Majotori is a quiz game with branching storylines. The game presents a character that makes a wish, and a witch appears before them and offers to make it come true if they win a game of trivia, but if they lose, something bad will happen instead. The ending of the story changes depending on whether you win or lose.
The title is a made up word formed from a contraption of the japanese words "majo" (witch) and both "toribia" (trivia) and "torihiki" (deal).
After making the usual mistake of trying to start one's game development career with one's way-too-ambitious dream game and wasting years on it with barely any progress, Majotori came as a desperate last attempt at releasing a videogame before being forced to give up and find another job.
It had to be small, it had to be good, and I had to be able to do it all by myself, but I barely knew how to code. I had previously relied on my gamedev partner Nosgoroth for that. I thought I could at least manage to make static images and text show up on the screen, so I decided on doing something similar to a visual novel, in which you only had to read text over images and click on buttons to make choices.
I framed it as the fantasy of being in a situation in which useless knowledge about your hobbies becomes critical. I thought that the context might be some sort of devil offering you a deal, with the catch that if you failed he would ruin your life. Searching for inspiration lead me to the instrumental version of Futurama's Robot Hell song, which immediately became a pilar for the game and the main reference for its soundtrack.
Since a devil sounded sorta generic, I thought of changing to Lariat the Witch. Lariat was based on the main antagonist of my previous game, a little girl that forced her classmates to play her games.
On November 15, 2015, I officially started working on the project. I made it in Unity solely because that was what Nosgoroth (my programmer partner on our previous attempt at making a game) was using on our previous game, although I had no idea how to use it.
Instead of looking up a tutorial, I just tried to figure it out. I looked for a menu that allowed me to create an image, and so the game was made showing images on a canvas and changing their sorting order to show the current screen on top; a completely wrong and highly inefficient way to do it, but it worked. I also wrote it all in a single script, since I didn't know how to pass data between scripts, and didn't see the reason to dive the code between scrips anyway.
Also, in this original version nothing moved at all; the whole game was made of static text and images.
I made the graphics in Inkscape using vector lines. If I was to make all the required pictures myself, it needed to be something quick and easy to draw. I came up with the style spontaneously trying to draw using straight lines to form basic shapes. It came out quite stylish.
I finished the game in 3 months and released it for free (pay what you want) on itch.io (Windows, Mac and Linux) on February 1, 2016, and a week later (08/02/2016) on the Play Store for Android, in English and Spanish.
I submitted the game to Steam Greenlight, the program Steam had before completely opening the platform to third-party publishing. The system required people to vote on games for them to be published. Majotori didn't get enough votes.
As the Steam Greenlight was starting to get shut down, a lot of games (maybe all?) got suddenly approved to be released. However, if the game had gotten disdain during its Greenlight campaign, the chances that it would get any success if it were to be published seemed narrow.
I made a small test cutting out Ava and Lariat from one of the static pictures and making them appear into the scene with a bounce-like animation. The result looked great and seemed easy and quick enough to make, so I decided to remake the whole game to make it more visually appealing (04/09/2016). This time I would also use Unity properly and divide the code into multiple scripts. I also remade the whole UI.
For the next few years after releasing it, I kept updating the game from time to time to fix reported bugs and add a few extra questions.
The result was way more polished and better looking. It took me 6 months to remake the whole thing and I finally released the new version on Steam, itch.io, Play Store and App Store on March 28, 2017. The App Store version, built by Nosgoroth, initially took up 2GB of space until we figured out how to make a compressed build.
I initially planned having a single main character, Ava (short for "Avatar"), that would start as a baby and have a branching point at each stage of her life. However, that soon proved impractical as the branching grew exponentially. Also, for subsequent runs, you would have to repeat every chapter up to the point in which you went a different route. So I changed it to be about multiple different characters with a few branching scenarios each. Ava remained the character with the most chapters, the longest route covering her whole life from infant to elderly, as I had already thought about most of them.
As I was writing the main characters, and naming each one with a different initial letter to make the name less confusing, I decided to close it at 26 characters, covering the whole alphabet. To make it more interesting and varied, some characters' storylines would spawn from others', so the game is usually finished without seeing all 26 of them.
I made the stories purposely short and easy to tell so they would flow better with the quiz gameplay. I also pursued both mundane situations and tropes or known stories. Mundane situations felt relatable and worked well with the fantasy of getting a wish fulfilled by your useless knowledge, and the magical alteration of tropes or known stories was also an amusing implementation of it.
In the original version, at the end of a character's chapter, the story of another random character would begin or continue. Right at the end of development, pushed by some misguided feedback and fears I don't quite remember, I added some stupid progression block. If I remember correctly, every 5 chapters or so, Lariat would challenge you (the player) and only if you won, new characters would be added to the pool of possible new chapters. Also, the game had no end, I think; you just played until you unlocked every character and then you could keep playing forever with a random character continuing or restarting their story.
