Return of the Obra Dinn: a detective story that takes answers three at a time
Lucas Pope, 2018. Sixty people, one ship, and a verification rule that holds the entire structure up.
score
Return of the Obra Dinn was made by Lucas Pope, the same person who made Papers, Please. It released on 18 October 2018 for Windows and macOS, and exactly one year later to the day on Switch, PlayStation 4 and Xbox One. It took the Independent Games Festival's Grand Prize, the Game Developers Choice award for narrative, two BAFTAs and The Game Awards prize for art direction.
You are an insurance adjuster for the East India Company. In 1807 the merchant ship Obra Dinn, missing for five years, drifts into port without a living soul aboard. Your job is to complete the insurance report: sixty names, sixty fates, and for each of the dead a cause and a culprit.
The instrument
You carry a watch which, held to human remains, moves you to the moment of that person's death — not to a scene, but to a single frozen frame. You walk around the freeze-frame studying faces, postures, hands, the line of a gunshot, tattoos, the cut of a coat. Sound precedes it and gives you a few lines of speech: sometimes an accent, sometimes a name, sometimes nothing usable.
There are no further hints. There is a crew manifest with sixty surnames, a group drawing in which every face is numbered, and your notebook. Everything else you derive: this man always stands beside that one, so they share quarters; this one has a Formosan earring and speaks Portuguese; there is exactly one Portuguese surname on the list.
The rule of three
The game verifies answers only in threes. Until you have three fates right at once it says nothing: it does not tell you a name is wrong, it does not highlight the error. That is the central design decision of the whole thing, and it solves the problem on which nearly every detective game founders.
Brute force is useless here: you will not work through sixty by sixty combinations, and there is no feedback on a single answer. One method remains that actually functions — understanding.
The game does not strand you, though. Every confirmed trio is a guaranteed piece of solid ground to push off from. Progress arrives in steps and always feels earned.
What does not work
The tail. The last eight or ten identifications come down to trivia: the shade of a uniform, a position in a line-up, a faint accent in one spoken line. Some players reach that stretch on honest deduction and finish it by elimination, and the difference between working it out and stumbling on it blurs at exactly the point where it should not.
The one-bit graphics are a precise choice — with no colour and no half-tones you look at shape rather than texture — but dithering in motion tires the eyes, and more than two hours at a sitting is physically hard. There is a style setting; it helps partially.
Why 8.8
Because this is one of very few games where deduction genuinely functions as a mechanic instead of being depicted as one. The rule of three is a rare case of a single systemic decision carrying an entire genre. A point and a bit comes off for the closing stretch, where the method breaks down, and for the physical strain of the image. You can play this once in your life; there will be no second time, because the knowledge is already in your head. All the more reason that the single pass is built almost flawlessly.
Source: en.wikipedia.org