Game Jams
Game jams are development sprints that are limited by a (usually short) time period and are (usually) centered around a specific theme or technology. The time limit could be hours, days, or months. Many game jams focus on making video games, but tabletop games are also popular.
Examples include:
Fallen Jam, where participants created new adventures, expansions, and supplements for the Fallen RPG created by Perplexing Ruins.
Postcards from the Front, hosted by Hissy Cat Studio, challenges participants to create an entire wargame that fits on a 5”x7” postcard. The jam already has 32 submissions, including one from noted game designer David Thompson.
TTRPG Art Asset Jam is all about submitting art that can be used in game development, rather than developing new games.
The key elements of a game jam are:
Limited time (start and end dates)
Common theme (even if wide or vague)
The final product is a game (or game related tools/assets)
There are countless ways to host and organize game jams, but hosting a jam on itch.io is simple, making it a popular platform. As of March 2023, itch.io claims that almost 259,000 games have been created for jams hosted on itch.io.
So when I found a Kaggle dataset called Game Jams hosted on itch.io by Patrick Klein that was updated just a month ago, I knew I had to take a look!
The Dataset
The total dataset includes 10,807 game jams hosted on itch.io with end dates from September 2013 through May 2023. The start dates are not included, so it’s impossible to look at duration for each jam. Here’s what it does have:
Title
Number of participants
Number of submissions
End date
Host
Ranked (True/False)
Featured (True/False)
Not all jams in the dataset have end dates listed, but it’s easy to see that the number of game jams hosted each year on itch.io is increasing. That’s a 73% increase from 2019 to 2022!
The number of participants really jumped up over the same period, particularly from 2019 to 2020 (+115%). There’s now about a quarter million participants in game jams on itch.io each year. Note, however, that one person might enter multiple jams.
Looking for outliers
The first thing that jumps out are the outliers. Out of 10,806 game jams with participant counts, the IQR (the middle 50% of the data) is 10 - 56 participants per jam, but the mean is 87. This is because while the vast majority of jams have less than 300 participants, there are a handful with thousands.
The GMTK Game Jams from 2020 - 2022 each have over 18,000 participants and over 5,000 submissions. The highest is the GMTK Game Jam 2020 with 22,000 participants and 6,063 submissions! That is far from typical.
This makes it difficult to show the full distribution of things using a simple histogram or box plot. So most of the following plots strip out any data above the 95th percentile (e.g. jams with > 2,732 participants) to keep it simple. The data was not excluded, however, for any summary statistics (e.g. mean, IQR, etc.)
Most game jams only have a few participants and just a handful of submissions. We can plot submissions vs. participants to see if there’s any correlation.
Perhaps unsurprisingly, having more participants is correlated with having more submissions to the game jam. Jams get about 0.286 submissions per participant (e.g. ~3 submissions per 10 participants) with a typical range of 0.189 - 0.470.
As an example, the URBN_LGND.exe Jam currently has 145 participants and 17 entries, with a 0.117 submission/participation ratio. We could expect that to go up as we approach the deadline, ending with with about 41 submissions if it is typical. This might be a good measure of engagement when hosting a game jam.
Featured or not
Itch.io selects some game jams to be featured on the main jam page. This flag (featured vs. not) was included in the dataset, so we can see if being featured makes a difference!
Here’s a similar scatter plot of submissions vs. participant, but split by if it was featured or not:
At first glance, it looks like the submission per participant ratio is about the same, regardless if the jam is featured or not. Certainly the larger jams were more likely to be featured, which is even more clear in this violin plot:
Game Jam Titles
What’s in a name? Based on the word cloud above, some of the popular terms are: game, jam, weekly, summer, week, hour, winter, indie, horror, and Halloween. One that stood out was Trijam, which is a series of game jams where you make something playable in only three hours!
Regardless of if it was a featured jam or not, titles are usually about 23 characters long, or around 15-28 characters.
Ranked or not
First, it doesn’t seem that ranked jams get significantly more participants than unranked jams:
Now let’s look at submissions per participate for all combinations of ranked vs. unranked and featured vs. not featured:
Interestingly, it would seem that unranked game jams have a higher submission per participant ratio (Median = 0.333) than ranked game jams (Median = 0.258). Perhaps people are more inclined to submit something when it’s not judged?
Conclusions
Some things to think about:
Game jams are becoming increasingly popular on itch.io. Hosting a game jam might be a good way to build communities and drive engagement.
In general, participation per jam is low. You can expect about 22 people to join your jam, and maybe get 10 - 56 participants total. This is might be a good benchmark to see how your jam is doing.
Submissions per participant are typically low. Getting people to join is half the battle, while getting people to submit something is the other half. Getting about 0.286 submissions per participant (IQR: 0.189 - 0.470) might be good benchmark for your jam.
What do you look for in game jams? Have you hosted a game jam and have advice to share? Please leave a comment!
See you next week!
— E.P. 💀
That was a very interesting read. Thanks for digging into the data. I wish I would have seen this before writing my game jam article, I certainly would have quoted it!
The entries/participant data is very interesting. Personally, I won’t even join a jam until my submission is ready! I enjoy writing for jams, but I don’t like there being the stress that someone might be expecting me to finish a project.
I also like how you displayed the featured or not data. It makes sense to me that featured jams would get more people signing up that end up not submitting anything vs jams that get their participants through narrower channels (like a topic-specific discord server).