My, limited, understanding of the US legal system is that pretty much everyone will be offered a plea bargain before going to trial. That would imply to me that the only people that end up in trial are either people who's case is truly ambiguous and it's not clear which way the the jury will vote, or people who probably should have accepted a plea bargain and stubbornly refused to.

So I'm wondering from those who have experience of jury trials how many fall into the former category and how many into the latter? How likely is it that you have a good idea which way the jury will vote before the trial even starts?

  • In theory, prosecutors are only supposed to bring cases that they believe they can prove beyond a reasonable doubt (and that are also in the public interest).
    – Davislor
    Commented Jun 9, 2022 at 16:54
  • 1
    @Davislor The standard of what kind of cases prosecutors may ethically bring is lower than that, although there is an ethical lower bound.
    – ohwilleke
    Commented Jun 10, 2022 at 1:37
  • @ohwilleke There probably are some corner cases. But, in general, you’d agree: bringing a case to trial that the prosecutors themselves don’t believe they can prove is bad on multiple levels?
    – Davislor
    Commented Jun 10, 2022 at 1:45
  • 1
    Rule of Professional Conduct 3.8(a): "The prosecutor in a criminal case shall: (a) refrain from prosecuting a charge that the prosecutor knows is not supported by probable cause;" americanbar.org/groups/professional_responsibility/publications/…
    – ohwilleke
    Commented Jun 10, 2022 at 5:48

2 Answers 2


Without being omniscient, it is impossible to pin down an exact percentage, and there are a lot of context specific reasons why some kinds of "easy" cases are more likely to go to trial than others.

But, there have been quite a few serious efforts to answer this question with data (putting aside the normative issue of whether you should go to trial when you are likely to lose).

One of the best statistical estimates comes from an analysis of criminal jury trials in Sarasota, Florida and the race of the jurors on the jury pool. Based upon that study, it is possible to infer statistically that an average juror of either race would reach the same aquittal or conviction decision in about 55% of cases involving black defendants and about 68% of cases involving white defendants. But, there are many other cases that are close enough on the merits given the likely available that the outcome depends upon the race of the jury, which basically means that the evidence can be reasonably viewed in different lights to reach different conclusions based upon your predispositions before seeing it.

For the sample as a whole, about 68% of cases where convictions results and a minimum of 14% of cases that produce acquittals, are sufficiently clear than the racial composition of the jury doesn't matter. Given that something like 90% of cases produce plea bargains generally, and that plea bargains are usually made before a jury pool is drawn, the random impact of the racial makeup of the jury pool that is selected for a case in Sarasota, Florida only directly matters in about 2% of all criminal prosecutions.

My criminal procedure professor from law school made one of the most comprehensive surveys of data pertinent to estimating wrongful conviction rate ever prepared. He concludes that wrongful conviction rates for murders and rapes are on the order of 2.3%-5%, and that wrongful conviction rates for other serious felonies are probably somewhat lower (since weak cases are less often pursued) but that it is harder to determine precisely what error rate is involved since the legal process and civic activism rarely takes the time and resources necessary to consider wrongful juvenile convictions or wrongful convictions for less serious crimes.

Other sources have suggested that wrongful conviction cases, and studies comparing jury outcomes with conclusions of the presiding judges in the cases regarding where the jury does and does not agree with the judge regarding the correct outcome generally speaking point to a similar level of uncertainty in decision making accuracy. suggest that as many as 10-20% of jury determinations are erroneous, although this is to some extent a product of samples biased for cases with a high risk of wrongful convictions.

For example, analysis of a special set of state court cases in 2000-01 from four jurisdictions in a study by the National Center for State Courts (Hannaford-Agor et al 2003) suggested that approximately 17% of jury verdicts were inaccurate, 7% of the all jury verdicts were wrongful convictions and 10% of all jury verdicts were wrongful acquittals, with corresponding rates of 10% wrongful convictions and 1% wrongful acquittals for the judges' verdicts (Spencer 2007).

Similarly, an abstract of one study stated that: "I examine . . . how the criminal system in the United States handled the cases of people who were subsequently found innocent through post-conviction DNA testing. . . . The leading types of evidence supporting their wrongful convictions were erroneous eyewitness identifications, faulty forensic evidence, informant testimony, and false confessions. . . . . few innocent appellants brought claims regarding those facts, nor did many bring claims alleging their innocence. For those who did, hardly any claims were granted by appellate courts. . . . courts often denied relief by finding error to be harmless on account of the appellant's guilt. Criminal appeals brought before they proved their innocence using DNA yielded apparently high numbers of reversals—a fourteen percent reversal rate. However, . . . the reversal rate is indistinguishable from the background rate in appeals of comparable rape and murder convictions[.]"

Another way to judge the ratio of easy to hard cases is to look at conviction or verdict rates in cases that go to trial. While the vast majority of criminal charges brought result in conviction of something and the vast majority of civil cases brought result in a judgment for the Plaintiff, in an hypothetical ideal world where the lawyers and parties on both sides of cases are rational actors with the best available information and there is no bias in the availability of information, you would expect pre-trial settlements to resolve, on average, all of the cases with an objective lean one way or the other, leaving only the cases that are, on average, coin flips left to go to trial, with 50-50 outcomes, regardless of the mix of cases originally filed.

