Scan: Will there be an at least 8.0 magnitude earthquake in Japan before 2030?

← Back to scans

Will there be an at least 8.0 magnitude earthquake in Japan before 2030?

KXEARTHQUAKEJAPAN-30 · Climate and Weather · 2026-05-08

48%

Agent

58%

Market Price

-10.0%

Edge

40%

Confidence

Volume: 23,214

Spread: 1.0c

Days to resolution: 1333

Markets in event: 1

Final Rationale

My Poisson base rate of ~30-35% over the 3.65-year remaining window, combined with Nankai Trough assessments (60-90%/30yr implying 10-25% Nankai-alone for this window) and active Japan Trench stress redistribution, supports a forecast above pure base rate. I accept Challenges 3 and 4 from the critique: I was likely underweighting the Nankai upper-bound and the broad geographic interpretation (EEZ + Kuril/Ryukyu boundaries), which justifies an upward adjustment to ~46-48%. I partially accept Challenge 1 (time window slightly shorter than I modeled, minor downward pressure) and reject the double-counting concern in Challenge 2 since sq2's elevated indicators reflect real updated stress states not fully captured by long-run base rates. The market at 58% likely overweights the recent April 2026 Sanriku quake and JMA advisory salience — advisories historically have low conversion-to-event rates, so I deviate ~10pp below market while moving meaningfully up from my independent estimate.

Price vs. prediction

How the market moved after the agent's call. Solid blue line is the actual market price; dashed line is the agent's static forecast at $t{=}0$.

$n = 58$ follow-ups

Re-scan Context

This market has been scanned before. Previous predictions:

Date	Predicted	Market Price	Confidence
2026-05-01	42%	66%	40%
2026-04-05	43%	54%	45%

1. Planner Opus · Fermi decomposition + tool plan

Tools chosen: wikipedia_lookup article_search web_search kalshi_data code_execution

Sub-questions (Fermi decomposition)

Based on historical frequency, is there at least a 50% probability that Japan experiences an M8.0+ earthquake in any given 4-year period? (w=0.40)
Are current seismic indicators (stress accumulation, recent seismic activity) elevated relative to baseline, suggesting an increased near-term probability of M8.0+ earthquake in Japan? (w=0.30)
Will Japan's Nankai Trough or other identified high-risk fault zones produce an M8.0+ earthquake before 2030? (w=0.30)

Planner reasoning

This question requires estimating the probability of a rare but recurrent geophysical event in a well-documented seismic region. Japan sits at the intersection of four tectonic plates (Pacific, Philippine Sea, Eurasian, North American), making it one of the world's most seismically active regions. **Reference class analysis**: The key is establishing the historical rate of M8.0+ earthquakes in Japan's territory. Since 1900, Japan has experienced roughly 8-15 M8.0+ events depending on exact catalog and geographic boundary definitions (~0.06-0.12/year). Using a Poisson model, the probability of at least one such event in a 3.65-year window is calculable. **Causal mechanisms**: Japan's seismicity is driven by subduction of the Pacific and Philippine Sea plates. The Nankai Trough is a particularly well-identified megathrust capable of M8-9 earthquakes, with a recurrence interval of ~90-150 years (last major event 1946). Japanese government assessments in recent years have given 70-80% probability of a Nankai earthquake within 30 years, implying roughly 10-15%+ per 3.65-year window from Nankai alone. Other sources (Kuril, Japan Trench) add additional probability. **What makes this different from superficial analogs**: The resolution criterion (epicenter in Japan OR territorial waters) is broad, including the expansive Exclusive Economic Zone and surrounding seas, which increases the effective area. **Temporal analysis**: With 1,333 days (~3.65 years) remaining, the question asks whether a low-probability-per-day but recurring event occurs over a multi-year window. Each year that passes without an M8+ event slightly adjusts the remaining probability, but the Poisson rate from historical data suggests meaningful cumulative probability (likely 30-55%). Current market prices and recent Japanese government warnings about Nankai Trough risk are critical calibration inputs.

