The Supreme Court — the record
The model's full walk-forward backtest, 1970–2023 (53,264 justice-votes), and how it stacks up against the published academic benchmark and every other forecaster we track. ← Back to the current court
How well it forecasts
Scored strictly walk-forward over 1970–2023. The honest bar is the Court's strong habit of siding with the petitioner (it reverses more than it affirms) — so "petitioner always wins" is a tough baseline, and it has only gotten tougher as recent benches grew more lopsided. The durable edge is at the justice level.
Per-justice vote accuracy vs the baseline, by term
The model holds a gap above the "petitioner always wins" line across the era. The gap narrows recently because the baseline itself climbs — modern terms break the petitioner's way more often, leaving less room to beat.
| By issue area | Model | Base | n |
|---|---|---|---|
| Criminal Procedure | 67.9% | 68.7% | 1,471 |
| Economic Activity | 63.5% | 63.4% | 1,087 |
| Civil Rights | 63.7% | 64.2% | 1,084 |
| Judicial Power | 63.4% | 61.6% | 803 |
| First Amendment | 67.4% | 67.0% | 436 |
| Federalism | 61.5% | 64.3% | 286 |
| Due Process | 69.0% | 69.4% | 271 |
| Unions | 57.7% | 59.3% | 194 |
| Federal Taxation | 64.1% | 55.2% | 145 |
| Privacy | 69.0% | 69.0% | 116 |
| Attorneys | 67.8% | 65.5% | 87 |
| Interstate Relations | 55.4% | 51.4% | 74 |
Against the field
SCOTUS has far fewer forecasters than elections, but there is a published academic benchmark — {Marshall}+ (Katz–Bommarito–Blackman, 2017), the model that put machine prediction of the Court on the map. Scored on the same cases, same walk-forward rules, our model edges a faithful reproduction of it.
aiscotus vs {Marshall}+ (reproduction)
same cases · 1971–2023 · 5,911 cases| Metric | aiscotus | {Marshall}+ | baseline |
|---|---|---|---|
| Justice-vote accuracy | 65.2% | 64.9% | 60.6% |
| Justice-vote Brier (lower better) | 0.217 | 0.218 | — |
| Case accuracy (vote aggregation) | 65.7% | 65.7% | — |
| Justice-vote accuracy by decade | 1970s | 1980s | 1990s | 2000s | 2010s | 2020s |
|---|---|---|---|---|---|---|
| aiscotus | 67.0% | 65.6% | 62.0% | 65.4% | 64.5% | 67.0% |
| {Marshall}+ | 66.0% | 66.6% | 61.9% | 64.4% | 62.8% | 67.8% |
| baseline | 61.4% | 58.0% | 57.2% | 64.2% | 63.4% | 66.1% |
{Marshall}+is our faithful reproduction of the Katz–Bommarito–Blackman (2017) method — extremely-randomized trees on the full set of raw SCDB codes — since their actual per-case predictions aren't published. Their reported headline was 71.9% justice / 70.2% case over 1816–2015; that span has a lower base rate than the modern window here, so it's context rather than a like-for-like target. On identical modern cases, our model edges the classic approach.
The model zoo — every forecaster, ranked
Rather than pick one design, we build them all and let the data rank them — forecasting October Term 2025 cases after oral argument. Our model is aiscotus (a per-justice model on the Supreme Court Database — not the election "VotePredictor" model), in a structural variant and one that adds oral-argument analysis, alongside the FantasySCOTUS crowd, prediction markets, and a consensus blend. Two honest takeaways: the informed crowd wins, and analyzing the oral argument is the strongest single signal we add — historically beating our structural model on its own.
| Forecaster | Accuracy | Brier | n |
|---|---|---|---|
| FantasySCOTUS crowd | 95.8% | 0.124 | 24 |
| Consensus (model + crowd + market)not independent | 87.5% | 0.135 | 24 |
| aiscotus | 83.3% | 0.183 | 24 |
| aiscotus + oral argument | 79.2% | 0.179 | 24 |
| "Petitioner always wins" baseline | 83.3% | — | — |
Does analyzing the oral argument help? (2013–2023, walk-forward)
On 637 argued cases with transcripts: the oral-argument signal alone (how hard the justices grill each side) is the strongest single input — it beats our structural model on both accuracy and calibration. None beats the lopsided base rate on raw accuracy, but the signal is real.
