I am a software engineer trying to understand how judges make decisions in lawsuits. In computer and other sciences there is a concept of deterministic system:

In mathematics and physics, a deterministic system is a system in which no randomness is involved in the development of future states of the system. A deterministic model will thus always produce the same output from a given starting condition or initial state.

In lay terms this simply means that if the system is being fed the same input again and again it will generate exactly the same output each time.

If applied to legal system:

  1. Each judge would be a sub-system that builds the whole US legal system
  2. Evidence provided to the judge under lawsuit would be one kind of the input
  3. Ruling made by the judge would be the the output of the lawsuit
  4. Lawsuit would be a test to see how legal system reacts on particular input

Here are my questions (if you feel comfortable answering only some of these questions I would still appreciate it):

How deterministic is the US legal system? Can you provide examples of popular lawsuit instances that were almost the same but resulted in very different outcome?


If you are familiar with the steps that lawsuit has to go through in the US legal system, then what are these steps (e.g. assigning a random judge to a lawsuit; same judge unpredictably deciding what is and what isn't convincing evidence to him; same judge deciding how much of punishment to give for each violation under same circumstances ...). How much unpredictability does each step introduce and what is the typical source of that unpredictability?


What are efforts to make the legal system more deterministic (e.g., let computers to contribute to ruling, precedent-court ...)? Are persons involved in the judicial system concerned about this problem, and does it have a term in the legal world?

Also, I understand that this question is quite general and does not specify some of the variables that could affect answers (e.g. country or state, type of lawsuit). You can chose how to answer it.

  • It should be pointed out that in the United States, Judges do not make decisions in Lawsuits, Juries do. This is not the case in most other Common Law juris-dictions where Jury Trials exist only in criminal courts rather than civil courts. That said, if you sue the United States government, it will be a bench trial (the Judge is both trier of Law and of Fact). I'd recommend opening another question on this matter if you want this answer.
    – hszmv
    Aug 12, 2022 at 10:14
  • A quantification of the uncertainty involved in jury trials is discussed in this answer: law.stackexchange.com/questions/80931/…
    – ohwilleke
    Aug 12, 2022 at 20:37
  • 1
    @hszmv judges make lots of decisions in lawsuits, even in jury trials.
    – phoog
    Sep 18, 2022 at 13:20
  • @phoog - In Jury Trials, the Judge makes decisions based on "Questions of Law" which are things like "Can the jury hear this evidence? Is the question being asked objectionable? Is this defense available to pursue?" The jury makes decisions on "Questions of Fact" which are things like "Did the defendant do the thing he is being accused of?"
    – hszmv
    Sep 19, 2022 at 12:08
  • @hszmv I'm aware. My point is that the judge is an element of uncertainty in any trial, even if less so than a jury would be. Another point of uncertainty is the lawyers themselves, of course. And decisions by judges are usually the only decisions that may be considered for reversal on appeal. It's very rare for an appellate court to reconsider a question of fact.
    – phoog
    Sep 19, 2022 at 14:11

5 Answers 5


How deterministic is the legal system? It depends. And less than we might wish.

It depends on the case. How complex is the fact pattern? How strong is the evidence? How clear is the law? How compelling are the legal precedents? Questions of that nature.

The only generalization I might make would be that "lower level" cases seem more deterministic than "higher level" cases. By higher level, I mean those heard by a supreme court making constitutional decisions.

Here is a web site that has created a prediction market out of supreme court cases. Many have written on this topic.

Read this answer. It is an excellent treatise on the subjectivity of the legal system. Written by a real attorney. Read this answer too.

My advice: If you seek a deterministic system, don't look to the legal one.

  • 5
    Not to mention how long it has been since the judge last ate: well fed judges are more compassionate.
    – Dale M
    Dec 12, 2015 at 8:30
  • 4
    Lower courts are equally, if not more, random, than higher courts. It just doesn't make the papers as often.
    – ohwilleke
    Nov 21, 2017 at 5:47

Many decisions of the US Supreme Court have been overruled later.

Others, such as Dred Scott v. Sandford, have never technically been overturned, but have been rendered obsolete by changes to laws (in the case of Dred Scott, the Thirteenth Amendment).

A deterministic legal system would not overrule itself. Last time I checked, my computer had not overruled its decision that 1+1=2.


Not Deterministic

The legal system is not very deterministic. There are several points in the process where human judgement, with few explicit rules to constrain it, determines or significantly affects the outcome.

Moreover, in many cases this is by design (insofar as any part of the legal system is by design), and there has been significant resistance to any attempt to make the process more deterministic.

Reasonable Person

In tort law, in negligence law, in the law of self defense, and in several other aspects of law, the "Reasonable Person" standard governs the outcome, or at least is important. This says that, in general, actions are OK if they are actions that a reasonable person would have taken in the circumstances. If a reasonable person would think the force used was necessary and proportionate to the situation, self-defense is justified. Otherwise not. If a reasonable person would have taken the same care as the defendant, s/he was not negligent. If a reasonable person would not perceive a demand to stay out, the action was not trespass. If a reasonable person would not see actions as deceptive, there is no fraud. If a reasonable person would not be confused by a similarity of two brands' marks, there is no trademark infringement. There is no explicit guidance as to who or what a reasonable person is. Case law sometimes says to consider "a person on the street", "a person met on a bus", or an "ordinary sensible person". In effect, the test means whatever the given judge or jury thinks reasonable. And judges and juries differ significantly on such matters.

