Birb World Tour: Royale is now available online!

Some notes on a post-jam game project 🐦

March 2026

Remember that game with birds? This is him now. Feel old yet?

Similar to the 2024 Jam version, you race against 99 other birds along a long racecourse, with powerups to collect. Unlike the Jam version, levels are now procedurally generated, and there’s many real-world bird skins to unlock. I’ve also added a tutorial level, and a practise room level without a timer or opponents.

If this is your type of thing, I’d love for you to give it a go and send me any feedback you have. If you’re more interested in the technical notes, carry on reading.

Notes about the development

Game design

In my head, the game has always been essentially a blend of modern 3D Sonic platformers, Mario Kart, and Fall Guys. (The ‘World Tour’ aspect takes particular inspiration from Sonic Unleashed.) This means I didn’t have to worry too much about core gameplay design, and so spent more time on art & level generation than core gameplay. This means that broad aspects of the design (character controller, UI, etc) are fairly unoriginal, but as a whole feels somewhat novel.

Since the jam version, I’ve actually ‘un-implemented’ some of the stacking upgrades that could be purchased in the Lobby. The new procedural levels can be difficult to navigate, so I’ve tuned the character controller’s parameters to suit the levels - which the upgrades would interfere with. I’d like to either re-implement them with less over-powered multipliers, or add some other way of spending money that doesn’t impact the character controller, which also isn’t solely Cosmetics-based.

Graphics

The game’s models & textures are low-poly & low-resolution, so I can worry less about being a bad artist 😛. It also means the game can run well on cheaper hardware (i.e. my laptop). To lean into the retro aesthetic further, I removed shadows & render the game at 0.5x resolution. I don’t think I’ll go any further on retro effects, in case it risks impacting the gameplay.

Interestingly, I experimented with ‘immersive’ weather effects like rain, snow & confetti, but they introduced too much visual noise to be able to ‘read’ the upcoming terrain. Perhaps I can re-introduce them with smarter screen placement?

Music

The soundtrack is mostly drum-n-bass, because I enjoy DnB! And also because I can’t play any rock instruments well enough.

Four in-race tracks & two lobby tracks were made during the Jam, in about 30mins per track. After switching laptops, I lost most of the effects, so I had to remake them in order to re-export. The game now has 12 tracks, and I’m sticking to roughly 1-4 hours per track - it’s a very refreshing way to start & finish new tracks!

‘Getting it done’

I started regularly working on the post-jam version in October 2025. At the height of development, I was putting in 1-2 hours of ‘work’ into the game most days, often after doing my normal job (also in front of screens… eek). I think by having a clear image of what you want the game to look like, and a decomposition of the required elements into small, 1-hour-ish tasks, I was able to make measurable progress every day. Having a mix of ‘artsy’ tasks (such as making a 3D model or a piece of music) and ‘technical’ tasks (adding a new C# script or Unity prefab) was also a nice way to stay enthusiastic about the work.

Regular builds/exports also helps. It’s nice enough to play your game from the Editor, but getting it into a distributable state lets you share it with other people and feel like

AI assistance

During development of the Jam Version in 2024, I had GitHub Copilot’s code completion enabled in my IDE. I have not been using any Copilot features since resuming the project in late 2025.

Why? With the Jam’s time pressure removed, I felt like AI assistance was sapping the fun out of making a game, and wasn’t giving much in return (especially when it tried to do vector math). Given I already know most of the core Unity APIs that I need to use, and the official docs are pretty good for the ones I don’t remember, it just didn’t seem beneficial.

If I entered another game jam, it’s a toss-up for whether I’d use AI to help with the code. Sloppy code design isn’t a huge concern on such short-lived projects. I’m capable of making playable Jam games without Copilot, but I’m not sure I want to spent the time hurriedly writing lots of (mostly boilerplate) game code in a weekend. It might be that a comprehensive toolkit of pre-made game scripts would be more beneficial/enjoyable.

I don’t use any GenAI tools for art or music, for similar reasons.

Asset packs FTW!

Human-made asset packs are very helpful and encouraged for hobby projects & jam games, particularly when you don’t feel like re-designing boilerplate assets that every game has in some capacity. BWTR uses a small handful of assets from Kenney - mainly Interface Sounds, Input Prompt icons and panels from the UI Pack.

What I’d like to add to the game next


As always, I thoroughly recommend that you try making a game of your own, whether in Unity, Godot, or any engine of your choice. Game jams are a good way of getting something started & finished in a short amount of time. I’ve enjoyed participating in Ludum Dare, particularly because of how it encourages giving & receiving feedback.

If you try out the game and have feedback, or want to know more about anything in this article, let me know!

Sam


Return to index.