Malheardle

Summary

I made a variant of a popular internet minigame. In this post, I describe the story and reasoning behind some of the design choices in my variation.

What is an Anime Heardle??

Malheardle is a spinoff I made of Anime Heardle, which was a spinoff of Heardle, which was a spinoff of Wordle. You can play Malheardle here!

The main idea of Heardle is that the player has 6 tries to guess a mystery song. It starts by playing a 1 second clip of the song, but every time the player guesses wrong, they get to hear more of the song. Anime Heardle follows the same rules, but the songs are limited to anime opening and ending songs and the player must guess the anime.

I discovered Anime Heardle a while back, near the end of Covid, I found it a fun game to play with other people. Grabbing a few friends and trying to figure out the song/anime together, there are a lot of good experiences I've had with Anime Heardle:

• In general, I enjoyed the sensation of going down a detective rabbithole — identifying the artist/voice actor behind specific voices, tracing the music style back to specific bands, browsing through huge lists of shows to find the exact one that fits, and even tracking down specific memes we associated with the song we've heard. Doing all of this and eventually narrowing down the correct answer feels incredibly satisfying.
• There were also times where someone will take a random shot in the dark and actually hit, which is very exciting.
  • It is equally exciting when someones guessing something correctly off of a very, very tiny piece of information, like a single breathe at the start of the song.
• Often times people will get bent out of shape because they can recall hearing something, but not where it was from. Seeing this happen was fun because it raised the stakes of the game, making it more exciting if they end up getting it and funnier if they don't.
• Playing as a group, we've also discovered the social potential of the game. We would often react to unknown songs with sayings like "oh, I kinda like this", or "I hate it already", or, sometimes, "this show is definitely something X person would watch". Playing the game in a group transitioned very naturally into chatting about anime, about specific bands, artists, shows, or studios as relevant songs appeared in the game. These were chats that I enjoyed.

There are also some things that I didn't like about the various implementations of Anime Heardle that we've tried:

• Song and anime selection often felt biased towards the creator of the site. If you watch specific genres, you might not be able to guess anything if the creator of the site did not watch the same shows as you did.
• Some of the sites stopped getting updates 💀, so they just started reusing the same set of songs over and over, songs that we already knew.
• Sometimes, very niche shows will show up in the game and will be nearly impossible to guess outside of pure luck. It's exciting to get these right, but in general, these are not fun to play because you feel like you never had a chance.
• Sometimes, there were just not enough clues to warrant diving that deep into the detective rabbithole. With at most 16 seconds of audio, there could be nothing significant to go off of.
• The search function was usually wonky, you had to search for the anime name by typing in matching words exactly (no fuzzy search) and the results were ordered alphabetically (instead of by relevance). This was not really a big issue but was something that I would improve if I had the code.

Why Make My Own Anime Heardle?

One day, in the middle of an Anime Heardle session, I made a casual remark that the game should let us enter our MyAnimeList (MAL) usernames (MAL is an anime-related social media platform where users can compile all of the shows and movies they've seen into a publicly visible list). That way, the game can pick songs that at least one of us should know for sure. Later on, I realized that I could probably make just that. And so, I started playing around with the MAL official API, and, a few day later, I presented the first version of Malheardle to my friends.

I made Malheardle as a variant of Anime Heardle that improves on the main problems I had with it:

• Songs are now selected from the player's MAL anime list (a list of anime they've watched or plan to watch), which helps the player narrow down the potential answers and ensures they at least know of the anime they are trying to guess.
  • This also helps refresh the game with new songs overtime, as the players add more anime to their lists.
  • The game can load multiple players' anime lists at the same time, which allows the game to be better enjoyed with friends. This also adds the social element of guessing "whose list is this anime on?" (sometimes, "who's fault is it if we lose?").
  • There is a technical problem though: if the song is randomly selected from an unknown list of anime, how can we provide the game with the right audio files to play? Well, we can search for the song on YouTube and play the first result we get — which feels a little janky, but it's good enough to make for a fun game.
• Instead of only playing a short clip of the song and playing some more when the player guesses wrong, songs are now played through all the way, without pauses or limitations. This potentially controversial change takes away the iconic "heardle" aspect of the game, but when my friends and I played with it early on, we all said it was a better experiences. This change turns the game into more of a rush than a slow paced slew of detetive work.
  • Funnily enough, there is no actual deadline for making a guess, yet players seem to want to hurry and get the correct answer before the song finishes playing. If they don't get it by then, they will then shift into detetive mode, much like before (where they investigate potentially clues from the song and gradually narrow down the answer). I liked how this played out, so I kept it this way.
  • This change also limits the chance of players getting excited from guessing correctly with minimal information, since they won't be asked to stop and think before hearing more of the song, but I think the tradeoff is worthwhile. Players still seem to feel some incentive to guess as fast and as early as possible (to get the answer before the song ends). And, when playing with friends, the game becomes a bit of a race anyway.
• After each failed guess, the game now display additional information, including the MAL score and airing season of the anime guessed, as well as indication of how close that information was to that of the correct answer. This restores some the gradual clue revealing design that is iconic of Worldle and its spinoff games, giving players more and more information as they make guesses. This additional information also helps players make miracle comebacks when they do not recognize the song at all (which can be exciting).
  • In a group, the incorporation of the MAL scores into the core gameplay leads naturally into discussions of peoples' perception of the shows, how good certain shows are, which shows are better, etc.
• Finally, I made various aesthetic and quality of life improvements as I developed the game, like fuzzy search, colorblind-friendly palettes, and more information about the anime on the results screen.

Conclusion

Overall, I think Malheardle is one of my most successful projects, based on how complete it feels and how much enjoyment my friends and I get from playing it together. I think the critical factor in its success was how easy it was to playtest. Having experience in web dev, I was able to whip up a playable prototype quickly, and with how easy and fun it was to get into the game, we were able to (and very willing to) playtest it almost everyday. I never had a project that got playtested as early and as often as this one, and making Malheardle convinced me that I should really try to do that more often.