weather

kalshiMarch 9, 202617d ago

California 8.0+ magnitude earthquake before 2035

Will there be an at least 8.0 magnitude earthquake in California before 2035?

Signal

SELL

Probability

Market: 41%Edge: -38pp

Confidence

HIGH

85%

Summary.

The market is pricing a California 8.0+ magnitude earthquake at 40.5% probability through December 2035, representing a massive 17.6x overpricing relative to scientific consensus. The USGS UCERF3 model—the gold standard for California earthquake forecasting—estimates just 7.0% probability over 30 years, which adjusts to approximately 2.3% for the remaining 9.8-year timeframe of this bet. My estimated probability is 2.5%, accounting for modest tail risk from the Cascadia Subduction Zone and model uncertainty. The market appears driven by recency bias from yesterday's routine M3.6 Santa Rosa earthquake and February's San Ramon swarm activity, evidenced by the 2.5 percentage point price increase on only 12 contracts of volume. Critically, no M8.0+ earthquake has occurred in California in modern seismological history, the San Andreas Fault's physical maximum is approximately 8.0 (requiring an unprecedented full-margin rupture), and recent seismic activity falls well within normal patterns that do not predict major events. The wide 35¢ bid-ask spread and low liquidity indicate market inefficiency, creating a strong opportunity to bet NO at significant value. The edge is substantial (38 percentage points) and backed by rigorous geological science, though capital will be tied up for nearly a decade.

Reasoning.

Step 1: Establish Climatological Base Rate

The UCERF3 (Uniform California Earthquake Rupture Forecast, version 3) is the gold standard seismological model for California earthquake risk, developed by USGS and partners. It estimates a 7.0% probability of a magnitude 8.0+ earthquake occurring anywhere in California over a 30-year period.

Step 2: Adjust for Actual Timeframe

The bet covers March 2026 to December 2035, approximately 9.8 years. Assuming a roughly uniform distribution (reasonable for rare seismic events), the adjusted probability is:

7.0% × (9.8 / 30) ≈ 2.3%

Step 3: Evaluate Physical Constraints

Critical context from seismological research:

The San Andreas Fault has a maximum magnitude capability of ~8.0, requiring simultaneous wall-to-wall rupture of both Northern and Southern segments (an extremely rare scenario)
Strike-slip faults like the San Andreas cannot produce 9.0+ megathrust earthquakes
The Cascadia Subduction Zone (beginning at Mendocino Triple Junction in Northern California) can theoretically produce 9.0+ events, but 10-year probability remains in low single digits
No magnitude 8.0+ earthquake has been recorded in California in modern seismological history

Step 4: Assess Recent Seismic Activity

Recent events that may be influencing market pricing:

March 8, 2026: M3.6 Santa Rosa, M2.7 San Ramon (yesterday's events)
February 2026: ~100 earthquake swarm in San Ramon Valley over 30 days
December 5, 2024: M7.0 Humboldt County earthquake

Important: These events are normal California seismic activity and do not indicate elevated risk of M8.0+ events. Small earthquakes and swarms do not reliably predict large earthquakes. The M7.0 Humboldt event, while significant, is still a full magnitude below 8.0 (10x less energy release).

Step 5: Market Psychology vs. Science

The market price of 40.5% represents a massive disconnect from scientific consensus:

Market implies: 40.5% probability
UCERF3 base rate: 2.3% probability
Mispricing: Market is 17.6x higher than scientific forecast

The 2.5 percentage point move in the last 24 hours (36% → 40.5%) on only 12 contracts of volume strongly suggests recency bias from yesterday's M3.6 earthquake. Low liquidity (23¢ bid / 58¢ ask spread) indicates the market is inefficient and susceptible to emotional trading.

Step 6: Uncertainty Adjustments

Reasons to adjust slightly upward from 2.3%:

Cascadia Subduction Zone potential (though epicenter location in California proper is uncertain)
Model uncertainty in tail risk events
Conservative rounding

Reasons confidence remains high:

UCERF3 is extremely well-validated and represents expert consensus
Physical constraints on fault mechanics are well-understood
No precursor activity suggesting elevated risk
Long timeframe (9.8 years) doesn't add much uncertainty to geological base rates

Final Estimate: 2.5%

This represents a modest upward adjustment from the 2.3% UCERF3 estimate to account for:

Cascadia uncertainty
Model tail risk
Conservative rounding given the catastrophic nature of the event

The true probability remains in the 2-3% range, making the current market price of 40.5% extraordinarily mispriced.

Key Factors.

