the water gets shallower they turn the ship they have a look at their data points and take an educated guess for where to explore next the ship keeps exploring with this method and then they lock in their best answer and

that's already better but the third Captain is a gigabrain the third Captain first explores all directions making use of the teleport ability from there two most promising measurements are at the top and at the bottom so that's where he continu use a search from there we can use the strategy of the second captain search around a bit more and settle on the best result design is a search algorithm as a game developer you're searching for the best game you can possibly make if your search algorithm is poor you won't find it how you search in

which order you try things when what you experiment with is absolutely crucial for your success in this video we'll talk about how you can optimize your search as a game developer and hopefully turn you into a big Brain Cap there are seven main problems holding your search algorithm back so let's talk about how to solve every single one of them the first problem almost all search algorithms suffer from is the speed versus accuracy trade-off your search algorithm can either be fast or it can be accurate but unfortunately it cannot be both you can either just

drive in a straight Lane and make occasional measurements to see if you're still on track very fast but not very accurate or you can drive a zag Lane all across the lake and map out every single inch of it very accurate but very slow that means both of these strategies are flawed in their own way and The Sweet Spot is somewhere in between a fairly reasonable approach is for example to start with a slow but accurate search in all directions and then to transition into a less accurate but much faster search go wi first and

narrow later this is not the only search strategy that works but for Indie Games in particular this is a very reasonable one you have an exploration phase in the beginning where you try to figure out the best direction for you to go in when Paul and I started making Throne fall we tried out a crap load of different ideas from multiplayer climbing games to games about golfing flowers every experiment every Pro protype is essentially a measurement How deep the water is to see if there's potential in this direction in the end the Prototype we found

most promising was this one here by Paul where you control a little King and have to defend your kingdom from enemies at night the next phase would be the pre-production where you still want to do some Active search to narrow in on the exact Direction you want to go in at this stage Paul and I prototyped a very simple economy and combat system in unity while simultaneously doing some very active search and experiments on our art style in pre-production you are still searching just in a smaller radius and then finally in production you sacrifice your

search accuracy for Speed cuz you got to finish your game someday for thrf fall this meant rebuilding all of our systems from scratch I rebuild the combat system Paul rebuilt the economy system we put it all together in a little blockout and from there we just kept polishing until it looked like this I think it is important to note that you should never stop search completely you still want to do your occasional measurements you want to do your occasional play tests you still want to find the deepest spot in the lake so every now and

then you should just give yourself the opportunity to correct CA if necessary the local minimum a very common problem in search algorithm it basically describes the illusion of having found the deepest spot because if you move forwards the water gets shallower if you move backwards the water gets shallower you logically conclude that you found the deepest spot already and that it doesn't make any sense to search any further what you don't know is that there's a much deeper spot just a few more meters down the line there's only one way to solve for this problem

and that is to DARE bigger jumps in your search tree every now and then your production might be moving in a trench that you can't break out of by making small adjustments but if you make a bigger jump and dare to cross the Hill you might discover that there's an even deeper trench on the other side attempting this is especially worth it whenever it is cheap to do so the game fornite was initially meant to be some kind of Base building tower defense game that's why you have all of these building mechanics in there but

the development team was brave enough to DARE one of those big jumps and try to make a completely different game mode and that mode became one of the most played games in history as the search space is so massive it's almost guaranteed you're stuck in some kind of minimum which means you might have a fortnite slumbering right around the corner and not even know about it surprisingly there are a lot of opportunities to make big jumps in your search tree for very little effort and whenever an opportunity like that presents itself to you you should

absolutely go for it an obvious example would be game modes game modes are often very easy to make you have all of the game systems all of the game mechanics all of the objects already laid out just change the objective change a small thing make a new game mode other obvious things could be completely rebalancing your game another way to get yourself unstuck from local minimums one very brutal thing about local minimums is that you have no way of knowing if you are in one so typically there doesn't even have to be a feeling of

being stuck or anything like it the opposite is the case right you're feeling like you're optimizing to the best of your ability that means even if you feel like things are going well give yourself the opportunity to break out of local minimums do it you're not on a lake you're on an infinite ocean to make metas even worth it's not even a two-dimensional ocean it's it's a multi-dimensional ocean there's an infinite amount of games you can make which raises the question how do you search an infinite search space well I have a very sophisticated answer

