Tools chosen: kalshi_direct kalshi_related polymarket_related wikipedia claude_news code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is an ultra-long-horizon geophysical tail-risk question; the base rate for a VEI-8 (supervolcano) eruption is roughly one per 50,000–100,000 years, implying a ~0.03–0.05% chance over 25 years, so the key uncertainties are (a) Kalshi's exact resolution definition of 'supervolcano', (b) the current market price (which on such long-dated markets often trades far above the physical base rate due to time-value/illiquidity), and (c) whether any current unrest (Campi Flegrei, Yellowstone, Taupo) meaningfully shifts near-term odds.

# Current state Kalshi's YES price (26¢) sits far above any base-rate-implied fair value (Poisson analysis puts true VEI-8 probability by 2050 at ~0.02–0.14¢), reflecting long-shot/catastrophe-premium pricing rather than volcanological odds; no VEI-8 eruption is imminent at any monitored caldera, though Campi Flegrei shows the most active unrest of any known supervolcano system. # Timeline of key events - 1950–1984 (confirmed): Campi Flegrei underwent three bradyseismic unrest episodes (1950–52, 1969–72, 1982–84) with major uplift/seismicity, none culminating in eruption. [PMC] - 2005–2018 (confirmed): Current Campi Flegrei unrest phase begins in 2005; seismicity notably increases from 2018. [Stanford/OPB] - 2022–2025 (confirmed): Retrained ML seismic catalog shows Campi Flegrei quake count grow from ~12,000 to >54,000; no evidence yet of magma migrating upward or seismicity below 3.7km ring-fault depth. [Science.org] - Jan–Aug 2025 (confirmed): Five M4+ earthquakes at Campi Flegrei — INGV director calls it "the strongest we've ever seen." [OPB] - 2022–2023 (reported): Taupō (NZ) showed unrest ("stirred") but has since calmed. [Orbital Today/VolcanosAtlas] - 2026 (reported): New Nature Communications Earth & Environment modeling study concludes a critically pressurized, potentially eruptible magma reservoir exists at ~4km depth under Campi Flegrei today, though eruption timing remains unforecastable. [Nature.com] # Event Will a supervolcano erupt before Jan 1, 2050? (Kalshi: KXERUPTSUPER-0-50JAN01) # Outcomes to forecast Yes / No # Kalshi market anchor Current YES price: **26.00%** (as of latest data). 7-day change +3.8%, 30-day change +2.0%. Average daily volume only ~48–114 contracts (thin, illiquid market). 157-day price range: 14.0%–32.9%, showing meaningful drift upward over time despite no actual eruption occurring. [kalshi_direct] # Sub-question answers 1. **Current Kalshi price/history?** YES trades at 26%, up from a 14% floor and toward a 32.9% high over the past ~157 days; volume is thin (~48–114 contracts/day), suggesting price is sentiment/news-driven rather than deeply liquid. [kalshi_direct] 2. **Definition used for resolution?** No explicit rules text provided; secondary reporting frames it as a "civilization-ending eruption," implying strict VEI-8 supereruption standard, not any large-but-sub-VEI-8 event (e.g., Campi Flegrei's historic max is VEI-7). Official resolution source/agency not specified in available data. [claude_news/covers.com; Wikipedia-Supervolcano] 3. **Geological base rate?** VEI-8 eruptions occur roughly every 17,000–100,000+ years historically (last one: Taupō ~25,600 years ago). Poisson modeling over the 24.1-year window to 2050 yields P(≥1 eruption) ≈ 0.02%–0.14%, i.e., a fair YES price of roughly 0.02¢–0.14¢. [code_execution; Wikipedia-Supervolcano] 4. **Current unrest at known calderas?** Campi Flegrei is the most active/"nervous" system — record seismicity (54,000+ quakes 2022–2025, five M4+ quakes in 2025) and a 2026 study suggesting an eruptible magma reservoir at ~4km depth; still no signs of imminent supereruption, and its historical max is VEI-7 (sub-supervolcano threshold). Yellowstone: USGS puts annual VEI-8 probability at 1-in-730,000, "not considered eruptible" currently. Taupō calmed after 2022-23 unrest; Toba shows only hot springs; no system shows imminent supereruption signs. [claude_news; Wikipedia-Yellowstone/Phlegraean Fields] 5. **Comparable markets/premiums?** No Polymarket matches found (0 markets on supervolcano/volcano/Yellowstone keywords). One aggregator (Octagon AI) cites a "prediction market" figure of 31.5% for supervolcano eruption by 2050 — likely reflecting a looser definition than strict VEI-8, consistent with broad tail-risk/catastrophe premium seen across Kalshi's "long-shot" series. [polymarket_related; claude_news] # Key facts (high-confidence, factual) 1. VEI-8 = eruption volume >1,000 km³; only defining criterion for "supervolcano" eruption. [Wikipedia-Supervolcano] 2. Most recent VEI-8 eruption: Taupō ~25,600 years ago; Toba ~74,000 years ago. [Wikipedia] 3. USGS: annual Yellowstone VEI-8 probability ≈ 1/730,000; not presently eruptible. [claude_news] 4. Campi Flegrei's largest known eruptions are VEI-7, not VEI-8 — it is popularly but not technically a "supervolcano." [Wikipedia-Phlegraean Fields] 5. Kalshi YES = 26%, thin volume (~50-100 contracts/day). [kalshi_direct] # Cross-market signals - Kalshi related: same market only (no sibling contracts found); series otherwise empty. - Polymarket: no matching markets exist for supervolcano/volcano/Yellowstone. - Sportsbook implied: none available. - Aggregator (Octagon AI, non-exchange): cites 31.5% figure, possibly loosely defined; claims "no compelling mispricing" — but this conflicts sharply with base-rate math and should be discounted as likely reflecting a different/looser event definition. # Analyst opinions and speculation - Octagon AI model: sees Campi Flegrei's unrest as more significant than Yellowstone's but stops short of predicting eruption before 2050. [claude_news] - Scientific consensus (Bulletin of Volcanology 2026 review): caldera unrest evolution is fundamentally hard to forecast; source of Campi Flegrei unrest (magma vs. hydrothermal fluid) remains debated. [Springer] - Historical base-rate modeling strongly implies markets are irrationally overpricing tail catastrophe risk (7x–100x+ fair value), a known behavioral pattern in "long-shot" prediction markets. [code_execution] # Directional lean per outcome - **Yes**: Supported only by generic unrest narratives (Campi Flegrei seismicity) and social/media salience; not supported by any VEI-8-specific precursor signal, and Campi Flegrei's historical max (VEI-7) doesn't even meet the "supervolcano" bar. - **No**: Overwhelmingly favored by geological base rates (fair value <<1%), USGS's explicit "not eruptible" assessments, and lack of any credible near-term supereruption forecast from volcanologists. # Gaps / unknowns - Exact Kalshi settlement/resolution criteria (source agency, VEI-8 requirement, qualifying volcano list) not confirmed in available text. - No transparency on why Kalshi price has drifted from 14% to 26%+ despite no change in base-rate science — likely media-driven sentiment (Campi Flegrei coverage) rather than new geological evidence. # Calibration anchors - Kalshi current YES price: **26%** (anchor, likely badly overpriced vs. physical base rate). - Poisson base-rate fair value: **~0.02%–0.14%** over the remaining window. - Precedent: long-shot Kalshi/Polymarket catastrophe markets (e.g., extinction, alien contact) typically trade at 1–5x geological/scientific base rates due to behavioral premium; a 26% price for a sub-0.2% event is an extreme outlier even by that standard.