2. Research Parallel tool calls · legacy v2

Tool	Status	Time	Items
wikipedia_lookup	OK	1.8s	-
article_search	OK	0.6s	-
web_search	OK	50.5s	-
kalshi_data	OK	1.8s	-
code_execution	OK	0.0s	-

3. Evidence Synthesis Sonnet · 12 items · legacy v2

#	Strength	Credibility	Direction	Source	Claim	Priced In
1	MODERATE	65	↓ DOWN	code_execution	A Poisson model using 11 M8.0+ events in Japan over 125 years (rate ~0.088/yr) yields a 27.5% probability of at least one M8.0+ event in a 3.65-year window; using a higher count of 15 events raises this to ~35.5%.	Yes
2	WEAK	50	NEUTRAL	wikipedia_lookup	Wikipedia lists of Japanese earthquakes confirm Japan is highly seismically active with frequent M7.0+ events, but exact count of historical M8.0+ events in Japan's territory is not precisely enumerated in the retrieved data.	Yes
3	STRONG	88	NEUTRAL	web_search	No M8.0+ earthquake has occurred near Japan in the past 10 years (as of April 2026), with the strongest recent event being the December 2025 M7.5-7.6 Aomori quake — meaning at least 10 years have passed since Japan's last M8.0+ event, extending a notable quiet period.	Yes
4	MODERATE	82	NEUTRAL	article_search	Following the December 2025 M7.5 Aomori earthquake, Japan's government issued a megaquake advisory estimating ~1% probability of an M8.0+ quake in the subsequent week — elevated but still very low short-term risk, and no megaquake occurred.	Yes
5	MODERATE	85	↑ UP	web_search	The April 2026 Sanriku earthquake (M7.4-7.7) prompted another JMA megaquake advisory, with officials stating the chance of an M8.0+ quake was increased tenfold to ~1% from a background of 0.1% — indicating elevated but still modest near-term risk.	No
6	MODERATE	72	↑ UP	article_search	Japan experienced a swarm of 1,800+ earthquakes in three weeks at the Tokara Islands in July 2025, reflecting heightened seismic activity in parts of Japan's territory.	Yes
7	STRONG	87	↑ UP	article_search	Japan's earthquake investigation panel (as of September 2025) assessed a 60-90% probability of a Nankai Trough megaquake within 30 years, reaffirming sustained elevated risk from this fault system.	Yes
8	STRONG	87	↑ UP	article_search	Japan's government assessment gives 60-90% probability of a Nankai Trough megaquake (M8-9 class) within 30 years, implying roughly 7-15% probability in any given 3.65-year window from Nankai alone.	Yes
9	STRONG	80	↑ UP	wikipedia_lookup	The Nankai megathrust has a documented history of recurring M8+ earthquakes, with the last major event in 1946; the fault is divided into multiple segments that can rupture independently or together.	Yes
10	MODERATE	84	↑ UP	web_search	Two M7.4-7.7 earthquakes at the Japan Trench (Dec 2025, Apr 2026) have not triggered M8.0+ follow-on events, but each prompted official megaquake advisories, showing active stress in the Japan Trench subduction zone.	No
11	MODERATE	90	NEUTRAL	article_search	A major M8.8 earthquake struck off Kamchatka, Russia in July 2025, causing tsunamis reaching Japan — this occurred in the broader Pacific Ring of Fire but epicenter was in Russian territory, not Japan, so would NOT resolve the question YES.	Yes
12	MODERATE	60	↑ UP	kalshi_data	The Kalshi market currently prices this event at 58%, up sharply +11% in the last 7 days and +4% over 30 days, with the price range over 152 days being 36-74% — suggesting market participants price significantly higher than the raw Poisson base rate of ~27-35%.	Yes

Information Gaps

Exact catalog count of M8.0+ earthquakes with epicenters specifically within Japan's territorial waters (not just broadly 'near Japan') since 1900 — critical for calibrating the Poisson rate
Whether the April 2026 Sanriku advisory period (until April 27) has now passed without incident, or whether any subsequent seismic activity has occurred
Current GPS-based strain accumulation data for the Nankai Trough and Japan Trench, which would quantify actual stress loading relative to historical averages
Precise remaining time window until January 1, 2030 (needed to recalculate Poisson probability with updated elapsed time)
Expert seismological assessment of whether recent M7.4-7.7 earthquakes at the Japan Trench have relieved or transferred stress in ways that alter M8.0+ risk
Any slow-slip events or tremor activity along the Nankai Trough recently, which can precede megathrust earthquakes
Why the Kalshi market jumped +11% in 7 days — whether driven by the April 2026 Sanriku earthquake and megaquake advisory or other factors