When remaking the game, I ditched that system and added a character selection screen so you could choose whether you wanted to continue the story of the last character or jump to a new one, while also teasing a third character. This made the game more engaging and made it feel more structured. However, I think not knowing what story would come next and being surprised by the reappearance of previous characters was a neat feeling that got lost.
Some of the characters and storylines are obvious references to popular culture or other works such as Romeo and Juliet, Indiana Jones, Beauty and the Beast or Sword Art Online.
One of the still-nameless characters reminded me of Woolie Woolz, a youtuber at the time part of the Super Best Friends Play channel, which I followed. Since I still hadn't named anyone starting with W, and his storyline started with him wanting to make internet videos, I named him Woolie. I made a further reference by the character dying his hair green if he lost the first chapter, as Woolie had dyed his hair green.
Fatima was made after Conchita Wurst, who had won the Eurovision contest a few years prior.
Oliver was made after my dog Trasto, who had died recently.
Kony was made after Kony, a bird-loving girl I had met during development that showed interest and offered to help me test the game. She later became part of Majorariatto.
What makes a good trivia question
When studying what makes casual trivia games fun, I found that they are a weird kind of game. They seem to mainly consist of displaying your knowledge to other players and/or proving that you know more than they do. You have a good or a bad time depending on whether you know the answers to the questions. This is weird because it's actually more a test than a game, and there's not much you can actually do to change the outcome during the game, as you either know the answers beforehand, or you don't. The fun and the outcome require the questions to cater to your current knowledge.
Having most questions be too obvious makes the game uninteresting (everyone knows the answer, you can't display knowledge), while having them be too hard makes it frustrating (no one knows the answer, no one can display their knowledge, feels unfair). A good balance is difficult, as it depends on the specific knowledge of the players.
Still, single-player trivia games can also be entertaining, even though you are not displaying your knowledge to anyone.
Is there a way to make good questions? Is there something more to it than whether the player knows the answer?
"Who won the Emily Wilson Award in 1973?"
This is the worst kind of question; you don't even know what it's talking about and could never guess the answer. What even is the Emily Wilson Award*? Who is supposed to remember who won it 50 years ago? And who even cares? A frustrating question.
In my experience, most trivial board games and online trivia quizzes feel often like this.
* I made it up for this example.
"How many legs does a spider have?"
This one, while not infuriating, is boring. You need to be particularly ignorant not to know the answer, so it's almost like no question was asked at all, but still preferable to the previous one.
"What village is Naruto from?"
This one is the worst of both worlds. For people familiar with Naruto it's way too easy, but for the rest of the world it's basically impossible to guess. It satisfies no one. Questions about a particular work or franchise should either test fans' knowledge or be so basic that just having heard about the work allows you to have a chance.
"How many planets are there in our solar system?"
This one, while still pretty basic, you probably don't know the answer to from the top of your head, and need to make a mental count. This makes it more interesting as there was some reasoning involved. This makes you feel somewhat clever, to a degree that depends on how hard did you find it to recall all the planets in our solar system.
"How many hearts does an octopus have?"
Chances are you don't know the answer to this question, but it's still sorta interesting. You know what an octopus is, and you know most animals have one heart, but octopussy are weird animals, and if the question is being asked at all it's likely because the answer is not 1, but either 0 or 2-4 (while 5 is not impossible, it sounds like way too many hearts).
Even if your answer is wrong, you had some room to reason and try to deduce the answer. You didn't just pick a random number, there was some thought behind it. That makes the question interesting.
"What is the capital of Portugal?"
This is your average trivia question. It's asking something a mildly educated person would be expected to know, yet not everyone does. Answering it correctly makes you feel knowledgeable. A good question, but not that exciting, since most people would either know the answer or not. You only have room to reason it if you know more than one important portuguese city but are not sure which one is the capital.
"What color is the first 'G' in the Google logo?"
Finally, this one is the perfect trivia question. Chances are you saw the logo dozens of times today, and yet you can't immediately recall the answer. You need to reason based on what colors seem to fit your mental image of the logo, and you probably won't have total confidence in your answer.
It's a question that everyone should know the answer to, yet few people do with confidence, while also leaving room for you to deduce the answer. It's risky and exciting, it's fair, makes you think, and makes you feel clever for getting it right, even if you are playing by yourself.
Sadly, you need a ton of questions for a trivia game (about a thousand for Majotori), perfect questions are mainly unicorns, and good questions take time to craft. I couldn't make every question a great question, but I could use this knowledge to significantly rise the overall quality and fun of them.
