Each scenario's past episodes held out against the cone of the others, and the house futuretests started by real events and tracked against their frozen cone. Misses included.
This is a leave-one-out check. For each scenario, every past episode is held out in turn. The cone is built from the other episodes, the ones after the held-out episode as well as the ones before it, leaving out every one within 63 sessions of it: their paths overlap its own, and counting them would make the band look better than it is. Because later episodes are used, this is not what the cone would have said at the time; the house futuretests below are that test. The held-out episode's real path is checked against that cone: inside the band, under it, or over it. Each count shows n, the episodes scored at that horizon. Once ten episodes remain the band is the 10th to 90th percentile, so about eight in ten should land inside if the cone is drawn at the right width. With fewer, the band is the others' min to max, so the two most extreme episodes always land outside it: a small record cannot score better than that, and the page says which band each count used.
| SCENARIO | BOOK | EPISODES | 5 SESSIONS: INSIDE / UNDER / OVER | 21 SESSIONS: INSIDE / UNDER / OVER | 63 SESSIONS: INSIDE / UNDER / OVER |
|---|---|---|---|---|---|
| War or a blockade in the Taiwan Strait | The S&P 500 | 42 | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) |
| War or a blockade in the Taiwan Strait | 60/40: the S&P 500 and long Treasuries | 42 | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) | 32 / 5 / 5 (n 42, 76% inside, 10th to 90th percentile) |
| Escalation in the Gulf | The S&P 500 | 20 | 13 / 3 / 2 (n 18, 72% inside, 10th to 90th percentile) | 13 / 3 / 2 (n 18, 72% inside, 10th to 90th percentile) | 11 / 4 / 3 (n 18, 61% inside, 10th to 90th percentile) |
| Escalation in the Gulf | 60/40: the S&P 500 and long Treasuries | 20 | 13 / 3 / 2 (n 18, 72% inside, 10th to 90th percentile) | 12 / 3 / 3 (n 18, 67% inside, 10th to 90th percentile) | 12 / 3 / 3 (n 18, 67% inside, 10th to 90th percentile) |
| Attacks on Red Sea shipping | The S&P 500 | 7 | 1 / 1 / 0 (n 2, 50% inside, min to max) | 1 / 1 / 0 (n 2, 50% inside, min to max) | 1 / 1 / 0 (n 2, 50% inside, min to max) |
| Attacks on Red Sea shipping | 60/40: the S&P 500 and long Treasuries | 7 | 0 / 1 / 1 (n 2, 0% inside, min to max) | 1 / 1 / 0 (n 2, 50% inside, min to max) | 1 / 1 / 0 (n 2, 50% inside, min to max) |
| Russia and NATO escalation | The S&P 500 | 6 | 2 / 2 / 1 (n 5, 40% inside, min to max) | 3 / 1 / 1 (n 5, 60% inside, min to max) | 3 / 1 / 1 (n 5, 60% inside, min to max) |
| Russia and NATO escalation | 60/40: the S&P 500 and long Treasuries | 6 | 2 / 2 / 1 (n 5, 40% inside, min to max) | 3 / 1 / 1 (n 5, 60% inside, min to max) | 3 / 1 / 1 (n 5, 60% inside, min to max) |
| A US tariff shock | The S&P 500 | 6 | 4 / 1 / 1 (n 6, 67% inside, min to max) | 4 / 1 / 1 (n 6, 67% inside, min to max) | 4 / 1 / 1 (n 6, 67% inside, min to max) |
| A US tariff shock | 60/40: the S&P 500 and long Treasuries | 6 | 4 / 1 / 1 (n 6, 67% inside, min to max) | 3 / 2 / 1 (n 6, 50% inside, min to max) | 4 / 1 / 1 (n 6, 67% inside, min to max) |
| An oil supply shock | The S&P 500 | 5 | 2 / 1 / 1 (n 4, 50% inside, min to max) | 2 / 1 / 1 (n 4, 50% inside, min to max) | 2 / 1 / 1 (n 4, 50% inside, min to max) |
| An oil supply shock | 60/40: the S&P 500 and long Treasuries | 5 | 2 / 1 / 1 (n 4, 50% inside, min to max) | 2 / 1 / 1 (n 4, 50% inside, min to max) | 2 / 1 / 1 (n 4, 50% inside, min to max) |
| A pandemic declaration | The S&P 500 | 5 | 1 / 0 / 1 (n 2, 50% inside, min to max) | 0 / 1 / 1 (n 2, 0% inside, min to max) | 1 / 0 / 1 (n 2, 50% inside, min to max) |
| A pandemic declaration | 60/40: the S&P 500 and long Treasuries | 5 | 1 / 0 / 1 (n 2, 50% inside, min to max) | 1 / 0 / 1 (n 2, 50% inside, min to max) | 1 / 0 / 1 (n 2, 50% inside, min to max) |
| A Fed decision | The S&P 500 | 46 | 36 / 5 / 5 (n 46, 78% inside, 10th to 90th percentile) | 34 / 6 / 5 (n 45, 76% inside, 10th to 90th percentile) | 33 / 6 / 5 (n 44, 75% inside, 10th to 90th percentile) |
| A Fed decision | 60/40: the S&P 500 and long Treasuries | 46 | 35 / 5 / 6 (n 46, 76% inside, 10th to 90th percentile) | 34 / 6 / 5 (n 45, 76% inside, 10th to 90th percentile) | 33 / 5 / 6 (n 44, 75% inside, 10th to 90th percentile) |
The same reference books, armed on every scenario. The desk starts one by itself when a matching event lands, freezes the cone that day, and tracks the book against it for 63 sessions. Finished ones stay here for good. 0 finished: 0 inside the band, 0 under, 0 over. 12 still tracking, not counted above (where they sit now, not a result): 4 inside, 0 under, 0 over, 8 too early to say.
