How to spot who wrote what
Ninety seconds, eight anonymous cards, and eight people you supposedly know well. Most players read the cards top to bottom and guess as they go. That is the worst available approach. Here is a better one — written for Masquerade, and most of it transfers straight to Mimic.
Last updated: August 2, 2026
Sort first, match second
The mistake almost everyone makes is treating each card as an independent puzzle: read card A, decide who wrote A, move to card B. This throws away the most useful information on the board, which is how the cards relate to each other.
Spend your first twenty seconds reading all of them without assigning anything. You are looking for structure, not authors — which two cards feel like the same person’s register, which one is the outlier, which is the shortest and which is the longest. Sorting the set is what makes the matching step easy.
Work from the extremes inward
In any set of answers there are usually one or two you are immediately confident about — the card that is unmistakably Priya, the card that could only have come from the person who has made that exact joke twice this month. Lock those in first.
This matters more than it sounds, because every card you place removes a player from the pool for every remaining card. The middle of the board — the three generic one-liners that could belong to anybody — is not solved by staring at it. It is solved by elimination from the edges. Confident cards first, ambiguous cards last, and by the time you get to the ambiguous ones you may have only two candidates left.
Tells that hold up
Mechanics before meaning
Capitalisation, punctuation, spacing, whether they use an ellipsis, whether they write numerals or words. These are motor habits, not choices — they survive the player’s attempt to disguise the content, because the player isn’t thinking about them.
Length as a baseline
After two rounds you have a rough sense of each player’s natural output. Someone who wrote two lines in both previous rounds is unlikely to have produced the six-word card.
Vocabulary level and register
Whether someone swears, whether they use hedges like “kind of” and “I guess”, whether they write formally under time pressure. Register is deeply habitual and hard to fake consistently across a whole answer.
Non-native and second-language patterns
In mixed groups, article usage and preposition choice are strong signals. This is genuinely one of the most reliable tells in multilingual rooms.
The answer that is slightly too much like someone
In Mimic mode this is the whole game, and it runs the opposite way to everything else here: the writer wants to be legible, so look for the card that is a shade broader than the real person would ever be. An impression overshoots by design. The exaggeration is the signature, not a mistake.
Tells that are traps
Topic knowledge. The card about climbing is not necessarily the climber’s. In fact it is often bait — experienced players deliberately write into someone else’s territory, because they know you’ll follow the obvious association.
The card that sounds most like the person. A card that is a perfect impression of Sam is more likely to be someone doing Sam than it is to be Sam, who is actively trying not to sound like himself. Treat a very strong resemblance as suspicious rather than conclusive.
Blandness read as innocence. A conspicuously flat, styleless answer is not a shy player. It is usually a careful one. The empty card belongs to somebody, and “I can’t read this at all” is itself information — it points at whoever is playing the most deliberately.
Use the constraint
One player per card, one card per player. This is the strongest tool you have and most players never use it deliberately.
If you are certain about card A, that certainty is worth more than one point — it removes that player from contention everywhere else. When you are torn between two candidates for a card, check what each choice forces elsewhere on the board. A guess that leaves you with a clean assignment for the remaining cards is usually right; one that forces you into an implausible pairing at the end is usually wrong.
Also remember your own card is excluded from your guessing. That is one fewer distractor for you than a naive count suggests — and it means the number of live candidates is always one smaller than the number of players.
Play the long game across rounds
The reveal at the end of each round is free training data and most players waste it on laughing. Whoever is quietly noting that was how Priya writes when she’s hiding will be measurably better by round four.
Watch specifically for how each player disguises, not just how they write normally. Disguise strategies are much less varied than natural voices. Someone who went terse and lowercase to hide in round one will usually reach for the same trick in round three, and once you’ve seen a player’s hiding mode, it stops being a hiding place and becomes the easiest tell on the board.