Key Uncertainties

The true historical rate of M8.0+ earthquakes in Japan's territory is uncertain depending on exact geographic boundary and catalog completeness — the Poisson estimate of 27-35% vs. market price of 58% represents a significant gap requiring explanation
Whether the recent M7.5-7.7 earthquakes at the Japan Trench (Dec 2025, Apr 2026) represent increased stress loading that raises near-term M8.0+ risk, or whether they partially relieved stress
The Nankai Trough's 60-90%/30-year probability implies wide uncertainty bounds — it could rupture before 2030 or not for decades more
The resolution criterion specifies 'epicenter in Japan or its territorial waters' — the breadth of Japan's EEZ significantly expands the effective seismic area vs. just mainland Japan
Whether the Kalshi market's 58% price reflects better-calibrated expert knowledge, emotional/media-driven overreaction to recent M7.5 events, or genuine incorporation of multiple fault zone risks

Market Context

Price: 0.58 · 7d: +0.11 · 30d: +0.04
Volume: MODERATE · Depth: MEDIUM
Reliability: MEDIUM — 92 avg daily contracts over 152 days with meaningful price history (36-74% range), but the 11% jump in 7 days suggests possible reaction to April 2026 Sanriku earthquake and megaquake advisory rather than purely fundamental reassessment. Market price of 58% substantially exceeds Poisson base rate of 27-35%, implying either broader geographic interpretation, higher historical rate assumptions, or elevated near-term risk weighting.

4. Independent Forecaster Opus · No market price · legacy v2 Did NOT see market price

Independent estimate: 41% (45% confidence)

base rate

42%

45% conf

evidence driven

40%

45% conf

contrarian

40%

45% conf

Spread

2pp

Low disagreement

sq1: Based on historical frequency, is there at least a 50% probability that Japan experiences an M8.0+ earthquake in any given 4-year period?

40%

Base rate from Poisson: 11 events/125 years = 0.088/yr gives ~30% for 4 years; 15 events gives ~38%. The question asks if it's ≥50%, which requires a rate of ~0.17/yr (one M8+ every ~6 years). Looking at post-1900 Japan M8+ events (1933 Sanriku, 1944 Tonankai, 1946 Nankai, 1952 Tokachi, 1958 Etorofu, 1963 Kuril, 1968 Tokachi-oki, 1994 Hokkaido, 2003 Tokachi-oki, 2011 Tohoku, plus a few near-Kuril boundary), counts vary 10-15. With broad EEZ inclusion, post-2000 rate appears higher (~0.12/yr), giving ~38% over 4 years. The 50% threshold requires being somewhat above what straight historical data shows. Estimate 40%.

sq2: Are current seismic indicators elevated relative to baseline?

70%

Multiple converging indicators suggest elevated state: (1) JMA issued two megaquake advisories in 5 months (Dec 2025, Apr 2026), explicitly stating 10x baseline elevation; (2) Tokara swarm of 1,800+ events in July 2025; (3) Nankai 30-yr probability at 60-90%, near upper bound of historical assessments; (4) Kamchatka M8.8 in July 2025 may have stressed adjacent zones. However, 'elevated' is somewhat subjective and Japan is always seismically active. Strong evidence points to YES on this descriptive question. Estimate 70%.

sq3: Will Nankai Trough or other identified high-risk fault zones produce M8.0+ before 2030?

30%

Nankai alone: 60-90%/30yr → ~10-14% per 3.65yr window (using -ln(1-0.75)/30 ≈ 0.046/yr → ~15% over 3.65yr). Japan Trench: post-2011 stress redistribution, recent M7.5+ events suggest active loading, perhaps 5-8%. Kuril/Hokkaido zone: ~5%. Sagami, Ryukyu: ~3-5%. Combined independent probabilities (1-product): ~25-30%. Recent megaquake advisories with elevated short-term probabilities provide modest upward push. Estimate 30%.

