Watermelon Game

Origin

When Suika Game took off on the Nintendo Switch, I wanted to play it but didn't own a Switch. Steam had no alternative at the time, and the mobile versions were cluttered with ads. So I decided to build my own version, and also add my own ideas to it.

Core loop

There are 10 fruits, each a different size. The player drops fruits into a container, and when two identical fruits touch, they merge into the next, slightly bigger fruit. This continues up to the watermelon. When two watermelons merge, they disappear and free up space in the container.

Every drop is a small physics puzzle: where will it roll, what will it push, and will it set off a merge?

Key design decisions

I kept the core idea but changed several things to make runs longer, fairer and more strategic:

• 10 fruits instead of 11: with 11 fruits, the extra size step made runs too short and too hard. One fewer step keeps the challenge while letting players reach the satisfying watermelon merges more often.
• A 5-second grace countdown: in the original, touching the top line ends the game instantly. Fruits sometimes bounce up there unexpectedly, which felt random and unfair. In my version a countdown starts instead, and if the player brings the stack back below the line in time, the game continues.
• Three skills with a rising cost:
  • Power Shot: fires the fruit with force to squeeze it through tight gaps or rearrange the container.
  • +1 Up: evolves whichever fruit the dropped fruit touches, even without a match.
  • Destroy: removes the fruit the dropped fruit touches, to make space.
  Skills cost points earned through merging, and each use raises its cost by 10%. Without that increase, players could spam skills with no real trade-off. With it, skills become strategic decisions, and skilled players can extend their runs significantly by timing them well.
• Combo counter: a well-built stack can trigger chain merges from a single drop. For example, with a lemon at the bottom, a strawberry above it and a cherry on top, dropping a cherry creates a strawberry, which merges into a lemon, which merges again. Merges in quick succession show a combo counter at the side. It has no gameplay advantage; it's a visual feedback that rewards smart stacking.
• Preview reward: the original only shows the next fruit. Merging two watermelons unlocks a preview of the fruit after next for 5 minutes. It's a strong advantage, but it's temporary, which adds a sense of urgency.
• Golden fruit: on rare occasions, a dropped fruit turns golden and destroys every fruit it touches on the way down. It has saved more than one of my runs from a near-certain loss. Players can also get a weaker version by using Power Shot to knock a fruit out the top of the screen: it returns as a golden fruit that destroys the first fruit it hits.

Multiplayer

In PvP, all the single-player rules apply, but each merge also fills a meter. When it's full, the player can drop a stone fruit into the opponent's container. Stone fruits can't be destroyed or merged and act as a permanent obstacle, which turns a calm puzzle into a direct duel.

Challenges & lessons learned

1. The art update that players rejected

I drew the original sprites myself in Paint, so they looked like "programmer art". I later bought professional art assets and released a visual update, and the review score dropped by about 5 percentage points. Most players preferred the old look, so I reverted it.

Lesson: first impressions set expectations. Players had chosen the game for its mechanics and its simple, recognizable style, which also made it stand out among more polished-looking alternatives. The new art wasn't the game they had signed up for, even if it looked "better".

2. Grindy achievements

I designed the achievements as something players would unlock naturally over a long time. But many Steam players care a lot about achievements, and some skip games whose achievements look too grindy. I couldn't remove them afterwards, because some players had already earned them and removing them would be unfair to those players. Some reviews still mention them today.

Lesson: achievements are part of a game's design and its store presentation, not just a bonus. For my next games I will plan them from the start, with the audience in mind.

Player feedback

Player reviews directly shaped the game: I reverted the art update as soon as the negative reviews came in, and the response to the achievements changed how I design progression and rewards. The game holds a Very Positive rating (91%) across more than 1,000 Steam reviews.

Links