Players choose dialogue, plant distractions, manage timing, and test branching social manipulations to avoid suspicion on a 1930s ship.
Why it may be useful
This is a clever, replayable puzzle about predicting people, worth playing for its dark comedy and tight design. What repeats is recall of a small cast, not a skill you can take much further than this one game.
Keep in mind
The premise means success requires blackmail, seduction, or framing an innocent person, so this recommendation depends entirely on being comfortable with that content.
Written 2026-09-15 · editorial assessment, not a measured effect · how we rate →
What the game is
Overboard is a short, heavily replayable mystery by the British studio inkle, released in June 2021 for PC at around twelve euros. It inverts the detective story: in the opening minute you push your husband over the rail of a 1930s ocean liner, and everything after that is the cover-up.
You have eight in-game hours before the ship docks and an inquiry begins. The clock only moves when you do, so every corridor you walk and every conversation you hold costs time you cannot get back. A single run lasts roughly thirty to forty-five minutes, which is the point: the game expects you to be caught, learn who was where and who knew what, and start again with that knowledge. The store page lists the verbs plainly, among them blackmail, theft, seduction, drugging a witness and framing someone else. The tone is dark comedy, the presentation is text with portraits, and Steam lists only English and Spanish.
What you actually do
You move between cabins, decks and the dining room, and you talk. Each passenger has a schedule, a set of things they know, and something they want, and your only tools are what you say and when you say it. Planting a suspicion on the wrong person requires that someone else be in the right place to notice it. Lies compound: an alibi only holds if a second person will confirm it, and confirming it usually costs you something. At eight hours the inquiry runs and you find out whether it held. Getting better is almost entirely a matter of knowing the ship and its people from previous attempts.
What it exercises
Predicts how a person will react to a line before saying it
Plans a route that spends limited time on what actually matters
Tracks what each character already knows and does not
Notices that an alibi only works if someone else confirms it
Treats a failed attempt as reconnaissance for the next one
Why we rate it Minimal Worksheet score 33 / 100
FitHow much real knowledge or skill does the play itself make you use?30weight 45%
Fit is low at 30. Reading what a character wants is the only way through, since there is no combat, no puzzle and no skill check. It stays low because a large share of improvement is simply remembering what happened last run, which is recall of a small fixed cast rather than reasoning in the moment.
TransferWill what you practise here work outside the game?22weight 30%
Transfer is lowest at 22. Modelling what another person knows and wants is genuinely useful, and this game drills it hard. It is held back by what it is drilled toward, which is deception, and by the fact that these characters are small rule systems, so what improves is prediction inside one fixed, invented model rather than anything about real people.
PracticeDoes the game make you use that skill again and again, as it gets harder?50weight 25%
Practice reaches 50. Repetition is the design rather than a side effect: you play the same eight hours over and over with better information each time, which is a great deal of practice at one narrow thing, predicting how a known person responds to a known move. It stays in the middle because that cast never grows.
What the evidence says
Editorial onlyNobody outside our desk has checked this. It is our reading of how the game plays.
The evidence is editorial only. The cited sources are two Steam addresses for the same listing, which confirm the developer, the June 2021 release, the price, the eight-hour structure, the thirty to forty-five minute run length, the replay-driven design, the listed verbs and the two supported languages. Steam shows no content descriptors at all despite the premise, so the store gives no warning of its own. Nothing on the page speaks to learning, no research on this game is cited, and the meaningfulness figures are our reading of the mechanics, which is why the level stays editorial.
Editorial assessment - not a measured scientific effect.
Fit
30/100
Transfer
22/100
Practice
50/100
Why it makes the cut
PASS - Its reverse-detective loop is distinctly game-first.
Editorial only
This is our reading of the mechanics, not a study of players. No research on this game is cited or exists, and the case rests entirely on what the replay-driven design demonstrably rewards.
Prices are US store listings in USD, last refreshed 2026-09-25; regional prices and sales differ.
No subscription requirement was found on the verified PC route.
Age guidance
Ages 13+ · ESRB T
Topics
Logic & strategy
Play styles
Single-player, Narrative, Strategy
Tags
Social manipulation · Branching alibis · Replayable deduction
Reading what a character wants is the only way through, but a large share of getting better is simply remembering what happened last run, which is recall rather than reasoning.
