For example, suppose there are two houses side-by-side. One day someone throws an incendiary device from one house to the other, causing a fire. The facts of the case are:

  • There are four people in the house from which the incendiary device is thrown
  • It must be one of the four that threw the device (e.g. because the device is thrown from the second floor and there is no sign of forced entry into the house)
  • All four people deny that they are the one that threw the device, although none of them can prove that they did not throw the device

What happens in situations like this where there's not enough evidence to pin the attack on one of the four people, but there is enough evidence to be certain that one of the four people is guilty? Do all four get convicted (i.e. three innocent people are convicted) or all four get released (i.e. one guilty person goes free)?

I feel like this must have happened in the past, but I'm unable to find any Google results for this. The closest is the Prisoner's dilemma, which implies that all four are convicted on a lesser charge. However, the Prisoner's dilemma is a thought experiment in game theory so I am not sure if it is legally correct.

  • 1
    Do you have a particular jurisdiction in mind? Detailed rules on such things vary from place to place. Commented May 11, 2022 at 4:48
  • @DavidSiegel not really, no. I don't mind answers for any jurisdiction.
    – Allure
    Commented May 11, 2022 at 4:51
  • One possible historical example is the Waco Siege. After the siege, some of the surviving Branch Davidians were charged with conspiracy to murder of federal agents and aiding and abetting in the murder of federal agents; but so far as I can tell, nobody was actually tried for the murder itself. (It's possible, though, that the people who actually shot the federal agents were known to have been killed in the standoff; I don't know for sure.) Commented Jun 19 at 17:44
  • This reminds me of a common trope in TV crime dramas: there's DNA evidence, but the DNA matches a set of twins, and there's no other evidence that pins it on a particular one of the twins. They each provide reasonable doubt for the other.
    – Barmar
    Commented Jun 19 at 20:22
  • 1
    Benjamin Franklin: "it is better 100 guilty Persons should escape than that one innocent Person should suffer". So 3-to-1 seems like a bargain.
    – Barmar
    Commented Jun 19 at 20:24

4 Answers 4


Unless there is enough evidence to convict one or more of the suspects, none will be convicted. In general the argument:

We know it must be one of you, but we don't know whch, so we find you all guilty.

is not allowed in any non-dictatorial jurisdiction.

Just how much evidence is needed for a conviction varies by jurisdiction in theory, and by judge or jury in practice.

Also, it would be possible to charge several of the residents with having acted jointly in the crime, but there would still need to be sufficient evidence against each defendant to obtain a conviction.


In any criminal justice system that follows the presumption of innocence principle, all four would go free.

Unless there is reason to believe that the four people acted together, they would have to be tried separately in 4 separate trials.

In each trial there would be the same burden of proof: The prosecution would have to prove beyond reasonable doubt that this specific defendant is guilty. When they can not do that in any of those four trials, then all four would be acquitted.

And no, "We already acquitted 3 people, so by exclusion principle this fourth one must be the culprit" would not work. An acquittal due to lack of evidence is not a proof that this person did not commit the crime. It just means that it wasn't possible to find out whether or not the person is guilty, so by the principle of in dubio pro reo, they should be treated as if they were innocent.

As Sir William Blackstone said:

It is better that ten guilty persons escape than that one innocent suffer.

The job of the criminal justice system is not to punish crimes; it is to punish criminals (and only criminals). When in doubt, then it is better for a crime to go unpunished than to risk punishing an innocent person for a crime they did not commit.


In most jurisdictions, if they all stick to their story, they all go free.

You are right that the Prisoners Dilemma isn’t law, but that very fact makes it universally applicable.

So, it is possible that they will all go to jail, or one or more of them will go to jail, and the ones that go to jail may or may not be guilty. Plea deals are the bane of justice, and it would be quite possible that the prosecution could get guilty pleas out of all four by a combination of lies, soliciting perjury, threats and lighter sentences.

  • I have several books on game theory, If you really want I can type in a quote, but trust me, if all those conditions are not fulfilled it isn't a true Prisoner's Dilemma. There are several other related scenarios. (Also i used to live just a few blocks away from the late Professor John Nash, who helped invent the Prisoner's Dilemma and did invent the Nash Equilibrium {a key game theory concept} and discussed it with him a few times.) Do you want a supporting cite? Commented May 14, 2022 at 23:04
  • What makes the Prisoner's Dilemma a dilemma, and an interesting problem in game theory is that it seems that a fully rational person will be forced rationally into a bad outcome. Joe assumes Kim has confessed, Joe should confess to avoid being on the short side of case D. But If Joe assume that Kim has NOT confessed he should confess to get case A instead of case B. If both follow that logic, they will wind in the unfavorable vase C, where they could have case B if they trusted each other and neither was tempted to betray the other. [...] Commented May 15, 2022 at 0:48
  • ]...] But all that logic works only if the inequalities between cases A, B, C and D hold. If any do not. the logic chains are different and it isn't a Prisoner's Dilemma any more, it is one of several other classic game theory scenarios, depending on which inequalities are changed. There is a whole chapter of Douglas R Hofsteader's Metamagical Themas on this, just to mention one of the more easy to follow works on the subject. Commented May 15, 2022 at 0:50
  • The situation described in the article is indeed a horrific perversion of Law. But it is not a Prisoner's Dilemma. It isn't even close. The Prisoner's Dilemma is a technical term in game theory, which is sometimes misused to describe other, vaguely related situations, but rarely so extremely misused as it is in this answer. A Prisoners Dilemma is always a situation involving exactly two participants, the two prisoners in the classic formulation, or two traders in more useful formulations., It does not refer to other situations in which people may have to make choices with serious effects.[...] Commented May 15, 2022 at 18:11
  • [...] In particular, a Prisoner's Dilemma. does NOT refer to any situation in which a participant has imperfect knowledge of the rules governing the situation, as the innocent accused did in the NYT story. Commented May 15, 2022 at 18:12

The case of Southwest Airlines flight 1763 may prove instructive.

On August 11, 2000, Jonathan Burton stormed the cockpit door of the Boeing 737 operating the flight, in an apparent case of air rage. The 19-year-old was subdued by six to eight other passengers with such force that he died of asphyxiation. No criminal charges were laid.

  • 2
    From further down in the article: "the conclusion of the U.S. Attorney's office [was] that criminal charges would not be filed because the death was not intended". Doesn't seem like a good example. Perhaps they could have charged those people for negligent homicide, but the perpetrators probably had a good chance to get away with self-defense.
    – Philipp
    Commented May 12, 2022 at 8:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .