Tools chosen: kalshi_direct kalshi_related polymarket_related wikipedia claude_news gdelt_news code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a long-horizon natural-hazard question best answered by a Poisson base rate from the historical catalog of M8.0+ events with epicenters in Japan/its territorial waters, adjusted for resolution-source ambiguity (JMA Mj vs USGS Mw, offshore trench events outside territorial waters) and for official Nankai Trough / Japan Trench hazard forecasts. The Kalshi price is the primary anchor, with cross-venue and related-market checks. Math tooling converts the catalog frequency into a ~4-year probability.

# Current state The market resolves YES if any M≥8.0 earthquake (epicenter in Japan or territorial waters) occurs before 2030-01-01. As of now (research dated mid/late 2026), no qualifying M8.0+ event has been confirmed since the 2011 Tōhoku quake; the window remaining is roughly 3.5 years. Kalshi currently prices this at 45% YES, well above historical base-rate estimates (~11-25%), likely reflecting elevated attention to Nankai Trough megaquake warnings. # Timeline of key events - 1900-2011: At least 10-14 M8.0+ earthquakes near Japan historically (1923 Kanto M8.1, 1933 Sanriku M8.5, 1946 Nankaido M8.1, 2003 Tokachi-oki M8.0, 2006 Kuril Islands M8.4, 2011 Tōhoku M9.0-9.1) — confirmed (Wikipedia, VolcanoDiscovery). - 2013-2024: Japan's Earthquake Research Committee (ERC) maintains ~70-80% 30-year probability estimate for Nankai Trough M8-9 event — confirmed (japantimes.co.jp). - 2024-08-08: M7.1 quake on western edge of Nankai Trough triggers JMA's first-ever "Nankai Trough Earthquake Extra Information" advisory; experts noted next-day risk still <1% — confirmed (Fox/AP, CNN). - 2025-01: ERC revises Nankai Trough 30-year probability upward to "more than 80%" (from 70-80%) — confirmed (gulfnews.com, aol.com). - 2025-09-27: ERC further revises range to 60%-94.5%+ for M8-9 Nankai event, alongside a separate 20-50% figure, causing public confusion over methodology — confirmed (japantimes.co.jp). - 2026-01-16: Japan Trench/Chishima Trench probabilities raised (Miyagi M7.4 to 79-95%; Nemuro/Hokkaido M7.8-8.5 to ~90%) — confirmed (japantimes.co.jp). - 2026-07: Kumamoto-area earthquake death toll reported rising; multiple outlets ran explainer pieces on why Japan is quake-prone (magnitude/scale of this event unspecified in research) — reported, magnitude not confirmed as M8+. # Event Will an M≥8.0 earthquake with epicenter in Japan or its territorial waters occur before 2030-01-01? # Outcomes to forecast Yes / No # Kalshi market anchor KXEARTHQUAKEJAPAN-30 currently trades at **45% YES** (kalshi_direct). 7-day change +5%, 30-day change -19% (range 31%-77% over 147 data points), avg daily volume ~178-303 contracts. High volatility and a wide historical range suggest thin liquidity and sentiment-driven swings rather than a stable consensus. # Sub-question answers 1. **Kalshi price/volume/trend?** — 45% YES, recent uptick (+5% 7-day) after a sharp prior decline (-19% 30-day); volume modest (~178-303 contracts/day) (kalshi_direct). 2. **Historical frequency per decade?** — Roughly 9-14 M8.0+ events in/near Japan since 1900 (125 yrs), implying ~1 event per 9-14 years broadly defined; narrower on-land/territorial-waters definition yields only 3-5 events (~1 per 25-40 years) (claude_news, code_execution). 3. **Magnitude scale & territorial waters issue?** — Rules don't specify Mw vs Mj explicitly; most historical/anticipated Nankai Trough and Japan Trench epicenters lie in the offshore subduction zone/EEZ, which is ambiguous relative to a strict 12nm "territorial waters" reading — a key resolution ambiguity not clarified in any source (rules text, Wikipedia "Territorial waters"). 4. **ERC Nankai Trough 30-yr probability?