Show full reasoning trace

base rate: Poisson model with λ≈0.088-0.12/yr for M8+ in Japan territory yields 30-38% probability over 3.65 years remaining. Post-2000 rate appears slightly elevated (Tokachi-oki 2003, Tohoku 2011) suggesting upper end of range is appropriate.

evidence updates: UP: Nankai Trough 60-90%/30yr assessment (+5-8% from base); recent M7.4-7.7 events at Japan Trench suggesting active stress (+2-3%); JMA megaquake advisories (+1-2%, mostly priced into structural assessments). DOWN: 10-year quiet period since last M8+ doesn't strongly affect Poisson assumption but very slightly suggests possible stress relief in some segments. Net upward adjustment ~8-10% from naive base rate.

combination method: Weighted average across three sub-questions, with sq3 (direct probability question) getting effective weight boost since it most directly maps to resolution criterion.

final: Combining ~33% Poisson base with ~10% upward adjustment for elevated indicators (Nankai assessments, recent advisories, broad EEZ definition) yields ~42%. Confidence is moderate (~0.45) given catalog ambiguity and inherent randomness of earthquake timing.

ensemble: {'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.42, 'confidence': 0.45}, 'evidence_driven': {'p_yes': 0.4, 'confidence': 0.45}, 'contrarian': {'p_yes': 0.4, 'confidence': 0.45}}, 'spread': 0.02, 'individual_p_yes': {'base_rate': 0.42, 'evidence_driven': 0.4, 'contrarian': 0.4}}

5. Devil's Advocate Sonnet · legacy v2 Market price revealed

Reasoning Flaw Missing Info

Challenges

Time window math may be inconsistent. Today is May 8, 2026; resolution is Jan 1, 2030 — that's ~3.65 years. But the forecaster sometimes uses 4-year windows in sq1 ('30% for 4 years', '38% over 4 years'). Using the correct 3.65-year window with λ=0.088-0.12/yr gives ~28-35%, slightly lower than what was used. Minor but worth checking.
Sub-question combination is opaque and possibly inconsistent. sq3 is described as 'most directly maps to resolution criterion' at 30%, but the final answer is 40.7% — higher than sq3. If sq3 is the most direct estimate, why does the final exceed it? The reasoning trace says 'weighted average' but doesn't show weights, and sq2 (descriptive question about elevated indicators) shouldn't directly increase the probability estimate — it's already captured in sq3's evidence updates. Risk of double-counting elevated indicators.
The forecaster may be underweighting the 60-90% Nankai 30-year assessment. Even taking the lower bound of 60%/30yr → λ=0.0305/yr → ~10.5% over 3.65 years for Nankai ALONE. The upper bound 90%/30yr → λ=0.077/yr → ~24.5% for Nankai alone. sq3's combined ~25-30% across ALL zones may be too low if Nankai upper-bound is realistic, especially given JMA's recent advisories explicitly citing 10x baseline elevation.
The 'broad EEZ inclusion' factor is mentioned but not quantitatively integrated. If resolution criterion includes broader Japan-region earthquakes (e.g., Kuril boundary, southern Ryukyu), the historical rate is materially higher than the narrow 11/125yr estimate. The forecaster acknowledges this could push the rate to 0.12/yr but doesn't fully commit to it.
The information gap about 'precise remaining time window' suggests the forecaster didn't lock in the exact 3.65-year figure, which propagates uncertainty into the base rate calculation.

Suggested adjustment: +5pp

Modest upward adjustment warranted due to (1) sq3 likely underestimates given Nankai upper-bound assessments and JMA's explicit 10x elevation statement, (2) the final 40.7% being higher than sq3's 30% suggests the combination logic adds something — but if that something is elevated indicators, those should push higher than 40.7%, and (3) broad geographic interpretation likely raises the base rate. This is NOT about market conformity — even ignoring the 58% market, the internal reasoning suggests the answer should be closer to 45-48%.

Market comparison: Forecaster: 40.7%, Market: 58%. The 17-point divergence is partially explained by the market reacting to the April 2026 Sanriku earthquake and megaquake advisory (+11% in 7 days). The forecaster has incorporated these factors but may underweight them. Some of the market premium likely reflects either (a) broader resolution interpretation, (b) traders pricing JMA's 10x elevation more aggressively, or (c) overreaction to the advisory. A reasonable forecast probably sits in the 42-50% range; full conformity to 58% is not warranted without stronger evidence the advisory-period risk is materially elevated.

6. Aggregator Opus · final calibrated forecast

48% (40% confidence)