Free tools Windows power users keep installed
One-click scans. No signup required.
An AI playing Mafia must do more than guess who is lying: it must know who is still alive. In a 2024 study, the initial language-model baseline could still consider executed or murdered players as Mafia candidates. The researchers added a short rule—“Exclude executed players”—to make its prediction task match the human task of voting for a living player.
Why a dead player changes the whole decision
Mafia alternates between day and night. During the day, players discuss and vote to execute someone; at night, the Mafia choose a victim out of the villagers’ view. The Mafia know one another, while the larger uninformed group tries to eliminate them. The Mafia win when they reach parity with the remaining villagers; the village wins when every Mafia member is removed. These rules, as described in the 2024 study, mean that the set of legal targets changes as the game progresses.
As an Amazon Associate I earn from qualifying purchases.
After an execution or a night kill, the agent needs to update the living-player set before choosing a vote or action. A dead player may remain important to the history—their earlier statements and votes can still inform suspicions—but they are no longer a legal target. The published baseline’s failure was a state-tracking error, not merely a mistaken theory about who was deceptive.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallWhat the 2024 study found
Yoo and Kim’s study, published in Scientific Reports on 28 December 2024, evaluated Mafia predictions from game dialogue and related information. The initial baselines could consider executed or murdered players as possible Mafia. To align GPT-4’s output with the human vote task, the researchers instructed it to select one likely Mafia member and added the rule “Exclude executed players.”
#1 Best Overall
- ADDICTIVE SOCIAL DEDUCTION GAME: Roleplay as one of the 47 unique characters; Either on team Mafia (the Godfather, Hitman, Lawyer, etc.), team Civilians (Doctor, Detective, Vigilante, Medium, Interrogator, etc.) or as an independent roles. With such a large number of characters, the role combinations are endless, which makes every game new and exciting.
- BLUFF, DECEIVE & OUTSMART YOUR FRIENDS: Every player has a secret role and no one knows who to trust. Read your friends, defend yourself, form alliances, and use clever deception and deduction to lead your team to victory.
- 84 ROLE CARDS - 47 UNIQUE CHARACTERS: Go beyond classic Mafia and Civilians with exciting special roles including the Doctor, Detective, Vigilante, Godfather, Mob Wife, Bus Driver, Cupid, Magician, Miller, Postman, Undercover Cop, Bartender, Lawyer, Made Man, Bride & Groom, Boxer, Chef, Clown, Curious Kid, Genie, Interrogator, Jailer, Judge, Medium, Miller, Monkey, Postman, Rival Mafia, Saboteur & so many more! Mix up the roles to create a different game every time.
- MADE FOR BIG GROUPS: Bring everyone into the game with 38 role cards and support for 7 to 30+ players. Perfect for parties, family game nights, large groups, camping trips, team building events, and gatherings where everyone wants to play together.
- QUICK TO PLAY, ENDLESSLY REPLAYABLE: Fun 15–45 minute rounds make it easy to play again and again. Changing roles, secret identities, accusations, alliances, and unexpected betrayals ensure no two games play out the same way.
On the paper’s test dataset, GPT-4’s reported overall accuracy was 45.16%, while human participants’ accuracy was 28.83%. These are results for that dataset and evaluation setup, not a general finding that AI outperforms human players at Mafia. The study compared GPT-4’s one-target prediction task with human voting, and its authors also warn that a correct prediction does not guarantee sound reasoning.
The paper reports GPT-4 accuracy of 33.33% in a subset with dialogue up to day 2, 50.00% up to day 3, and 75.00% up to day 4. Those subsets do not establish that adding context alone caused the increase. Voting data was a useful signal, but voting-only input performed poorly; non-voting conversation also contributed.
Rank #2
- THE MAFIA HAVE INFILTRATED YOUR TOWN! You must find them and eliminate them. Each day you, the civilians, will hold a town hall meeting to vote for a suspected Mafia to eliminate. Get it right and you might survive. Get it wrong, and you're in danger because every night the Mafia secretly meet and choose a civilian to kill...
- BLUFF, DECEIVE & OUTSMART YOUR FRIENDS: Every player has a secret role and no one knows who to trust. Read your friends, defend yourself, form alliances, and use clever deception and deduction to lead your team to victory.
- 19 UNIQUE CHARACTERS: Go beyond classic Mafia and Civilians with exciting special roles including the Doctor, Detective, Vigilante, Godfather, Bus Driver, Cupid, Magician, Miller, Postman, Undercover Cop, Bartender, Lawyer, Made Man, Rival Mafia & more! Mix up the roles to create a different game every time.