for you you have to take a guess but you don't have to do so blindly you can do so using measurements from your database luckily for you you're not the first person to ever make a video game there are tons of video games out there already and we roughly know if they succeeded or if they failed on top of that you have the experiences from your own past projects and this is your personal database now in guessing a direction the safest thing to do is generally to pick a direction that has worked in the past

but to then search around those data points to make sure you don't make a game that exists already controversial statement unique selling points are overrated cuz when somebody tells you you need a unique selling point what they actually mean is don't freaking Park your ship on top of another ship don't make a game that exists already but verse that is a horrible idea on the other side of the spectrum this ship is floating in a swirl of color oh my God so creative but nobody can even tell it's a ship anymore this is the equivalent

of making a super Innovative artsy fartsy game that eventually nobody's interested in do not make a unique game for an alien species we have no innovation on one side we have too much Innovation on the other side both of them are horrible The Sweet Spot is in between in order to find that sweet spot you want to be near other successful boats but you do want to give them plenty of space and not really get too close to them when somebody tells you you need a unique selling point for your game what they really really

mean is you need to position yourself correctly on this Innovation scale the reason why I don't really like the unique selling point framing of this anymore is because I feel like it encourages too much Innovation or sometimes also pointless Innovation so while I technically agree and I'm talking about the exact same thing I just feel like this Innovation scale is a bit of a nicer way to frame it nurture your database a little bit like play some other games read the reviews get some data from other games you don't want to do all of the

exploring yourself and then freaking use that data to navigate this infinite search space cuz the search space is massive there are different search algorithms that make sense in different situations for example if you want to make a new indie game and you have no clue what to make then go wide first and narrow later whereas if you're trying to search a specific Corner then go there first and then open up your search the only thing that matters is that you open up your search eventually cuz without search you can't find a good local minimum chasing

after the wrong thing is another very common problem in search algorithms and that's also the case in game development you think you want to make a viral game whereas deep down maybe you just want some recognition and laugh congratulations you found the deepest spot in the ocean but now you figure out that you're a fishing boat and that there are no fish here making a game that brings you Joy is a very different story from making a game for your portfolio which is a very different story from making a game that is meant to commercially

succeed what you optimize for really matters what gets measured gets improved most game developers try to optimize for Revenue which I guess makes sense cuz a if you want to keep making games you need revenue and B in the indie game scene revenue is a very good proxy for the quality of your game anyway that means if your game is better if more people enjoy your game if people have a better time with your game you're going to make more Revenue so it's a win-win with the players anyway whatever goal you might have you can

probably split it into smaller derivative goals in order to make Revenue in my opinion you need three things you need fun for keeping players you need appeal for acquiring players and you need scope for making sure that you can actually finish your game I very briefly go through all of them cuz I know a lot of you will be optimizing for these factors the best theory of fun that exists so far is the one of Flow by Mii chick sent Mii flow is the channel between borom and frustration where difficulty and skill are in a

perfect equilibrium it's honestly quite simple to wrap your head around and the more you think about it the more you realize that this applies to everything like any kind of player skill that can be tested like comprehension reaction times strategic thinking puzzle solving is going to create a lot of fun if you manage to put it into the flow Zone yes even the story of your game like if you spell everything out it's way too boring and if it's way too complicated then it's frustrating one more framework that I think is useful to know is

the octalysis framework by yukai Chu it describes a related concept which is motivation it names eight factors that can be used to motivate people we have the intrinsic motivation sources at the top the extrinsic motivation sources at the bottom the creative motivation sources on the right and The Logical motivation sources on the left so everything is quite neatly organized it's always allowed to do some of your own Theory crafting as well of course I made some cool videos about this already like the the fun button video or a much older one where I talk about

how to design choices that are fun all in all fun is a very well covered subject in game development it's also very easy to measure you just do some play tests either with yourself or with other players remember that what gets measured gets improved so if you want to improve the fun then you have to measure the fun moving on to the next Point appeal is essentially another word for marketability while fun is for keeping players appeal is for gaining players AAL is the quality of a game that makes you go I have to play