| Model | Accuracy | Brier |
|---|---|---|
| aiscotus (structural) | 57.9% | 0.245 |
| Oral argument only | 63.6% | 0.238 |
| aiscotus + oral | 62.6% | 0.245 |
| baseline | 68.0% | — |
| Case | aiscotus | +oral | crowd | market | consensus | result |
|---|---|---|---|---|---|---|
| learning resources v. trump Roberts 67Thomas 66Alito 71Sotomayor 47Kagan 52Gorsuch 64Kavanaugh 68Barrett 67Jackson 54 | 62% | 62% | 57% | — | 59% | Petitioner |
| louisiana v. callais Roberts 63Thomas 46Alito 51Sotomayor 78Kagan 79Gorsuch 69Kavanaugh 64Barrett 69Jackson 74 | 66% | 66% | 41% | — | 50% | Respondent |
| chiles v. salazar Roberts 80Thomas 78Alito 76Sotomayor 46Kagan 47Gorsuch 70Kavanaugh 75Barrett 75Jackson 55 | 67% | 59% | 73% | — | 68% | Petitioner |
| wolford v. lopez Roberts 63Thomas 46Alito 51Sotomayor 78Kagan 79Gorsuch 69Kavanaugh 64Barrett 69Jackson 74 | 66% | 79% | 64% | — | 69% | Petitioner |
| noem v. al otro lado Roberts 73Thomas 81Alito 78Sotomayor 40Kagan 42Gorsuch 69Kavanaugh 75Barrett 73Jackson 55 | 65% | 72% | 68% | — | 69% | Petitioner |
| cox communications v. sony music Roberts 59Thomas 55Alito 57Sotomayor 70Kagan 72Gorsuch 58Kavanaugh 59Barrett 59Jackson 67 | 62% | 82% | 94% | — | 90% | Petitioner |
| fcc v. at&t Roberts 59Thomas 55Alito 57Sotomayor 70Kagan 72Gorsuch 58Kavanaugh 59Barrett 59Jackson 67 | 62% | 62% | 70% | — | 67% | Petitioner |
| chevron usa v. plaquemines parish Roberts 75Thomas 71Alito 76Sotomayor 44Kagan 45Gorsuch 68Kavanaugh 75Barrett 73Jackson 57 | 65% | 55% | 72% | — | 66% | Petitioner |
| monsanto v. durnell Roberts 67Thomas 66Alito 71Sotomayor 47Kagan 52Gorsuch 64Kavanaugh 68Barrett 67Jackson 54 | 62% | 77% | 52% | — | 59% | Petitioner |
| hikma pharmaceuticals v. amarin Roberts 59Thomas 55Alito 57Sotomayor 70Kagan 72Gorsuch 58Kavanaugh 59Barrett 59Jackson 67 | 62% | 58% | 93% | — | 81% | Petitioner |
| exxon mobil v. corporacion cimex Roberts 52Thomas 42Alito 45Sotomayor 52Kagan 57Gorsuch 59Kavanaugh 61Barrett 57Jackson 52 | 53% | 72% | 61% | — | 64% | Petitioner |
| bost v. illinois state board of elections Roberts 52Thomas 42Alito 45Sotomayor 52Kagan 57Gorsuch 59Kavanaugh 61Barrett 57Jackson 52 | 53% | 69% | 79% | — | 76% | Petitioner |
| villarreal v. texas Roberts 56Thomas 44Alito 41Sotomayor 77Kagan 77Gorsuch 60Kavanaugh 58Barrett 58Jackson 72 | 60% | 77% | 16% | — | 39% | Respondent |
| barrett v. united states Roberts 56Thomas 44Alito 41Sotomayor 77Kagan 77Gorsuch 60Kavanaugh 58Barrett 58Jackson 72 | 60% | 72% | 76% | — | 75% | Petitioner |
| bowe v. united states Roberts 56Thomas 44Alito 41Sotomayor 77Kagan 77Gorsuch 60Kavanaugh 58Barrett 58Jackson 72 | 60% | 73% | 61% | — | 65% | Petitioner |
| usps v. konan Roberts 75Thomas 71Alito 76Sotomayor 44Kagan 45Gorsuch 68Kavanaugh 75Barrett 73Jackson 57 | 65% | 50% | 55% | — | 54% | Petitioner |
| olivier v. city of brandon Roberts 82Thomas 55Alito 67Sotomayor 80Kagan 83Gorsuch 76Kavanaugh 78Barrett 80Jackson 74 | 75% | 83% | 88% | — | 86% | Petitioner |