There's a trick I used a lot, which is to ask a question about the available answers.
"In which of these video games can you ride a horse?"
If the game offers 4 possible answers, there are 4 chances for you to have some info to work with. Even if you only know one of the games, you can either know that's the correct question or know that it's not and discard it, allowing you to make a more informed guess.
One more thing I dislike about the usual trivia question is how they tend to be about stuff surrounding media instead of the media itself. Who directed it? How many awards did it win? Who's that actor married to? You can't answer any of that by just having seen the movie. On writing the questions, I focused a lot more on asking about the works themselves.
When previously making a trivia game to play with my friends, I established that 4 was the ideal amount of possible answers. 3 are too few, having a high chance for you to get the correct answer by sheer luck, but 5 are too many, making it bothersome to have to read them all. 4 is a great balance.
Majotori, however, is a single-player game, so players getting lucky is not nearly as detrimental to the game as it is in a multi-player competition. It might have actually been beneficial, because it rises the chances of players feeling good about winning, but I didn't consider it at the time and went with 4.
I never really considered the possibility of letting the user input the answer, as that carries a lot of problems (having to guess every possible way a player might input an answer that could be considered valid, frustration over mistyping the answer...), and makes the game slower and the implementation more complicated.
The number of questions per round is 9. I thought the game worked better with an odd number of questions (so either you got most of them right or most of them wrong, no ties) and 10 felt a bit too long anyway. I could have also gone with 7, but 9 made the game (which was already going to be fairly short) last a bit longer, so I went with 9.
I also made it so a roulette would decide whether you win or lose, each possibility in the roulette being one of your answers, so the more questions you got right, the better your chances.
There were 3 reasons for this: The first one was about consistency between rounds. If I made it so that you needed 5 out of 9 correct answers, some rounds would be almost double the length of others, depending on how many mistakes you made.
The second one was about helping the player embrace failure, reducing their possible frustration. Trivia games are inherently unfair; you can get a batch of questions that are easier or harder for you. You can risk it and get it right, or risk it and get it wrong. Also, gamers would consider the outcome of the story as punishment or reward, so they would feel it unfair that they got a bad ending because they were unlucky with the questions.
I purposely made Majotori so that a bad ending is more like a different route than a punishment. Emphasizing the randomness of the game would help players embrace that the outcome was not totally under their control, so it's OK to get a bad ending. The roulette made that pretty clear, since even with 8 correct answers you still had a chance to get a bad ending, and that's obviously just bad luck. It worked but most people, but not all of them.
Finally, it added tension, surprise, and most importantly, hope. Even if you get a single correct answer, you still have a chance. It's a clever system, since the amount of randomness is determined by the amount of correct answers. You can always get all of them right and secure the result.
Occasionally, the roulette will play a trick on the player by going backward and choosing the previous result or keeping spinning so it lands on the next one. However, the game never actually cheats and it's not rigged in any way; the result is always decided beforehand by picking one of the answers at random. Still, in order to generate more pleasant surprises than frustration, there's double the chance that one of these tricks will play by "changing" a bad outcome into a good one than vice versa.
If I had to write all the questions, they had to be about things I knew about. Also, why is literature considered a perfectly acceptable topic, but anime or video games too quirky and niche? Not on my game!
I decided to focus the questions mainly on audiovisuals, which I considered to be pretty much "general culture" for a younger audience. The categories were video games, cinema, animation and miscellaneous, the later being the actual "general culture" category (literature, history, geography...), and also a wildcard to add any kind of question that didn't fit into the other categories.
Anime was a risky one since I know it's not as mainstream as the rest. I mixed it with wester animation to broaden its scope and make it a bit less niche.
I thought anime ought to be as valid a category as any other; ignorance about it being no lesser than ignorance about literature. I also internally defended the decision by appealing to the game's fantasy: If a nerdy witch offered you a trivia game, she would ask about anime and you would have to deal with it. I didn't even want to allow for any customization, but eventually decided it was better to let people customize the categories a bit, letting the player reduce the frequency of any of them.
I also eventually realized and accepted that I had way too much anime knowledge and the questions tended to be too obscure even to casual anime enjoyers. To fix it, I removed some of the hardcorest questions and added a lot of questions about the most popular anime.
After the launch of the game, I made an update to allow you to play using custom questions by adding a txt file with the proper formatting on the game's folder. It has probably been used by just a handful of people, but I thought it was nice to have.
This was my first time having to deal with hiring someone to do work for me. I looked for video game musicians on the internet and found a post by composer Esther García announcing her services on some forum. I listened to her portfolio and liked some of the songs better than those of the rest of the composers I had found, so I contacted her. I decided on asking for a single piece with multiple variations so it would be cheaper and hopefully not too repetitive.