that immediately to demonstrate this I brought you two screenshots one is from my game Islanders and one is from my game or robas you could run the exact same marketing campaign on both of those games but one of them would do significantly better can you guess which one having an appealing game makes marketing infinitely easier it multiplies all of your marketing efforts if you have a game that is 100 times more appealing it is 100 times easier to Market some developers only focus on the fun and forget about the appeal and if you're trying to

be commercially successful that is crazy you need appeal Throne fall so much easier to Market than will you snail like marketing wise we did almost the exact same things right there are only a couple of standard things you can you can do for indie game marketing and yet one of them is so much easier to sell and so much easier to Market and the only difference is appeal there's not really any good theory crafting about appeal already out there so I had to do my own this is a Jonas original formula appeal is presentation plus

fantasy multiplay by readability let's break it down presentation is exactly what you would expect its art its sounds its visual effects its polish and its juiciness here Illustrated with a screenshot from Bal gate there are also more Indie friendly ways to achieve the same thing with low poly art styles like here with our game Islanders or you can even just go for Pure juiciness like in one finger death punch fantasy is a bit more difficult to explain but I would say it's basically about the question whether it's an experience people actually crave and more specifically

it's about role playing a good fantasy is something that people people would like to experience in real life but can't experience in real life either because it is too dangerous or because it is not practical a great example for this is super flight where you fly in a wing suit flying in a Wings suit is something a lot of people would like to do but unfortunately this just too dangerous and not practical for most people and guess what that's exactly where games can help you can really find fantasies that work in very unexpected places sometimes

the other example I brought for you is power wash simulator a game about pressure washing what matters are the positives of a fantasy cuz the negatives are something you're just going to remove from the game Power watch there are no Angry customers you have to deal with you don't have to do any bookkeeping you just do the fun part the lesson from this and also from games like eurotruck simulator is that people are interested in real life fantasies that are somehow grounded in reality also all of those typical childhood fantasies being a king being a

pirate building a theme park being a hero going on some kind of epic adventure and saving the world all in all just things we would all love to do in real life but for one reason or another can't it is important to note that not all fantasies are created equal some fantasies attract more players than others that's why it's important to pick your fantasy very carefully and not only does it matter which fantasy you pick it also matters how much you manage to live up to that fantasy players want to feel it they want to

be fully immersed in the fantasy you're delivering can you make a game with no fantasy at all yes you can it's called abstract games but in that case your presentation better be on point you can however not make an appealing game without having readability readability is absolutely key readability means how easy is the game to understand from just seeing a screenshot or from just seeing a few seconds of gameplay can you tell which fantasy the game delivers just based on seeing a screenshot can you tell what the game play is about just from seeing a

few seconds of gameplay in case you do not know about this game here then try to guess what it is about all right and now we'll look exactly at 2 seconds of game playay and I think at this point all of you will have guessed that this is indeed a game about climbing this game is called a difficult game about climbing by our fellow YouTuber Ponty pant put it on your wish list in my opinion the readability of this is absolutely excellent you can just have a very quick glance at this and you immediately know

what the game is about there is some intentional resemblance to Bennett fodies getting over it being compared to other games is not always pleasant but if it helps your readability it can actually be worth it readability is the bridge between presentation fantasy and gameplay the presentation must communicate the fantasy and the gameplay this is a screenshot from snake birds and just by having a glimpse at this you immediately know that this is a puzzle game because of the blocky nature of it you know that it's kind of a cute game and you know that you're

likely going to control these little sausage bird things so from one single screenshot we could just immediately understand and grock the entire game for our indie game Throne fall the readability is good but people sometimes think it's a classical RTS game just cuz it looks like that so what I'm really just trying to say is the more you can get your presentation to explain the game for you the better that was The quick summary for how to create a pie the only other thing you need to optimize for revenue is scope because it turns out

we don't just care about Revenue we care about revenue or return per investment meaning that if there are two equally sized pools of fish we want to go to the closer one if you go for the patch of fish that is further away you might even run out of fuel and not make it at all the good old keep it small keep it simple keep it tiny for some reason a lot of developers still think that tin games cannot succeed but the opposite is the case like in the Indie Arena tiny games can do extremely

