Gamemaker’s Toolkit is a fantastic educational YouTube channel for anyone interested in the design of games. In 2017 GMTK hosted it’s first GameJam, which ended up being the most popular jam in itch.io history. Every year since it’s been even bigger. This was the third GMTK Jam, but our first time entering, and first time doing one that was 48 hours.
We wanted to do something a bit different, breaking the formula of puzzle games that we always made in the past. The idea of making a dungeon crawler was quite interesting, so with that in mind we awaited the theme. The theme was “Only One”, which was to take something that normally you have many of, and change it so you only have one. Examples given were a shooter with one bullet, or a game with only one enemy. We spent a few hours trying to come up with a way to make this work with a dungeon crawler.
We wanted to have a game where you’d be presented with multiple options for which party member to add, but you could only pick one. The game would then branch off depending on your choices, leading to a different set of rooms each time you played. Each room would have multiple ways to solve depending on the party you’d picked. We’d have some control over that since we knew which rooms the player must have come through to get to the current one.
We wanted to push the “Only One” aspect even further though, in relation to the party members you could have. Normally in a dungeon crawler the character might be able to attack, and block, or use magic. We decided each character has only one ability. That meant that one class could ONLY attack, and another class could ONLY jump. Not only that, but each character would only be able to use their ability once (per room). As a final touch we made the characters only one colour, emulating a 1-bit art style.
We tried to plan out the rooms in a branching tree design. It would mean from any one room we would know exactly which rooms the player had gone through, because there was only one possible route there. It would also mean though that every extra level of depth we adding would require doubling the total number of rooms. This quickly became unmanageable and we had to make the tree converge at certain points, especially towards the end.
The trouble with the doubling room problem is that the further it goes, the less of the game the player will see (as a percentage). It’s not an efficient use of time in a GameJam because we’d end up making 10x as much content as any one player would see. Even as it already is, a single play-through misses half of the content. We could have resolved this by having the choices every few rooms, rather than every single room. This would also give us the chance to make easier rooms that introduce the players to concepts, rather than expecting them to know advanced information straight away.
Unfortunately, due to lack of time, we didn’t manage to make the final room work for all possible party loadouts. Some players got stuck on this last room because they didn’t have enough combat classes, and others breezed through it because they had a single rogue. We left these levels until the last minute and the overall game suffered because of it.
At the end of the 48 hours we kinda realised that despite trying otherwise, we’d ended up making a puzzle game again. We had started with a dungeon crawler template, and turned it into a puzzle game by making it about taking your time and thinking a lot. It was different enough from our previous puzzle games, but much like the game itself, we’d taken a different route and ended up in the exact same place at the end of it.
In the days following the GameJam, Mark Brown (the host of the Jam) was streaming entries. We added our game to his “to play” list that he was randomly picking from, but with over 300 games on the list we didn’t expect to show up. Imagine our surprise when our number got picked just a few minutes later. He played it all the way through, really liked it, and rated it a full 15 stars out of 15 stars. Happy noises were made.
There were over 2600 entries to this GameJam, and after getting 113 votes we ranked #293. We felt like we did a pretty good job, and the 4.0 stars we got as our score I think is more than fair, but it’s a testament to just how incredible so many of the submissions were that over 300 of them were higher rated than 4.0. We played 100 games over the next week and gave feedback – it was totally exhausting and totally awesome.
You can try it out for yourself from the web browser.