And, that model isn't horrible. Less than 2% of civil cases and less than 10% of criminal cases go to trial. Civil case outcomes vary by type of case, but the overall result in those the go to trial is close to 50-50. But, in criminal cases, convictions greatly outnumber acquittals, because "easy cases" where a conviction is likely often still go to trial, because neither side is paying for their lawyers from their own funds in most cases, because there is little incentive to offer favorable settlements in close cases, and because, as discussed below, there is a significant irreducible risk of an inaccurate outcome.

In federal criminal cases that actually go to trial, the Pew Research Center’s data shows that defendants who pursue a trial experience different outcomes based on whether they choose a bench or jury trial. The acquittal rate in bench trials is 38% (a very small and unrepresentative sample), whereas it’s 14% for juried trials (the vast majority of cases). This would suggest that about 72% of federal jury trials are "easy" cases, while about 28% are "hard" cases, in line with the Sarasota study in order of magnitude.

Note that this is different from "conviction rates" which compare the percentage of cases charged that produce guilty verdicts or plea bargains, rather than conviction rates in the subset of cases that go to trial. The percentage of people charged with some federal crime who end up being convicted of something is very, very high, compared to state court, but that is almost entirely due to the ability of federal prosecutors to cherry pick strong cases with high mandatory minimum penalties and to secure plea bargains as a result, rather than from different rates of criminal jury trial outcomes.

Still, in "2018, the Bureau of Justice Statistics reported that among defendants charged with a felony, 68% were convicted (59% of a felony and the remainder of a misdemeanor)" and the conviction rate at trial is lower than the conviction rate of all people charged criminally with felonies, since plea bargains are more common than unilateral government dismissals of all charges. But, in almost every court system, more than 50% of criminal trials result in convictions. So, the percentage of "easy" criminal cases going to trial in state courts is still significantly lower than in federal courts. Maybe the percentage of state criminal cases going to trial that are "easy cases" in the range of 5%-20% in this kind of analysis (which is quite a bit lower than the more rigorously designed Sarasota study).

The challenging thing, of course, is knowing in advance which cases judges or juries will get wrong. I typically conceptualize the issue as a certain irreducible uncertainty of outcome any time you actually roll the dice on going to trial on the order of 10%-20% whether you are in front of a judge or a jury, and an additional uncertainty in cases where there is some specific reason to think that the outcome is a close call, or that special risk factors for inaccurate verdicts (like heavy reliance on cross-racial eye witness identification of suspects) is present.

From the perspective a client, even in a seemingly secure case, this impacts how plea bargains and settlement offers are evaluated. If your client is actually in the wrong and facing very severe punishment and there are no offers to settle that don't also involve very severe punishment, going to trial and hoping to benefit from the irreducible inaccuracy of trial determinations can be rational even in a quite weak case.

Likewise, in either a civil or a criminal case, if getting some win is much more important for the person bringing the case, than getting a "home run" maximal win, making a lenient deal even in a fairly solid case can make sense to avoid the risk of rolling the dice and the irreducible risk of error any time there is a trial.

Also, of course, lots and lots of parties to both criminal and civil cases are not rational actors and make bad decisions. These characteristics of parties to legal cases frequently play a large part in the fact that these parties ended up having to deal with the legal system in the first place.

One of the difficult systemic and institutional issues, however, is that the behavior of people who are irrational because they are dumb or crazy, and the behavior of people in the system who are innocent and have excessive but not necessarily unreasonable trust in the accuracy of the judicial process, can look very similar. People who are factually innocent systemically insist on going to trial even in the face of lenient plea bargains at rates much higher than people who are factually guilty, even in cases that seem to have identical strength before a neutral third-party, and are, as a result, over represented in the ranks of people who actually go to trial.


I think virtually every trial lawyer will tell you that the verdict is never clear until it's in.

There's just so much that can go right or wrong in a trial that it becomes impossible to make reliable forecasts. You don't know who's going to be in your jury pool, and you don't know who's going to be seated from that pool. You don't know if the prosecution's witnesses will actually show up, and you don't know if your witnesses will actually show up. You don't know whether the witnesses will actually come across well on the stand. You don't know if the judge is going to be in a pissy mood and rush the presentation of evidence. You don't know how strictly or leniently the judge is going to exercise his discretion to admit or exclude evidence. As a lawyer, you don't know if you're going to have a good day or a bad day. You don't know if you're going to have a stress-induced fainting spell during opening statements (true story).

Every trial lawyer is aware of these variables, and they're a huge reason that so few cases go to trial.

Beyond that, though, you need to also consider the defendant's incentives. Even if the evidence is so clear that he could accurately predict a conviction, that doesn't always mean that the defendant "should have" taken the plea. The circumstances of a defendant's life might be such that taking a 30-year deal isn't really that much better than a 40-year sentence after conviction.

Likewise, defendants in sex-offense cases face such miserable consequences for the rest of their lives that they might be willing to roll the dice and hope for some incredibly unlikely trial, thinking that the additional time out of prison will not be so spectacular that they're willing to give up any chance of exoneration.

And there are also cases where the likely punishment is so trivial that a defendant doesn't mind risking it. If a defendant is charged with disorderly conduct or trespassing, she may realistically be looking at nothing worse than 40 hours of community service, even if she goes to trial and loses. Those consequences may not matter to her at all, so she may be perfectly comfortable putting the government to the task of convicting her, even if the evidence against her is overwhelming.

You must log in to answer this question.

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