well put quality over quantity keep the scope small optimizing for the wrong thing hurts a lot so what should you optimize for if you're going for Revenue then you should optimize for fun appeal and scope and that means right from the beginning not as an afterthought I have to admit I'm a bit of a formula lover so for all of the formula lovers this is the formula you want to optimize problem five another very common problem in search algorithms noisy measurements you're all excited about all of the fish you found whereas in reality you're just

floating on top of a bunch of poop particles this most commonly happens if you have only one play tester yourself cuz all of our judgment is clouded sometimes but even if you have more play testers this is a very common problem you will have some noise and some inaccuracies in your measurements the solution is to become a scientist and to just simply measure twice where it matters let some time pass let the feeling cool down the poop might just float away exploration costs how can you afford all of this expensive exploration wouldn't you arrive at

your goal much faster if you left out all of these unnecessary search paths there you have the quick answer no you wouldn't you wouldn't even know the best direction to go in you might have the illusion of going fast but you would be going in circles you would notice your mistakes too late and you would end up with a suboptimal design don't be afraid of exploration costs the lack of exploration is far more expensive with that out of the way I have to agree though you want want to minimize the cost of exploration as much

as possible minimize the cost of exploration so you can explore more not so you can explore less so you can explore more cuz the cheaper your exploration the more of it you can afford to do the cheapest form of exploration you'll use most often is prototyping a prototype is like a tiny scouting boat your main boat is big heavy and slow and it's not very suitable for fast exploration but your scouting boats are a prototype is a very quick and Scrappy version of whatever you want to try so should you build perfect production ready code

when prototyping hell no cuz I've been asked this before the only thing you care about when sitting in a speed boat in a scouting boat is going fast and making a halfway decent measurement that's roughly in the correct ballpark those are the only two things you care about when prototyping so take shortcuts go as fast as possible do not focus on writing clean code that does not matter at this point depending on your specific search strategy you can send scouting boats throughout the entire process whenever you want if you're already a bit further into the

process then it can sometimes make sense to actually Scout with your main fishing boat if whatever you want to build relies on a lot of existing systems then obviously you prototype that right into your game you don't start a new Unity project and make everything from scratch just to try out an idea the rule of thumb is you do whatever is fastest whatever is fastest that's what you do and for that exact reason you should always prototype art and gameplay separately you want to know what happens when you build gameplay and art at the same

time it's called making the game you're not prototyping anymore you're just making the game and you're losing all of the speed advantages that come with prototyping so building art and gameplay separately is crucial it gives you a massive massive massive speed boost I'll give you an example when we made our game Islanders this was our visual prototype and there's no gameplay in this whatsoever you can move the camera and toggle a bunch of building buildings on and off but that's it as there was no gameplay to worry about our artist freedoman could explore a lot

more freely and a lot faster than he otherwise could have the gameplay was prototyped separately with a bunch of blocks and boxes that look like really ugly programmer art that allowed us to experiment with a lot of different ideas and to iterate quickly rule of thumb systems that are separate from one another should be prototyped separately from one another that way you are much faster and you stay more agile during the process additional bonus your measurements will be more precise because you're not adding any unnecessary noise to your prototypes we did the same for throne

fall fall our gameplay prototypes did not have any art and our art prototypes did not have any gameplay and we were significantly faster because of that another thing you can do to reduce exploration costs is to parallelize send multiple scouts at once especially in Indie teams ever only put one person into a scouting boat it's better to have more scouting boats than to have bigger scouting boats if you ever have idle team members sitting around put them into a scouting boat immediately speed up the evaluation of your prototypes take shortcuts wherever you can when scouting

and speed up the the decision- making in your team I have observed this with my own eyes a lot of teams love discussing they love the decision making process they're they're like great if we talk we don't have to work and then they spend hours and hours and hours discussing where to send their scouting boats and it would have arrived two days ago if they had just sent it perfect gateway to the next problem most game of teams these days have multiple decision makers this is the multiple captains problem both captains are sent on a

