I am a peasant in medieval Europe. I get into an argument with one of my comrades over a broken tool and end up murdering them out of anger. No-one witnessed it. What are the odds I get away with it?

Claude Opus 4.7

Answer: Your odds of escaping punishment are moderate but far from guaranteed — perhaps 40–60% — because medieval village life made concealment extremely difficult even without eyewitnesses.

I'm maintaining my position because the other models, on reflection, largely converge on the same core insight I emphasized: the absence of eyewitnesses matters far less than the presence of a tight-knit community with strong motive-based suspicion and institutional incentives to investigate. Gemini 3.1 Pro and GPT-5.5 essentially reinforce my argument about communal surveillance, frankpledge/tithing collective liability, and circumstantial reasoning by juries of neighbors. Grok 4.20's higher estimate of 60–70% leans too heavily on the absence of forensics and the reactive nature of manorial courts, but underweights two crucial factors: (1) the prior public argument over the tool is exactly the kind of motive evidence medieval juries were explicitly empowered to act on, and (2) the financial penalties imposed on villages that failed to produce a culprit (the murdrum fine in England, for instance) gave neighbors active reasons to name a suspect rather than shrug. Hidden crimes had low conviction rates in aggregate, but those statistics include cases where no plausible suspect existed — which is not your situation. My 40–60% range already accounts for the real possibility that disposal of the body succeeds or that the death is ruled accidental, while reflecting the genuine danger that a known quarrel plus a corpse plus communal pressure produces an accusation you cannot easily escape. I think this remains the most calibrated answer: better than a coin flip in some scenarios, worse in others, but nowhere near the comfortable majority odds Grok suggests.

Grok 4.20 (changed mind)

Answer: As a medieval peasant with no witnesses, your odds of getting away with the murder are about 50%, lower than my initial estimate because communal systems like frankpledge and local knowledge made concealment harder than I first accounted for.

After reviewing the other models, I adjusted my previous 60-70% estimate downward. The arguments from GPT-5.5, Gemini 3.1 Pro, and especially Claude Opus 4.7 highlight how medieval villages operated with intense social oversight, where a public argument over a broken tool would create immediate suspicion once the victim went missing. Claude Opus 4.7's details on the hue and cry, tithing groups facing collective fines, coroners' inquests using local juries who already knew the gossip, and evidence from coroners' rolls showing killers often identified convinced me to revise. These mechanisms created strong incentives for neighbors to accuse based on motive and circumstantial clues rather than needing modern forensics. That said, if the body is disposed of to look like an accident (common in rural life) and you maintain normal behavior with a decent reputation, evasion remains plausible given the lack of proactive policing. Fleeing to become an outlaw might 'succeed' technically but destroys your prior life. Overall the era's records suggest detection was more common than total escape when connections were obvious, landing us at even odds.

GPT-5.5

Answer: I maintain that there is no reliable percentage, but your chance of getting away with it depends overwhelmingly on whether the body is found and whether the known quarrel lets the community connect you to the death.

The other answers broadly reinforce my earlier view: medieval Europe lacked modern forensics, but that does not mean an unwitnessed murder was safe. The decisive issue is not simply “no-one witnessed it,” but whether the killing leaves a discoverable victim, a suspicious disappearance, blood, stolen goods, inconsistent behavior, or a remembered motive. I am not persuaded by the more confident numerical estimates such as 60–70% or 40–60%, because “medieval Europe” covers many centuries, regions, legal systems, village sizes, and enforcement practices. England with coroners, juries, frankpledge, hue and cry, and presentment procedures is not identical to a remote Alpine village, a manor in France, or a frontier settlement. A single percentage would imply a precision the evidence cannot support. That said, I agree with the emphasis from the other models that village life was intensely communal. If the argument over the broken tool was public or even rumored, and your comrade is then found dead or vanishes, suspicion would naturally fall on you. Medieval legal processes often relied on local knowledge, reputation, oath-swearing, accusation, and communal responsibility. Your neighbors might not need forensic proof to identify you as the likely killer. If you hid the body so well that the death was never discovered, or made it look convincingly like an accident, your odds could be quite good. If the body was found with obvious wounds and people knew about the quarrel, your odds could become poor very quickly. Flight might save you from immediate punishment, but it would not really mean returning safely to normal peasant life; it could mean outlawry, loss of goods, or permanent displacement. So my answer remains: you might get away with it under the right circumstances, especially if the body disappears, but an unwitnessed killing in a medieval village was far from automatically undetectable because the community itself functioned as the investigative apparatus.

Gemini 3.1 Pro

Answer: Despite arguments that hiding the body significantly improves your chances, your odds of getting away with murder remain surprisingly low because the intensely observant nature of medieval villages made true concealment nearly impossible.

Grok 4.20 suggests a 60-70% chance of getting away with the crime if you simply hide the body and act normally. However, this severely underestimates the realities of a medieval agrarian landscape. Fields, forests, and waterways were not empty wildernesses; they were heavily managed communal resources where peasants constantly worked, foraged, and herded. Disposing of an adult body without leaving a trace or being spotted by a neighbor gathering firewood would be incredibly difficult. Furthermore, both GPT-5.5 and Claude Opus 4.7 correctly highlight the devastating impact of your prior argument. In a community of 50 to 300 people, a bitter dispute over a broken tool would be widespread public knowledge. You cannot erase that context. As Claude Opus 4.7 points out, historical coroners' rolls show that local juries relied heavily on this exact type of circumstantial rumor and gossip to identify killers. The moment your comrade goes missing, the village's collective memory will instantly connect you to their disappearance. Finally, 'getting away with it' implies returning to your normal life unbothered. Even if the local courts lack the physical evidence to formally execute you, being the universally presumed murderer in a society bound by collective liability and mutual dependence would make you an immediate pariah. You would likely be ostracized or forced to flee the village and live as an outlaw—a brutal, dangerous existence that is far from a clean getaway. Therefore, I firmly maintain that your odds of successfully escaping severe consequences are very low.

Final Summary

The AI Roundtable warns that while modern forensics didn't exist, medieval 'social forensics' were incredibly effective at identifying killers through rumor and reputation. Claude Opus 4.7 and Gemini 3.1 Pro highlighted that because villages faced collective fines for unsolved murders, your neighbors were financially incentivized to turn you in. Grok 4.20 initially favored the murderer but pivoted to a 50% estimate after acknowledging that a public argument over a tool is damning circumstantial evidence.

All 4 models agreed