1. Both forecasters converge near 3-4x the base rate (0.04, 0.025 vs ~0.02-0.14%) citing "resolution ambiguity," but neither quantifies how likely a loose resolution actually is — the brief flags an aggregator citing 31.5% "likely reflecting a different/looser event definition," which could imply resolution ambiguity risk is much larger than a mere 20-30x multiplier, or conversely could be irrelevant noise; this key uncertainty is asserted but not decomposed. 2. Neither forecast explicitly engages with why Kalshi drifted from 14%→26% over 157 days despite "no change in base-rate science" — if this reflects a real update mechanism (e.g., market-makers pricing in resolution-source risk they know but the brief doesn't), both may be underweighting information asymmetry rather than assuming pure sentiment/media noise. 3. Both forecasts land far below the Kalshi anchor (26%) without seriously grappling with the possibility that Kalshi's resolution language could default to "any officially classified supervolcano-scale event" reported by media/USGS during the window — given 24 years is a long horizon and volcanic classification disputes are common (e.g., Campi Flegrei's VEI-7 vs "supervolcano" popular label), the tail risk of a favorable-to-YES resolution technicality may deserve more than a 20x bump over base rate. 4. Neither model addresses the long 24-year horizon's compounding effect on "unknown unknowns" — e.g., potential newly identified caldera systems, reclassification of VEI thresholds, or discovery of another unrest episode elsewhere (Toba, Taupō re-activation) that could independently contribute non-trivial tail probability beyond just Campi Flegrei. 5. Both forecasters treat the two considerations (base rate vs. resolution ambiguity) as simply additive without giving a probabilistic breakdown (e.g., P(resolution is loose) × P(loose-criteria event happens by 2050)), which is a modeling gap since a rigorous decomposition might justify a rate anywhere from 1% to 10%+ depending on that split — the point estimates of 2.5-4% appear to be intuition-anchored rather than derived from explicit sub-probabilities. 6. Neither critique interrogates the "long-shot behavioral premium" precedent cited in the brief (1-5x base rate for similar catastrophe markets) — since 26% is 185x-1300x the base rate (far outside that reference class), this extreme outlier should either prompt higher suspicion of an information gap in resolution criteria (favoring higher YES) or reinforce that the market is uniquely irrational here (favoring even lower YES than both forecasts show), and neither forecaster fully resolves this tension.

4% (75% confidence)

Yes: 4% No: 96%

Geology dominates: VEI-8 supereruptions recur every ~17,000–100,000 years, giving a Poisson probability of ~0.02–0.14% over the remaining 24-year window, with USGS rating Yellowstone 'not eruptible' and no monitored caldera showing supereruption precursors (Campi Flegrei's record unrest tops out historically at VEI-7, below the supervolcano bar). The 26¢ Kalshi anchor is a long-dated, illiquid contract where 24-year capital lockup blocks NO-side arbitrage, so the price reflects carry cost plus media-driven catastrophe premium rather than volcanological odds — its 14%→26% drift tracked Campi Flegrei coverage, not new science. Taking the critique's decomposition seriously: P(loose/ambiguous resolution counting a VEI-7 or a 'supervolcano caldera' eruption) ≈ 15–20%, times P(such an event by 2050) ≈ 10–20%, yields ~2–4% from definitional risk, plus a negligible strict-VEI-8 term and a small allowance for unknown-unknowns over a long horizon. That justifies a final ~4.5% YES — slightly above both forecasters to honor resolution ambiguity and the market's persistent premium, but far below 26%, which is 200x+ the physical base rate and an extreme outlier even for long-shot markets.

Total pipeline time: 120.9s