I spent three days building a combo multiplier system for Bubble Pop. It tracked consecutive matches, calculated exponential score bonuses, triggered screen effects at milestone thresholds — the whole package. When I finished, I felt like a genius. I showed it to Max and he said, "Cool, show me how it works." I couldnt. Because in three days of testing, I had never actually triggered a combo of more than three.
The system was designed for a theoretical player who would match colors in rapid succession, chaining combos like a professional Tetris player. Real players? They popped a few bubbles, paused, looked around, popped a few more. Average combo chain: 2.4. The elaborate animation I built for the 10-combo milestone played for exactly zero players during the entire beta. Zero.
The Sunk Cost That Almost Ruined the Game
Here's where it gets embarrassing. When I realized the combo system was useless, I didnt delete it. I doubled down. I spent another day making the combo windows larger, increasing the time limit, adding visual hints to encourage faster play. I was trying to make players behave the way my feature needed them to, instead of accepting that the feature was wrong.
The game got worse. The larger combo windows made normal play feel sluggish. The visual hints cluttered the screen. Players who were happily popping bubbles at their own pace now felt rushed. Play time dropped. Bounce rate went up. I had made the game strictly worse for everyone because I couldnt admit that my three-day project was a mistake.
Max finally sat me down and said something I still think about weekly: "If a feature needs the player to change how they play, it's not a feature — it's homework." I removed the combo system that night. Bubble Pop's retention rate went up 40% the next week.
What I Should Have Done Instead
Looking back, the warning signs were there from the start. On Day 1, when I sketched the combo system on paper, Max said "that seems complicated." I heard him but I didnt listen. On Day 2, when I showed him the first implementation, he said "who is this for?" I had an answer ready — "power players" — but I couldnt name a single person who fit that description.
I now have a checklist I run before building any feature:
- Can I name three real players who have asked for this?
- Does this feature work if the player ignores it?
- Would removing it make the game noticeably worse?
If the answer to any of these is no, the feature doesnt get built. It sounds obvious written down. It took me three wasted days and a nearly-broken game to learn it.