The middle of summer has come and gone, and yet another GMTK Game Jam is behind us. As always, it was bigger, it was better, it was all the things, and it crashed Itch.io (albeit only for a tiny bit this time). I also took part, as I have almost every year of the past eight.
The theme was “COUNTDOWN,” and this year I made a roguelike deck-builder game, loosely inspired by Balatro (2024), I suppose. I say “I suppose,” because I did not realise it during development. However, I have been playing plenty of Balatro the past few months, so it only makes sense.
The Count is Down has you assuming the role of a jester for a borderline bipolar medieval count, who may or may not chop your head off at the end of any given day. And while I do not believe it was my best and most original idea for a jam game, it appears it was one of the more successful ones. The jam had a grand total of 10,387 entries, and I placed as follows in each of the categories:
| Criteria | Rank | Score |
| Narrative | #1513 | 3.240 |
| Creativity | #2923 | 3.600 |
| Enjoyment | #2964 | 3.280 |
| Artwork | #4158 | 3.160 |
| Audio | #5589 | 2.560 |
Overall, I would say this was a good score, considering the number of games made and the “what went wrong” part that you will read below. So, let’s get straight into it.
What Went Right?
The jam started at 8:00 PM local time on Wednesday for me, so usually I do not start working right away. Instead, I spend the first evening thinking it over and figuring out what I want to make. This year, unlike previous ones, I landed on the idea immediately and ran with it.
Knowing exactly what I wanted to make meant that I had a very clear direction, and I didn’t spend that much time iterating on the core mechanics. Therefore, I had far more time I could use to code the systems I needed, which was more than welcome, because card games can be a pain to develop.
I also had some extra time, since I am out of work right now. So, I could spend the entirety of Thursday and Friday working on the game. And by the end of the day on Friday, I already had the majority of mechanics coded relatively bug-free.
Furthermore, I decided to do all of the art myself (except for the icons). This gave the game a visual style that was distinct and memorable, if not very artistically solid. I made the art by taking photos of myself in various poses and tracing them in a Krita vector layer. Then, made them into silhouettes with white outlines and linework. Overall, I believe it worked well, considering my utter lack of artistic talent.
The one thing that was absolutely crucial to my ratings not being chopped was that I had to meet the median number of votes during the voting period. Usually, I fail at this, so this year I made sure that all of the design is web-friendly. Having a web build, as well as a Windows build, meant that a lot more people were willing to give the game a chance and with the striking high-contrast capsule art, I managed to get the required number of votes to leave my raw rating unchanged.
In the end, I had a solid foundation of engaging mechanics and interesting art that drew people in, and I am quite happy with it.
What Went Wrong?
As always, a jam project has a lot of rough edges. It is a matter of sanding those out whether you make it to the top 100 or top 1,000. I should have done more sanding.
The mechanics were solid, and I worked very, very hard for three days straight. By the end of Saturday, I was absolutely knackered, but I had not playtested the game with anyone other than myself and a friend who played a late Friday build.
Therefore, I had several problems, the most glaring of which was the balance. The game was simply too punishing to play. You are always going to lose, sure, but most people could barely make it through day 3 before losing. And a card game lives or dies on its balance.
Furthermore, the UI was not only bog-standard (buttons, fonts, etc.), but its design was also not fleshed out enough. A lot of bugs reported in the comments were not bugs at all, but badly designed UI/UX that did not communicate the mechanics at work clearly enough.
There were also a few bugs that popped up that I didn’t have the energy to hunt down on Sunday, because I was simply burned out from working on the game and I just wanted to ship it.
So, on Sunday, just after noon, with eight hours left on the clock, I decided that it is what it is and I am done. Just before that, I added some card drawing animations and floating text that shows the effects of cards when played, which added some clarity during gameplay. However, the end-of-day screen and almost all other elements were extremely static, and they didn’t communicate what was happening at all.
Lessons Learned
I do not want to end this postmortem on a sour note, because I do believe that this is one of my better jam projects and I am proud of it. Therefore, I will not focus on a couple of lessons that I walked away with from the GMTK 2026 jam.
Firstly, stick to your guns. Once you pick a solid idea, run with it. Do not doubt it, do not overcomplicate it, keep the scope manageable and do not give in to the “you know what would be cool?” temptation. I did this very well this year, and I have been pretty solid in this regard in previous years, too.
Secondly, I should focus more on self-care. If I had taken some more time to relax and unwind at the end of the day, instead of crunching further, I may have had more energy to do something more on the last day, instead of stopping eight hours short. Of course, I may have also found myself with no time left, so it is definitely a balancing act.
And last, but by no means least, make a web build! Lower the barrier to entry as much as possible for jam games. There were over ten thousand games to play, so any unnecessary friction meant that people would immediately move on to the next game.
Well, that’s it for this year. Fingers crossed, next year will be even better.