** — Revised Sept 2025 to 60%-94.5%+ (or alternatively 20-50% per a separate calculation method), up from ~70-80% pre-2025; translating a ~75% 30-yr figure to a 4-yr window via Poisson gives roughly 12-15% (japantimes.co.jp; code_execution methodology). 5. **Recent elevated risk signals?** — Aug 2024 JMA Nankai Trough advisory (first ever) after M7.1 quake; Jan 2026 upward revisions for Japan Trench (Miyagi) and Chishima Trench (Nemuro) M7.4-8.5 probabilities; no confirmed M8.0+ event yet (claude_news, gdelt_news). 6. **Other venues?** — No Polymarket market exists for this specific event (only an unrelated Iran airspace market surfaced). Kalshi's own California M8+ markets price much lower: California-2035 at 30%, California-2028 at 15%, suggesting Japan's 45% is not directly benchmarked against comparable US markets (kalshi_related). # Key facts (high-confidence, factual) 1. [Wikipedia] 2011 Tōhoku M9.0-9.1 is the largest and most recent confirmed M8+ Japan event. 2. [claude_news/VolcanoDiscovery] ~9-14 M8.0+ quakes near Japan since 1900, averaging roughly one per 10-15 years (broad definition). 3. [japantimes.co.jp] ERC Nankai Trough 30-yr M8-9 probability revised to 60-94.5%+ in Sept 2025 (from ~70-80%). 4. [Fox/CNN, Aug 2024] JMA issued first-ever Nankai Trough megaquake advisory after M7.1 quake; next-day risk still <1%. 5. [japantimes.co.jp, Jan 2026] Japan Trench (Miyagi) and Chishima Trench (Nemuro) 30-yr M7.4-8.5 probabilities raised to 79-95% and ~90% respectively — these are sub-M8.0-adjacent, not necessarily M8.0+. # Cross-market signals - Kalshi related: California M8+ by 2035 = 30%, by 2028 = 15% — Japan (45%) prices meaningfully higher, consistent with Japan's higher seismic hazard profile. - Polymarket: No directly matching market found. - Sportsbook implied: N/A. # Analyst opinions and speculation - CNN (Aug 2024) cautioned against overreacting to short-term risk upticks despite advisories; day-after probability remains "less than 1%." - Japan Times/Tokyo Weekender flagged public confusion from ERC's dual probability figures (60-94.5% vs 20-50%), indicating genuine expert disagreement on methodology, not just a media artifact. - code_execution analysis suggests fair value ~11-24% based on Poisson extrapolation of historical rates; Kalshi's 45% appears materially above this range. # Directional lean per outcome - **Yes**: Elevated ERC Nankai Trough probabilities (60-94.5% over 30 yrs), 2024 advisory precedent, active subduction zones (Nankai, Japan Trench, Chishima Trench) all with rising 30-yr estimates; Japan is among the most seismically active regions on Earth. - **No**: Historical base rate over any 4-year window is low (~10-25% per Poisson modeling); the 2011 Tōhoku event released substantial regional stress reducing near-term likelihood of an immediate repeat in that segment (2019 assessment: near-zero chance of Tohoku-scale repeat soon); ambiguity in "territorial waters" definition could exclude many offshore epicenters, lowering effective resolution rate. # Gaps / unknowns - No official Kalshi rules clarify magnitude source (JMA vs USGS Mw) or precise geographic boundary ("territorial waters" vs EEZ/trench) — critical for resolution given most historical/anticipated large events are offshore. - Unclear whether the 2026-07 Kumamoto earthquake mentioned in gdelt_news reached M8.0 (no magnitude given) — needs verification. - No Polymarket comparable market to cross-check pricing. - ERC's 30-year Nankai figures don't map cleanly to a 4-year window; no official government short-window forecast exists. # Calibration anchors - Kalshi current YES price: **45%** (anchor, but volatile: 31-77% range historically). - Poisson-derived fair value from historical frequency: ~11-25% (code_execution). - Comparable Kalshi California M8+ markets: 15% (2028), 30% (2035) — much lower than Japan's 45%, despite Japan's higher hazard, suggesting Japan market may be overpriced relative to structural base rates but could reflect genuine elevated regional risk (Nankai/Japan Trench advisories).

