Update 11/1/25:
The updated model (experience-3) continues to predict reversal, with roughly the same probability (70 percent). Vote ordering remains roughly the same, with possible dissents from a conservative bloc.
Experience-3 predictions
Update 10/22/25:
The updated model (experience-2) continues to predict reversal, now seeing this as less of a close case (probability of reversal 68 percent after argument). The model continues to see a conservative bloc as the most likely to break off and vote to affirm.
Updated vote predictions
Initial post:
Barrett v. US is another interesting criminal case involving the boundaries of double jeopardy. Normally, a person may not be tried twice for the same offense—this is the meaning of the double jeopardy clause of the Constitution. However, Congress may create multiple offenses for the same conduct. For instance, someone who robs a store and destroys a safe in the process may be criminally liable for both robbery and an offense related to the destruction of property.
This case boils down to a question of statutory interpretation—did Congress create two offenses or one? The facts involve an armed robbery that resulted in the death of someone trying to stop the robbery. One statutory provision criminalizes using a firearm to advance a federal crime of violence; another provision relates to this same class of conduct, but where a death is caused. Normally, courts try to figure out statutory meaning in this context by looking at what facts or elements would need to be shown. If the second candidate offense requires different elements to be shown, then it is considered a separate offense; if not, then a single offense. The lower court (Second Circuit) determined that Congress created two separate offenses that may be punished cumulatively, pointing to the structure of the statute and the different penalty structure for each provision.
Barrett appealed to the Court, and the government actually filed a brief supporting the defendant on this question. Both before and after argument, the model predicts a reversal, though it is a close case. Alito and Jackson appear to be the closest votes.
Pre- and Post-Argument Predictions