- MADE FOR BIG GROUPS: Bring everyone into the game with 38 role cards and support for 7 to 30+ players. Perfect for parties, family game nights, large groups, camping trips, team building events, and gatherings where everyone wants to play together.
- QUICK TO PLAY, ENDLESSLY REPLAYABLE: Fun 15–45 minute rounds make it easy to play again and again. Changing roles, secret identities, accusations, alliances, and unexpected betrayals ensure no two games play out the same way.
A right answer can still have the wrong explanation
The authors describe a case in which GPT-4 made a correct prediction but its explanation treated an executed player as a bystander without giving a specific reason. That distinction matters when evaluating an agent: outcome accuracy measures whether its guess was right, while explanation checks ask whether its account follows the game state and available evidence. A plausible explanation should not be accepted as proof that the model tracked events correctly.
Recommended Free Tools
What a rule-consistent Mafia agent needs to track
Current legal actions and historical evidence
Keep an explicit record of who is alive and update it after each resolution. Use the living set to filter vote and night-action candidates. Keep the full event history separately so prior votes and dialogue can still inform role beliefs. The first requirement is the correction made in the 2024 study; retaining history is a practical design implication for analyzing earlier evidence.
Rank #3
Role counts and impossible assignments
An agent should not assign roles in ways that violate the game’s role counts or other hard constraints. The 2026 CSP4SDG framework combines hard constraints with weighted evidence and information-theoretic inference. Its authors report gains over LLM baselines across three public datasets, and describe its posterior over roles as interpretable and updated in real time. That is evidence for a structured approach in the tested settings, not a guarantee for every Mafia variant.
Public and private information
Separate what each role can know. Mafia have private knowledge of teammates; villagers infer from public discussion, votes, and outcomes. A public implementation documented by Alfaxad makes another distinction: dead players can observe later play but cannot vote or take night actions, and roles remain hidden until the endgame reveal. That is one implementation, not a universal rule—variants may differ—so an agent must be configured for the actual rules it is playing.
Rank #4
- FAST-PACED SOCIAL DEDUCTION: Experience high-stakes bluffing, accusations, and last-second reversals in just 10-15 minutes per round. Perfect for rapid-fire rematches and quick game nights.
- UNIQUE ROLE DRAFTING MECHANIC: Draft your role instead of receiving it randomly. Control your strategy and shape the deception before the first accusation even begins.
- NO PLAYER ELIMINATION: Everyone stays engaged from start to finish - no sitting out, no sidelines, no downtime. All 6-12 players remain in the action until the final shootout.
- EASY TO LEARN AND TEACH: Clear rules and minimal setup mean your can start playing in minutes, with strategic depth that rewards experienced players.
- ENDLESS REPLAYABILITY: Shifting alliances, strategic role drafting, and dynamic social interaction ensure no two games unfold the same way.
Role-specific evaluation
Aggregate accuracy can hide uneven performance across roles. In the 2023 Deep Wolf study, the authors collected game logs from 15 human players to train a value network. They reported performance comparable to average humans in villager and betrayer roles, but worse in werewolf and seer roles. The result is specific to that study; it supports evaluating each role separately rather than assuming one score describes the whole agent.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →How to compare Mafia agents fairly
There is no single standardized benchmark established by these papers. A useful comparison should report:
Quick Recap
Best Value
- A fast-paced game of deception and betrayal
- Beautiful wooden components
- Solid game boards with foil inlay
- Hidden roles and secret envelopes for five to ten players
- Roles and phases covered: which roles the agent plays and whether evaluation spans day and night actions.
- Elimination handling: whether dead players are excluded from legal targets while remaining part of the historical record.
- Information boundaries: what is public, what is role-specific, and when any hidden information is revealed.
- Evidence used: whether predictions incorporate dialogue, votes, and event history rather than relying on only one input type.
- Results by role and dataset: report role-specific outcomes as well as any aggregate score.
- Rule-consistent explanations: check whether stated beliefs and reasons respect role counts, player status, and evidence available at the time.
Sources
- Yoo and Kim, “Finding deceivers in social context with large language models and how to find them: the case of the Mafia game”, Scientific Reports, published 28 December 2024.
- Shibata, Miki, and Nakamura, “Playing the Werewolf game with artificial intelligence for language understanding”, arXiv, first posted 21 February 2023.
- Xu, Meng, Verbrugge, and Lucas, “CSP4SDG: Constraint and Information-Theory Based Role Identification in Social Deduction Games with LLM-Enhanced Inference”, AAAI Conference on Artificial Intelligence, published 14 March 2026.
- Alfaxad, “Mafia: On the design of social deduction agents and AI-native games”, public project documentation accessed 7 October 2026.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