UCERF3 30-year base rate of 7.0% adjusted to 2.3% for 9.8-year timeframe
Physical constraints: San Andreas Fault maximum magnitude ~8.0 requiring full-margin rupture (extremely rare)
No M8.0+ earthquake recorded in California in modern seismological history
Recent seismic activity (M3.6 Santa Rosa, swarms) is routine and not predictive of major events
Market showing extreme recency bias and psychological overreaction to normal seismicity
Low market liquidity (12 contract volume, 35¢ spread) enables mispricing
Cascadia Subduction Zone represents most plausible M8.0+ scenario but has low single-digit 10-year probability

Scenarios.

Base Case - Normal Seismic Activity

98%

California experiences normal seismic activity over the next 9.8 years, including moderate earthquakes (M5-7 range), swarms, and aftershock sequences. No M8.0+ event occurs. The San Andreas Fault does not experience a full-margin rupture, and the Cascadia Subduction Zone does not produce a major event with epicenter in California waters.

Trigger: Continuation of current patterns: scattered M3-5 earthquakes, occasional M6-7 events (consistent with historical rates), localized swarms. UCERF3 forecast accuracy continues to hold.

Cascadia Subduction Zone Event

A major Cascadia Subduction Zone megathrust earthquake (M8.0-9.0+) occurs with epicenter located in California territorial waters near the Mendocino Triple Junction. This represents the most plausible scenario for an M8.0+ event in the specified geography, though still highly unlikely over the timeframe.

Trigger: Increased seismic activity near Mendocino Triple Junction, slow-slip events detected on Cascadia interface, pattern of precursor earthquakes. However, note that earthquake prediction remains unreliable.

San Andreas Full-Margin Rupture

An extremely rare simultaneous rupture of both Northern and Southern segments of the San Andreas Fault occurs, producing a magnitude ~8.0 earthquake. This is at the absolute physical limit of what strike-slip faults can produce and would require unprecedented coordination of rupture across 800+ miles of fault.

Trigger: Increased seismicity along multiple SAF segments, geodetic strain measurements showing locked sections approaching failure threshold, pattern of M6+ events on adjacent segments. However, such precursors are not reliable predictors.

Risks.

UCERF3 model may underestimate tail risk of extreme events due to limited historical data
Cascadia Subduction Zone timing is highly uncertain - could occur anytime from tomorrow to 300+ years
Earthquake science cannot predict specific events - a M8.0+ could occur without precursors
Resolution criteria includes California territorial waters, which extends epicenter possibilities
Induced seismicity or previously unknown faults could theoretically produce larger-than-expected events
Recent seismic activity could (though unlikely) represent genuine precursors we don't yet understand
Climate change impacts on crustal stress are not well-modeled
My analysis assumes market is irrational, but informed traders may have access to non-public geophysical data

Edge Assessment.

STRONG EDGE: Market significantly overpriced at 40.5% vs. estimated 2.5%

Quantitative Edge:

Estimated probability: 2.5%
Market implied probability: 40.5%
Raw edge: 38 percentage points of overpricing
Expected value of NO position: +37.5 cents per dollar

Edge Sustainability: The market moved UP 2.5 points in the last 24 hours on tiny volume (12 contracts) following yesterday's routine M3.6 earthquake. This is classic recency bias and suggests the market is driven by emotional reactions to news rather than seismological fundamentals.

Key Edge Factors:

Massive scientific consensus gap: UCERF3 at 2.3% vs market at 40.5% is a 17.6x discrepancy
Low liquidity inefficiency: 35¢ bid-ask spread and 12-contract volume means market is easily manipulated by uninformed flow
Recency bias: Recent swarms and yesterday's quakes are causing predictable overreaction despite being routine California seismicity
Long timeframe works in NO favor: 9.8 years feels like "a lot can happen" psychologically, but geologically it's a blink

Risks to Edge:

Market could stay irrational longer than expected (resolution not until 2036)
Additional moderate earthquakes (M5-7) could push price even higher temporarily
Opportunity cost of capital tied up for years
"Big One" fear is deeply embedded in California psychology and hard to arbitrage

Trading Recommendation: This is an exceptionally strong NO bet, but be prepared for price to move against you in the short term if California experiences normal M5-7 earthquakes. The fundamental edge is enormous and backed by the best available seismological science. Consider sizing appropriately for a long-term hold (9.8 years) and expect volatility around earthquake news cycles.

Caveat: If you observe large informed volume coming in on YES side, reassess whether there's non-public geophysical intelligence you're missing. Current movement appears purely emotional.

What Would Change Our Mind.