| SCENARIO | BOOK | STARTED BY | SESSIONS | BOOK NOW | CONE MEDIAN | WHERE |
|---|---|---|---|---|---|---|
| Russia and NATO escalation | The S&P 500 | Russian drone and missile strikes on residential buildings and other civilian sites in Kyiv, Ukraine, kill seven people, including a 14-year-old boy, and injure 46 others. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| A US tariff shock | The S&P 500 | The United States and China reach an agreement to cut tariffs on $30 billion of goods and launch a dialogue on artificial intelligence. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| A US tariff shock | 60/40: the S&P 500 and long Treasuries | The United States and China reach an agreement to cut tariffs on $30 billion of goods and launch a dialogue on artificial intelligence. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| An oil supply shock | 60/40: the S&P 500 and long Treasuries | Strait of Hormuz: Iran offers US 7 - day plan to end war , reopen Strait of Hormuz 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| An oil supply shock | The S&P 500 | Strait of Hormuz: Iran offers US 7 - day plan to end war , reopen Strait of Hormuz 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| A pandemic declaration | The S&P 500 | A laboratory-confirmed mpox case from a 21-year-old woman is reported in Bunia, Ituri Province, Democratic Republic of the Congo. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| Russia and NATO escalation | 60/40: the S&P 500 and long Treasuries | Russian drone and missile strikes on residential buildings and other civilian sites in Kyiv, Ukraine, kill seven people, including a 14-year-old boy, and injure 46 others. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| A pandemic declaration | 60/40: the S&P 500 and long Treasuries | A laboratory-confirmed mpox case from a 21-year-old woman is reported in Bunia, Ituri Province, Democratic Republic of the Congo. 2026-09-25 | 0 of 63 | - | - | TOO EARLY |
| Attacks on Red Sea shipping | 60/40: the S&P 500 and long Treasuries | France announces that it will deploy its soldiers and air defense systems in Yanbu, Saudi Arabia, to support the Saudi Arabian Armed Forces against the Houthis. 2026-09-24 | 1 of 63 | -0.5% | +0.0% | INSIDE THE BAND |
| Escalation in the Gulf | The S&P 500 | The Israel Defense Forces and Shin Bet say that they have killed Hamas's chief financial officer in the Khan Younis area of the Gaza Strip. 2026-09-24 | 1 of 63 | -0.0% | -0.1% | INSIDE THE BAND |
| Escalation in the Gulf | 60/40: the S&P 500 and long Treasuries | The Israel Defense Forces and Shin Bet say that they have killed Hamas's chief financial officer in the Khan Younis area of the Gaza Strip. 2026-09-24 | 1 of 63 | -0.5% | +0.0% | INSIDE THE BAND |
| Attacks on Red Sea shipping | The S&P 500 | France announces that it will deploy its soldiers and air defense systems in Yanbu, Saudi Arabia, to support the Saudi Arabian Armed Forces against the Houthis. 2026-09-24 | 1 of 63 | -0.0% | -0.1% | INSIDE THE BAND |
4 armed and waiting.
0 tracking right now: 0 inside the band, 0 under, 0 over, 0 too early to say. 0 more have run their full window (each member picks its length) and are not counted as tracking. Their books are theirs; only the counts are shown.
RUN YOUR OWN FUTURETEST