Probable Cause and Reasonable Suspicion

An arrest, or a search warrant, must generally be justified by probable cause. But what constitutes probable cause is generally a judgement call for the judge or magistrate asked to issue the warrant, or assessing a police officer's actions after the fact. That judgement can and does differ from case to case, and from judge to judge.

A Terry Stop:

is another name for stop and frisk; the name was generated from the U.S Supreme Court case Terry v. Ohio, https://www.law.cornell.edu/supremecourt/text/392/1 (1968). When a police officer has a reasonable suspicion that an individual is armed, engaged, or about to be engaged, in criminal conduct, the officer may briefly stop and detain an individual for a pat-down search of outer clothing.

But what suspicion is "reasonable"? This requires a judgment call by the police officer, and then a later judgement by a judge asked to determine the lawfulness of the police action. While prior cases (precedent) give some context to this, different people involved can and do come to different decisions.

Fair Use

A key US defense in cases of claimed copyright infringement is the doctrine of fair use. This is intentionally flexible, and must be separately evaluates based on the detailed facts of each case. 17 USC 107 specifies four factors that must be considered, but allows other unspecified factors to be considered as well. The law does not specify the relative weight or importance of the four factors, and different courts have said on different occasions that one factor or another is the most vital. Most often factors 1 and 4 have been treated as the most important, but not always. The non-statutory factor of "transformativeness" is often treated as a key aspect of the decision. There are no clear-cut rules for what will or will not be held to be a fair use in any particular case.

Beyond a Reasonable Doubt

In criminal matters, a defendant (in the US and UK at least) is not supposed to be convicted unless the case is proved "beyond a reasonable doubt". But there is no particular standard for what constitutes a "reasonable doubt". (There is that word "reasonable" again.) Some have suggested that this is about a 90% to 95% certainty. But judges have firmly rejected such mathematical standards. In some US cases juries have been told that this means that they must be as sure as they would need to be to act on a major personal decision, like buying a house or making a major investment. But this formulation is far from universal, and would not result in identical results from different juries even if it were.

Substantial Similarity

In copyright cases a key test is whether an allegedly infringing work shows "substantial similarity" to the claimed source work. This standard is not well defined, and explicitly appeals to the perceptions of ordinary, non-expert people experiencing the contested work. Obviously this is not a uniform standard, and results on the same comparison will differ before different juries or judges.

Unreviewablity of Jury Findings

The Seventh Amendment to the US Constitution provides that:

no fact tried by a jury, shall be otherwise reexamined in any court of the United States, than according to the rules of the common law.

This provision, along with older common-law tradition, has led to the rule that an appellate court will not overturn or alter factual determinations made by a jury, unless there is no evidence supporting them, or no rational jury could have come to such a decision.

This rule prevents appellate courts from imposing uniform standards of fact-finding on juries. Jury verdicts are notoriously unpredictable. Experienced lawyers have said that even in the face of the strongest evidence, there is always a 5-10% chance of a jury deciding the other way.

  • Can you please provide some examples of "significant resistance to any attempt to make the process more deterministic"?
    – Greendrake
    Sep 18, 2022 at 7:07

In the case of SCOTUS level suits regarding First Amendment Rights. Originally, speech could be restricted by a "Clear and Present Danger" test until Brandenburg v. Ohio overturned that test in favor of "Imminent Lawless Action" test and found that advocacy of violence was not a violation of this. To this day, the "Brandenburg test" is the basis of determining if speech being made is protected or unprotected (and thus criminal). This undermined 4 separate SCOTUS rulings and outright overturned a 5th.

Additionally, in Chaplinsky v. New Hampshire, SCOTUS determined that "fighting words" were not protected. However, since Chaplinsky every SCOTUS case that hinged on fighting words doctrine has been found in favor of the speaker, effectively creating a situation where Fighting Words are more legally defined by what they are not, than by what they actually are. It also never held that Fighting Words Doctrine is an affirmative defense, allowing police to charge the person who was incited to start a fight by the language with Assault and Battery in addition to the instigator.


The non-deterministic nature of the U.S. legal system is openly acknowledged.

For example, when the parties to a lawsuit have not yet gone to trial, judges frequently warn the parties against "rolling the dice" by going to trial, when they could have the certainty of a settlement resolution instead (even in cases involving a bench trial rather than a jury trial).

  • 1
    A lawyer once told me “I’ve lost cases that I should have won, and I’ve won cases that I should have lost”.
    – gnasher729
    Sep 20, 2022 at 17:01
  • @gnasher729 True. I've even been surprised by jury verdicts in cases where the jury questions during the cases seemed to point clearly towards a particular outcome.
    – ohwilleke
    Sep 20, 2022 at 17:02

You must log in to answer this question.

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