This is the best-written game in the file. The case is a disappearance that turns out to be about a debt, the town is drawn through eleven people who all know each other too well, and the deduction system trusts you to notice things rather than highlighting them.
It is also, structurally, a series of conversations. And every one of them loads.
The rhythm problem
Three seconds is nothing on its own. What makes it worth a review is where it falls: between the end of one conversation and the beginning of the next, which in this game is exactly the point where the player is holding a new piece of information and wants to take it somewhere.
You leave the harbourmaster with a name. You walk twelve feet to the tavern. The screen fades, holds for three seconds, and fades back, and in that interval the momentum of having just learned something goes flat. The information survives; the urgency does not.
| Measurement | Harbour Notes | File median |
|---|---|---|
| Cold boot to first input | 19s | 34s |
| Median scene transition | 3.1s | Not usually measured |
| Transitions per playthrough | 404 | Not usually counted |
| Total transition time | 21 min | Not comparable |
| Case file open | 1.4s | 2.6s |
Counting four hundred
We logged every transition, which is the reason this review took three weeks rather than one. Four hundred and four in a single completion, including revisits, and the distribution is heavily weighted toward the middle act where the game expects you to shuttle between four locations chasing a contradiction.
One stretch of about forty minutes contained thirty-one transitions. That is roughly a hundred seconds of fade in the densest, most interesting part of the case, and it is the only part of this game we found tiring.
A short wait in the wrong place costs more than a long wait in the right one. Three seconds between conversations is worse than thirty seconds at launch, because only one of them interrupts a thought.
Streaming ahead
The fix is well understood and this game is unusually well suited to it. There are never more than three plausible destinations from any location, the scenes are small, and the player telegraphs their intent by walking. Loading the two most likely next scenes while the current conversation is still running would remove almost all of this.
We are not guessing at feasibility: two other games in this file do precisely that, one of them with larger scenes, and their transitions are imperceptible. This is a choice that was never revisited rather than a limit anybody hit.
The case itself
Excellent, and the score reflects it. The deduction board is legible, the contradictions are genuinely findable without a guide, and there is a moment in the fourth chapter involving a misremembered date that is the best single piece of writing we have encountered in this genre in years.
The ending is slightly weaker than the middle — it explains one thing too many — and there is a chapter where the game withholds a document for pacing reasons in a way that feels like the author rather than the world. Neither is enough to matter much.
| Element | Assessment |
|---|---|
| Writing and case design | Best in the file |
| Scene transitions | 3.1s each, 404 of them |
| Deduction board | Legible, no hand-holding |
| Ending | Explains one thing too many |
| Options and rebinding | Complete |
Where this leaves it
Recommended without hesitation, and the timing complaint is a design argument rather than a warning. The lesson we take from it is that the placement of a wait matters more than its length: twenty-one minutes spread across four hundred moments of interrupted thought did more damage here than a single minute at launch would have done.