scouting mission to find the perfect pool of fish and they both return with great news they both claimed that they just found the perfect spot and that the ship should follow and now we have a problem cuz now they will get stuck in Endless discussions and in Endless arguments the problem is humans suffer from a lot of biases one of them is zost fallacy so if you invest a lot of time into something then you do not want to throw that work away that is a very human thing to do and to want so both

of these captains are simultaneously suffering from sunk cost fallacy and this can get surprisingly emotional if you don't address it in the subtext there can easily appear something like I'm the better Captain you went in the wrong direction you are a verse Captain there are a couple of naive solutions for example one I can think of is let's do both if we can't agree then let's just do both let's meet in the middle and that is one of the worst things you can do it clutters up your game it leads to scope creep and in

the end nobody will be entirely happy with it anyway you can make the entire team vote but that does not solve the emotional aspect of this problem I've really only come across one good solution for this problem so far the two captains need to use the teleport ability switch places and continue the search from the other Captain's position for a bit this does two fantastic things firstly it removes the emotional aspect cause now both captains have invested an equal amount of time into each idea and secondly it brings A New Perspective to each of those

search branches making those ideas only better and better you will be surprised how much of a rational argument people can all of a sudden have before we decided on making Thro fall I made this prototype a cute little game about golfing flowers and I was convinced there was a lot of potential in this idea so we switched prototypes and Paul kept developing the idea a bit further turning it into something that looks like this and Paul reported back to me with some roadblocks he discovered that he thought would be very difficult to overcome so even

though I was very in love with the idea at this point this was not my baby anymore this was our baby and after talking it out I had to agree this was not going to work another thing that can be very useful when having m Captain is to split responsibilities a little bit for example you can have one captain that always has the last word on Art decisions and one captain that always has the final say in gameplay decisions in this case both of the captains can run their own experiments in their own domain they

send out their own scouting boats and they're not going to get into direct conflict as much also don't have so many captains you can also have some regular crew members crew members kind of need to act like mini captains they need to run a little bit of search in their own domains but where to steer the ship as a whole that is for the captains to decide the decision- making process needs to be fast and having too many decision makers leads to way too much friction here are some red flexs that tell you exactly when

your search algorithm is not working efficiently number one you never scrap any of your work cuz that means you don't even have any search branches at this point you're just guessing the best Direction and following your nose while this feels efficient and it feels like you're not throwing anything away it will lead you into a very poor local minimum I can already think of two cases where this happened to me one time with will you snail where I did not spend any time on the character controller I just made one character controller for my jump

and run game and just rolled with it I did not experiment with it whatsoever there was no search there I just winked it and built a fouryear project on top of it that is crazy the other time where this happened to me is with the economy in Throne fall I pretty much only had one or two attempts at balancing the economy in Throne fall and we pretty much ended up with an economy that is extremely snowb so either it will completely expon exponentially escape and you become way too powerful or you don't build any economy

and you become way too weak so it is extremely snowball in both directions and now it is too late to change it I feel like I would step on too many players foot if we make any big changes now cuz a lot of players like it the way it's now and it's you know it's kind of too late to make any adjustments so I just wish I had run more experiments with the economy of Throne fall fall early on to make it less snowball if you never scrap any of your work that means you're not

taking full advantage of the search space available to you and you're likely missing out on a lot of great opportunities to improve your game secondly you constantly scrap your work that means one of three things either you have commitment issues you have decision-making problems or your database is flawed and you are exploring in completely nonsensical directions thirdly you scrap your work way too late this is usually just a different symptom of the same problem you keep going in One Direction without any exploration and then suddenly you realize you ended up in a dead end before

making Throne fall we spend two months on making a card game just before scrapping the whole thing this can also happen when you don't measure correctly you make some noisy measurements and you waste two months of your time two months is the best case scenario if you're uncareful you can waste years fourth red flag your game failed you end up with no fish if that ever happens to you you should never think oh crap my game failed you should think oh crap my search algorithm failed what can I do to improve my search algorithm next

time there are only three main things to remember from this video you are running a search algorithm because design is a search algorithm you're searching for the best game possible if you want to Delight as many players as possible or maximize your Revenue then aim for fun appeal and a reasonable scope fun is for keeping players appeal is for gaining players and scope is for making sure you can finish the game being optimize your search cuz the success of your project depends on it