Lost City Of S

Lost City Of S looks like one of the best-produced entries reviewed on this shelf. A lone explorer in a pith helmet moves through a lush isometric jungle — palm groves, ancient stone monoliths, prowling gorillas — that later opens into a desert and a river crossing, all rendered in the same clean flat-vector style this author uses across a whole run of js13k entries. js13kgames.com's own judged results back up the eye test directly: 19th place out of the entire 2024 field for graphics, a genuinely strong result.

The touch controls work — confirmed by watching the world move

The core interaction is a virtual joystick: touch and hold to plant a stick, drag to set a direction, and the character walks that way. This was tested properly rather than assumed, because a drag gesture needs to be held and moving, not just tapped — two separate drag inputs (up, then right) were each confirmed to move the character by comparing full-screen screenshots before and after: the gorillas, rocks and tree layout were completely different each time, meaning the world had genuinely scrolled a meaningful distance rather than the character twitching in place. That's real, working touch-driven movement, not a control scheme that looks right in the source but does nothing on a real screen.

The twist that undoes it

The game's hook, per its own description, is that "a mishap happens here every 13 seconds." What that actually means, discovered only by reading through the competition page's own comment section, is that your movement controls silently reverse on a 13-second timer — up becomes down, with no visible warning. That would be a clever idea for a survival game built around memorizing a rhythm. In practice, it's a clarity failure serious enough that it shows up independently, described in almost the same words, across a dozen separate comments from the competition's own judges: "I didn't understand why... my movement speed suddenly would flip and instead of going up I would go down," "is mishap a control reversal when the timer runs out? ...very confusing," "not sure what the goal is." This isn't one confused player — it's a pattern serious enough that the game's own controls score (104th of the field) sits well behind its graphics score (19th), and several judges reported dying repeatedly without ever understanding what was killing their attempts. On top of that, the game is close to silent throughout — judged 155th of the field for audio — which several of the same comments call out directly.

To the author's credit, he's been direct about the other half of this in his own replies: the desktop mouse control was "really designed for touch input on mobile devices," left in by mistake rather than intended to be used with a mouse, which explains why several judges playing on desktop reported being unable to figure out movement at all.

Final verdict

The touch implementation itself passes the test this shelf runs: dragging genuinely moves the character, confirmed by watching the world change under it. But a shelf that promises "genuinely good" games has to weigh more than input plumbing, and a central mechanic this consistently misread by the game's own competition judges — not just casual players, but people used to evaluating jam games — is a real design problem that beautiful art doesn't cover for. Worth revisiting if a future version telegraphs the control-reversal clearly and adds even minimal sound; as shipped, it doesn't clear the bar.