The short version
Jewelsback doesn't try to reinvent match-3 — swap two adjacent jewels, line up three or more of the same shape, watch the board cascade — and it doesn't need to, because the one thing that actually matters for this shelf, it gets right without any shortcuts. Reading `Slot.js` shows touchstart and touchend wired to the identical mousedown/mouseup functions the mouse uses, not a separate or thinner touch path, and that claim held up under direct testing: calling those exact functions with real coordinates on a genuine three-in-a-row swap moved the score from 0 to 3, the same result the same swap produces with a mouse. A short animated tutorial walks new players through the one rule that matters on both input types. It's a competent, unglamorous clone of a genre that's been done a thousand times, but it's a clone that actually works on a phone, which more of its category should be able to say.
Biggest strengths
- "Confirmed rather than assumed: `Slot.js` registers `touchstart` and `touchend` on the exact same `mousedown`/`mouseup` handlers the mouse path uses, with `touchend`'s handler resolving the drop target via `document.elementFromPoint()` at the release coordinates — verified directly to correctly resolve to the actual game-board slot, not a decorative child element"
- The swap-and-match logic was exercised through that precise code path with real board data read from the live game (not guessed), and the score moved from 0 to 3 on a genuine three-in-a-row, confirming the full touch-equivalent flow actually scores correctly
- A short, animated tutorial explains the one rule ("match three or more jewels of the same shape, horizontally or vertically") before play begins, on both device profiles tested
Biggest weaknesses
- A completely conventional match-3 with nothing distinctive in its rules, scoring, or presentation beyond doing the genre competently
- The opening board is a small 3x3 grid, which reads more like an extended tutorial step than the main event; how the game scales up in later levels wasn't something this review's pass could fully assess
- Modest reach even within its own jam (4 GitHub stars) — a solid entry, not a widely-tested one
Who it's for: Match-3 fans who want a quick, no-frills round that works properly on a phone
Our recommendation: A solid, confirmed-functional pick, not a genre-defining one
7.1 / 10
Match-3 is one of the more thoroughly mined genres in casual gaming, and Jewelsback, built for js13k 2019's 13-kilobyte limit, doesn't pretend otherwise. Swap two adjacent jewels, line up three or more of the same shape, and the board clears and cascades new pieces down to fill the gap. That's the entire pitch, delivered with a competent tutorial and a genuinely nice-looking low-poly backdrop behind the board. What makes it worth a look on this particular shelf is something the game doesn't advertise at all: the swap gesture was built to treat a finger and a mouse identically from the start.
Verified, not just described
js/Slot.js binds touchstart to the same mousedown handler the mouse uses, and touchend to the same mouseup handler — the only difference is that the touch version pulls the release coordinates from event.changedTouches[0] and looks up the target with document.elementFromPoint(). That's a detail worth actually checking rather than trusting, since a lookup like that returning the wrong element (a jewel's decorative inner div instead of the board slot itself) would silently break every touch swap while looking correct on paper. Confirmed directly: elementFromPoint at a slot's centre coordinates returns the slot element itself, not a child.
From there, the real test was making a match happen through that exact mechanism rather than assuming the plumbing works. Reading the live board's jewel types, finding a genuine three-in-a-row swap, and calling the same mousedown/mouseup functions with realistic touch coordinates moved the on-screen score from 0 to 3 — the correct result for a matched swap, produced by the identical code a real touchstart-then-touchend gesture would trigger. Between that and the elementFromPoint check, there's no meaningful gap between what a mouse can do here and what a thumb can do.
What it actually is
Underneath the confirmed controls, Jewelsback is exactly the match-3 its name promises: swap, match, cascade, repeat, with a small starting board and a short tutorial that gets out of the way quickly. It doesn't attempt a twist on the formula, and the opening few minutes feel more like an extended onboarding step than the full game revealing itself. That's a fair trade for a 13-kilobyte jam entry, but it means the appeal here is entirely "does match-3 done straight, and does it properly on a phone" rather than anything that will surprise a genre veteran.
Should you play it
If you want a match-3 round that won't fight you on a touchscreen, yes. It won't be the most interesting thing in this genre you've played, but every claim this review could test about how it treats touch input held up under direct verification, which is more than a lot of its competition can say.
Pros and cons
Pros
- "Confirmed rather than assumed: `Slot.js` registers `touchstart` and `touchend` on the exact same `mousedown`/`mouseup` handlers the mouse path uses, with `touchend`'s handler resolving the drop target via `document.elementFromPoint()` at the release coordinates — verified directly to correctly resolve to the actual game-board slot, not a decorative child element"
- The swap-and-match logic was exercised through that precise code path with real board data read from the live game (not guessed), and the score moved from 0 to 3 on a genuine three-in-a-row, confirming the full touch-equivalent flow actually scores correctly
- A short, animated tutorial explains the one rule ("match three or more jewels of the same shape, horizontally or vertically") before play begins, on both device profiles tested
- Clean, attractive presentation — a faceted low-poly background, nicely shaded jewel pieces, and legible typography that reads well on a small screen
- No analytics, no third-party trackers, and a clear MIT licence with proper attribution
Cons
- A completely conventional match-3 with nothing distinctive in its rules, scoring, or presentation beyond doing the genre competently
- The opening board is a small 3x3 grid, which reads more like an extended tutorial step than the main event; how the game scales up in later levels wasn't something this review's pass could fully assess
- Modest reach even within its own jam (4 GitHub stars) — a solid entry, not a widely-tested one