A credible update to UCERF3 or new USGS/SCEC analysis significantly raising the probability of M8.0+ events beyond current 7% 30-year estimate
Sustained anomalous seismic activity patterns (e.g., weeks of M5+ earthquakes along San Andreas, unusual deep tremor at Cascadia) that credible seismologists identify as potential precursors
Detection of significant slow-slip events or aseismic creep acceleration on major fault segments indicating imminent rupture risk
Publication of peer-reviewed research identifying previously unknown megathrust potential in California geology
Large informed volume (thousands of contracts) entering YES side from sophisticated traders, suggesting non-public geophysical intelligence
Major M7.5+ earthquake occurring in California that changes scientific assessment of fault connectivity and rupture potential
Evidence that market mispricing is correcting downward toward fundamentals, reducing edge below acceptable threshold

Sources.

Market History.

Market moved up 2.5 percentage points in the last 24 hours (from 36¢ to 39¢). 7-day range: 27¢ – 40¢.

Get This Via API.

Access real-time prediction market analysis programmatically. Every analysis on this page is available through our REST API.

curl -X POST https://api.rekko.ai/v1/analyze \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"category": "weather", "platform": "kalshi"}'

Get API Key Read Docs

Related Analysis.

weatherkalshi

SELL

Will a supervolcano erupt before 2050?

The market prices a supervolcano eruption before 2050 at 17.5%, but my analysis estimates the true probability at only 0.15% — a staggering 117x overpricing. This massive discrepancy stems from critical definitional ambiguity and apparent market confusion. Geological base rates show VEI 8 supereruptions occur globally once every 50,000-100,000 years, yielding just 0.02-0.05% probability over the 24-year timeframe. Current monitoring data from USGS and INGV as of March 2026 shows no precursor signals at any major caldera: Yellowstone operates at background levels with insufficient molten material (6-15%), and even elevated Campi Flegrei activity (Yellow alert, peaked 2023-2025) suggests at most a VEI 3-4 eruption, not a civilization-altering VEI 8. The market appears anchored to an ice core study showing 16.7% probability of a VEI 7 eruption by 2100 — conflating a smaller magnitude event (10x less than VEI 8) with a different timeframe. However, the resolution criteria stating "if a supervolcano erupts" without specifying VEI threshold creates substantial risk: if any eruption at a supervolcano site qualifies, the probability could increase 10-20x to 1-3%. This definitional uncertainty, reflected in my low 0.45 confidence level, is the primary risk factor preventing maximum conviction despite apparent massive mispricing on geological merits alone.

0%Mar 8, 2026

weatherkalshi

SELL

Will a supervolcano erupt before 2050?

The market is pricing a supervolcano eruption at 23.5% probability through 2050, representing a catastrophic ~587x overvaluation relative to geological reality. Based on well-established VEI 8 eruption base rates (once per 50,000-100,000 years globally) and current monitoring data showing zero elevated risk signals, the true probability is approximately 0.04% (1 in 2,500) over the 24-year period. As of March 2026, Yellowstone remains at NORMAL alert level with a "nearly solid" magma chamber, paused ground uplift, and background seismic activity. Recent geyser eruptions are routine geothermal processes being sensationalized by media, driving uninformed speculation (volume spiked 2.2x with 4-point price movement in 24 hours). VEI 8 eruptions cannot occur without years-to-decades of detectable precursors (massive ground deformation, thousands of earthquakes, gas emission spikes)—none of which are present. The 23.46 percentage point edge represents extreme market inefficiency driven by sensationalist headlines, geological innumeracy, and the 24-year capital lockup deterring informed traders.

0%Mar 9, 2026

weatherkalshi

SELL

Wind as largest source of global primary energy consumption in 2030

The market prices wind becoming the largest global primary energy source in 2030 at 4%, but the estimated true probability is approximately 0.01% (1 in 10,000), representing a ~400x overpricing. This extreme mispricing reflects confusion between electricity generation metrics (where wind is growing rapidly) and total primary energy consumption (where wind represents only 3-4% currently vs oil at 31%). For wind to resolve YES, it would need to grow 10-fold in just 4 years while simultaneously displacing oil, coal, and natural gas—all of which are projected by IEA, BP, and other authoritative agencies to remain dominant through 2030, collectively maintaining ~73% of the global energy mix. No credible forecasting scenario supports wind overtaking fossil fuels by 2030; the mathematical and infrastructure constraints make this outcome virtually impossible within the 4-year timeframe. The market price likely reflects retail confusion, longshot bias, and minimum tick size effects rather than informed probability assessment.

0%Mar 24, 2026