| first choice women's resource centers v. platkin Roberts 82Thomas 55Alito 67Sotomayor 80Kagan 83Gorsuch 76Kavanaugh 78Barrett 80Jackson 74 | 75% | 72% | 94% | — | 87% | Petitioner |
| hencely v. fluor Roberts 59Thomas 55Alito 57Sotomayor 70Kagan 72Gorsuch 58Kavanaugh 59Barrett 59Jackson 67 | 62% | 56% | 64% | — | 62% | Petitioner |
| landor v. louisiana dept. of corrections Roberts 82Thomas 55Alito 67Sotomayor 80Kagan 83Gorsuch 76Kavanaugh 78Barrett 80Jackson 74 | 75% | 78% | 48% | — | 59% | Respondent |
| galette v. new jersey transit Roberts 66Thomas 56Alito 57Sotomayor 76Kagan 77Gorsuch 64Kavanaugh 66Barrett 64Jackson 73 | 66% | 46% | 51% | — | 50% | Petitioner |
| t.m. v. university of maryland medical system Roberts 52Thomas 42Alito 45Sotomayor 52Kagan 57Gorsuch 59Kavanaugh 61Barrett 57Jackson 52 | 53% | 67% | 65% | — | 66% | Respondent |
| cisco systems v. doe i Roberts 73Thomas 81Alito 78Sotomayor 40Kagan 42Gorsuch 69Kavanaugh 75Barrett 73Jackson 55 | 65% | 94% | 74% | — | 80% | Petitioner |
| havana docks v. royal caribbean Roberts 59Thomas 55Alito 57Sotomayor 70Kagan 72Gorsuch 58Kavanaugh 59Barrett 59Jackson 67 | 62% | 85% | 60% | — | 67% | Petitioner |
| trump v. slaughter Roberts 75Thomas 71Alito 76Sotomayor 44Kagan 45Gorsuch 68Kavanaugh 75Barrett 73Jackson 57 | 65% | 65% | 63% | 89% | 72% | pending |
| west virginia v. b.p.j. Roberts 73Thomas 81Alito 78Sotomayor 40Kagan 42Gorsuch 69Kavanaugh 75Barrett 73Jackson 55 | 65% | 26% | 71% | — | 56% | pending |
| little v. hecox Roberts 73Thomas 81Alito 78Sotomayor 40Kagan 42Gorsuch 69Kavanaugh 75Barrett 73Jackson 55 | 65% | 89% | 68% | — | 75% | pending |
Cells show each forecaster's P(petitioner wins); green/red= right/wrong on decided cases. Under each case are aiscotus's per-justice predictions — each justice's % chance of siding with the petitioner (name colored by appointing party, D/R; hover for detail). Forecasts are made after oral argument only. "aiscotus" is our SCOTUS model (per-justice, on the Supreme Court Database) — distinct from the election "VotePredictor" model. Consensus blends model + crowd + market, weighted by each source's track record on this term's decided cases — it ingests the others, so it can't be benchmarked against them. Crowd = FantasySCOTUS; markets = Polymarket / Kalshi.
How this works
A gradient-boosted model predicts, for every justice on a case, the probability they side with the petitioner, using only pre-decision features (issue area, jurisdiction, the lower court's ruling and its ideological direction, cert reason, case origin) plus that justice's own prior record and the Court's composition. A separate case-level model produces the calibrated win probability. Both are trained walk-forward — to forecast a term, only earlier terms are used — so the accuracy above is out-of-sample. Vote data: Supreme Court Database(Spaeth et al.), the standard source for this work. "Ideology" in the profiles is the share of conservative votes, a descriptive proxy rather than a Martin–Quinn ideal point.