1. Both forecasters converge tightly (0.28-0.31) despite acknowledging multiple independent hazard zones (Nankai Trough, Japan Trench/Miyagi, Chishima Trench/Nemuro) each carrying elevated 30-yr probabilities — treating these as a single Poisson process rather than summing near-independent trigger sources likely understates tail probability, since three active zones each with even a 10-15% four-year-equivalent chance would push combined probability meaningfully higher than either single-zone estimate suggests. 2. Neither forecast grapples with the resolution ambiguity flagged in the brief (Mw vs Mj magnitude source, and "territorial waters" vs EEZ/subduction zone definition) as a probability-moving factor — this is treated as a footnote, but if the strict 12nm reading is applied, nearly all historical and anticipated Nankai/Trench-zone quakes would be excluded, which should push the estimate toward the low end of the historical range (~1 per 25-40 years, i.e., <15%), yet both landed near 28-31%, implying they didn't fully price in this exclusionary scenario. 3. Both cite the California M8+ comparables (15% for 2028, 30% for 2035) as evidence Kalshi's Japan price is inflated, but neither addresses that Japan's baseline seismic hazard is objectively far higher than California's, making this cross-market comparison weak grounds for anchoring down — the comparison may be a red herring rather than a calibration check. 4. The brief notes a 2026-07 Kumamoto earthquake with "death toll reported rising" and unconfirmed magnitude — neither forecaster explicitly assigns residual probability mass to this event potentially already being a qualifying M8+ (or triggering follow-on activity), an unresolved fact that could materially move the estimate if verified. 5. Both forecasters shade below the Kalshi anchor by similar magnitudes (14-17 points) using nearly identical reasoning chains, suggesting insufficient independent variation in method — given the explicit expert disagreement noted in the brief (ERC's dual 60-94.5% vs 20-50% methodologies), a wider or bimodal treatment of uncertainty (e.g., a low scenario ~15% and high scenario ~45%) might better reflect genuine unresolved methodological disagreement rather than collapsing to a single point near 30%. 6. Neither rationale addresses the recent 7-day uptick (+5%) after the 30-day decline (-19%) as a potential signal of fresh information (e.g., new advisories or the Kumamoto event) that the market is currently pricing in real-time — both treat the volatility purely as noise/thin-liquidity, but recent directional momentum could reflect legitimate updating that shouldn't be fully discounted.

30% (low-medium confidence)

Yes: 30% No: 70%

The historical Poisson base rate for M8.0+ events in/near Japan (~9-14 since 1900, i.e., ~0.07-0.11/yr) over the ~3.3 years remaining gives roughly 21-31%, and a strict 'territorial waters' reading would push the effective rate lower still. The ERC's elevated Nankai Trough 30-year figures translate to only ~12-15% over this short window, and the Japan Trench/Chishima revisions concern M7.4-8.5 events that need not reach M8.0 — the critique's point about summing independent zones is largely already embedded in the aggregate historical count, so it does not justify a large upward move. I discount the 45% Kalshi anchor substantially given thin volume (~200 contracts/day), a 31-77% historical range, and evident headline-driven sentiment around Nankai advisories, but I do not collapse to the narrowest base rate: the market anchor, genuine expert-flagged elevated hazard, and the unverified 2026-07 Kumamoto activity warrant some upward allowance. Balancing the exclusionary offshore-resolution scenario against the possibility of generous broad-region resolution, I land at 30% Yes — slightly above the two forecasters' midpoint and well below the market.

Total pipeline time: 306.1s