Even though she advertised video game composition as part of her services, it looked to me like she hadn't really worked on one before; she seemed to be used to producing a mostly final piece from an initial description, but I'm very particular with what I like, so constant revisions were going to be unavoidable. I had very little money and was very risk-averse, so I was filled with anxiety about the prospect of having to pay for music I didn't like because I didn't have enough money to keep asking for revisions. I wanted to work very incrementally to reduce costs and make changes easier, but she either didn't understand my request or didn't agree with that way of working and sent me mostly finished pieces from the start. This created plenty of stress and friction.
At one point, after requesting a revision, she told me that I didn't know what I wanted. I took that badly because I knew exactly what I wanted: something that sounded like Futurama's Robot Hell song, which I sent her as reference, but I understand that I was not paying enough for her to have to deal with my detailed expectations. Thankfully, I eventually got something I was satisfied with.
The original version of the game came out with a single track. People loved it and it worked just well enough to last through the whole game, but having it repeat constantly was clearly one of the weaker points of the game.
For the remaked version, I contacted Ether again to make a few more tracks. Somehow I had a clear hearing of what I wanted for the character selection track, so I wrote the melody myself and sent it to her. Kony had the brilliant idea to make a special version of the main theme with dog barks for Oliver's story. The reference for the credits track was something from Splatoon, but I don't remember which one ("Ink Me Up!"?). The refence for Tsubasa's song was Tsubasa wo kudasai, which was also the inspiration for Tsubasa's story.
Other design considerations
Being against the idea of video games being a set of content to be consumed, I didn't want players to think they had to go through every single route; in fact, I wanted to discourage that idea, as going through the game over and over for completionist purposes would be a tedious experience.
Apart from not encouraging it by making the routes hard to parse and not giving the player any aid, I also made some routes variate at random. Kony's route, for example, features 1 of 3 possible injured animals, and Queralt makes 1 of 2 possible different wishes.
Inspired by my experience using video games to learn English, I added a way to instantly translate the in-game text from Spanish to English or vice versa, by pressing the "T" key or selecting the "quick translation" menu button. I'm pretty sure this feature has been barely used.
Although the game received external derision for its poor visuals and basic gameplay, those who played liked it a lot. That's what pushed me to work on remaking the game.
The original version released for free ("pay what you want") on itch.io (Windows, Mac and Linux) on February 1, 2016, and a week later (08/02/2016) on the Play Store for Android, in English and Spanish. Also on GameJolt, although I removed it from there eventually.
Downloads (first 6 months):
Google Play: 55
There were 11 paid downloads on itch.io from friends, for a total of $69.30 (gross).
I paid 240€ for the music.
On Steam Greenlight, it got 297 "yes" votes (31%) and 651 "no" votes (69%).
A few very small youtubers covered it.
Steam sales (v2 remake)
The game released with 289 wishlists on Steam.
Day 1: 106
Month 1: ~1.300
Trimester 1: ~3.000
Year 1: ~7.700
Until 2023 (6 years): ~35,600
Regional (until 2023): Argentina 18%*, Spain 15%, United States 14%, Mexico 9%, Chile 8%, Russian Federation 4%
*Most likely due to how cheap the price was, based on Steam's recommendation.
Gross revenue (until 2023): $76,100
Reviews (2023): Overwhelmingly positive (1.441), 97% positive.
Google Play (v2 remake)
Gross revenue: ~25.000€
Regional: United States 23%, Mexico 21%, Chile 12%, Spain 11%, Brazil 4%
*It has been temporarily free to participate in some promotions
On 2021 it joined Google Play Pass, and up until 2023 it's made ~17.500€ (net).
App store (v2 remake)
Net revenue: ~4.600€
Regional: United States, Spain, Mexico, United Kingdom
itch.io (v2 remake)
Gross revenue: ~$320
Net revenue: ~78.000€
I was discouraged by its first-day Steam sales, and my understanding at the time was that launch day should be the one where most of the profit is done, and then it's a steep downhill from there. However, the game maintained an average of 20 copies sold per day, which made for a small monthly wage.
The game slowly grew its popularity among small Spanish youtube creators, which helped maintain its sales. In 2020, top Spanish streamer Rubius played it live, widening its popularity and reach.
For a first game, and made relatively fast, I think it came out pretty good. The premise is novel, the questions are good, and the art style is quite compelling for how simple it is.
The game was (and continues to be) a financial success.
I think the overall writing and the plot of the stories are weak, and there's a good bunch of them that make me cringe. While the music is good, a single repeating track is not enough to carry the whole game comfortably. Also, I made a quick and poor job with a lot of the illustrations.
Overall, I think it's a pretty fun little game that executes its premise well enough.