Move one price, revalue every book it touches, and test each account against its maintenance requirement.Note 16
Forced exits read 2026-11-22 01:12 UTC, once a day rather than on each visit. The register keeps the newest 5,000; it holds fewer than that, so this is all of them.
Every instrument held by a measurable account (179), the closest to a liquidation first; the figure is the smallest move, either way, that breaks one.
| Shock | APE price | Liquidated | Equity that goes | Share of measured |
|---|---|---|---|---|
| As APE rises | ||||
| 50% | 0.10868 | 0 | — | — |
| 40% | 0.10143 | 0 | — | — |
| 30% | 0.09419 | 0 | — | — |
| 20% | 0.08694 | 0 | — | — |
| 15% | 0.08332 | 0 | — | — |
| 10% | 0.0797 | 0 | — | — |
| 5% | 0.07607 | 0 | — | — |
| now | 0.07245 spot | 0 | — | — |
| As APE falls | ||||
| -5% | 0.06883 | 0 | — | — |
| -10% | 0.06521 | 0 | — | — |
| -15% | 0.06158 | 0 | — | — |
| -20% | 0.05796 | 0 | — | — |
| -30% | 0.05071 | 0 | — | — |
| -40% | 0.04347 | 0 | — | — |
| -50% | 0.03623 | 0 | — | — |
Only the named instrument moves. This is the observed damage under that one scenario, not a bound: a market-wide fall can make a correlated book worse than this and a hedged book better, so it is neither a floor nor an estimate of what a real fall would do.
A scenario, not a bound.
Only APE moves here; every other price a book holds stays where it is, so a book hedged in another asset looks worse than it would, and one exposed to assets that fall together looks better. Each book is judged on its whole account, its equity against the maintenance margin all of its positions require at the venue’s ceiling, not on the leverage one position was opened at. Books with no margin reading are counted above and not simulated.
The same computation is at GET /api/v1/stress, where ?scale=sigma places the rungs by how far out they are rather than at round numbers. The API also has the cascade, what the forced selling from these liquidations does to everyone else.