Devlog — Odyssos
Snapshot
Odyssos is one of the first games I’ve actually developed and released for real players. It’s a browser incremental, so no Steam release circus—just get it in front of eyes and see what happens.
It still needs balancing, but it took the better part of a year to build, using GPT-3.5 + GPT-4 plus a decent chunk of custom/manual coding. It’s also one of the last projects where I did a lot of the manual coding myself, before AI got “good enough” that doing it all by hand stopped feeling necessary for something like this.
Timeline (how it actually happened)
- ~3 months: initial build push
- ~6 months: stepped away for other work/projects
- ~3 months: came back to finish and get it publishable
- ~1 week: final cleanup sprint, then release (because I could feel myself losing steam and I wanted to finish something)
Reception (the good + the annoying)
It got mixed reviews, and the negative ones were almost entirely about balancing. At the same time, it also got a lot of praise—people seemed to genuinely enjoy the gameplay, and I racked up a lot of logged playtime across a couple incremental game sites.
What I think I did right
The “gently unfolding” UI
A ton of work went into the UI and tutorial pacing—specifically: don’t dump a thousand menus/options on players immediately. I hate that, and I’m pretty sure most players do too, so the plan from day one was a UI that gradually unfolds instead of jumpscaring you with complexity. I think I pulled that off pretty well.
Actually finishing (rare event)
For my first game that I’ve released—hell, one of the first I’ve even pushed past the “20% done” zone—I’m legitimately proud of it.
What’s still broken / unfinished
Balancing (obvious) + math (worse)
The balancing problems aren’t just curve tuning—some issues are in the actual calculations. And yeah, a modern model (Opus 4.5 / GPT-5.2) would probably annihilate the cleanup and balancing work if I sat down and did it properly. The limiting factor is just me setting aside the time.
Player feedback pattern
Most feedback threads were basically: “Hey, this game explodes after X level.” I already knew about that, but it was still useful—and now I have real playtesting data to work from.
Architecture notes (or lack thereof)
I don’t think there’s anything especially notable architecturally here. This isn’t the “clever systems engineering” project—it’s the “make a real game and release it” project.
Next steps (when I stop procrastinating)
- Do the real balancing pass (curves and calculation fixes)
- Finish/polish until it’s actually “well received” across the board
- Port/import it into an Android app and maybe monetize it a bit (not the core goal, but still nice)
Why I care (the real reason)
The best part isn’t money. It’s the proof: I can make something that other people actually enjoy, and I’m not insane for thinking my ideas are fun.
You can play it at odyssos.io.