Scan: Magnitude 6.5+ earthquake in LA before 2027?

date	predicted	market price	confidence
2026-08-04	3%	8%	65%

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

What is the current Polymarket YES price and how has it trended, and how much time remains until Dec 31, 2026?
How many M6.5+ earthquakes have had epicenters within 50 miles of 34.0522N, 118.2437W since 1900 (e.g., 1971 San Fernando, 1994 Northridge; does 1933 Long Beach M6.4 or 1952 Kern County fall outside), and what annual rate does that imply?
What does the USGS UCERF3 model give as the 30-year probability of an M6.7+ (or M6.5+) earthquake in the Los Angeles region, and what per-year rate does that imply?
Has any M6.5+ earthquake already occurred within 50 miles of LA since the market's creation, per USGS?
Is there any recent elevated seismic activity, swarm, or foreshock sequence in the LA basin / San Andreas / Puente Hills / Newport-Inglewood faults that would raise short-term probability above baseline?
Do related Kalshi or Polymarket earthquake markets (California, San Francisco, global M7+) imply annual rates consistent with or divergent from this market's price?

Planner reasoning

This is fundamentally a base-rate question: historically only two M6.5+ quakes (1971 San Fernando M6.6, 1994 Northridge M6.7) have struck within 50 miles of downtown LA in the past ~100+ years, implying roughly a 2-3% annual Poisson rate, though UCERF3 hazard models may imply a somewhat different rate for the region. The key work is anchoring on the Polymarket price, checking whether any qualifying event has already occurred in the window, converting published USGS/UCERF3 probabilities to the remaining window, and checking for longshot bias versus the actuarial base rate.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.6s	1	## This Market's Polymarket Data Magnitude 6.5+ earthquake in LA before 2027? - Current price (probability): 7.20% - 7-day price change: -5.00% - 30-day price change: -3.95% - Total volume: $18,136 (USD notional) - Price range: 2.40% - 37.00% - Data points: 140 days
polymarket_related	OK	24.5s	0	Scanned 100 active Polymarket markets, kept 0 matches. keyword 'earthquake': 0 markets \| keyword 'Los Angeles earthquake': 0 markets \| keyword 'California earthquake': 0 markets
kalshi_related	OK	24.3s	3	3 related markets / summaries. keyword 'earthquake': ok \| keyword 'California earthquake': ok \| keyword 'magnitude': ok
claude_news	OK	21.7s	7	Current date for this research: September 5, 2026. Summary of findings: - No M6.5+ earthquake has occurred within 50 miles of downtown LA in 2025–2026. A live-tracking aggregator (data sourced from USGS) explicitly states "No significant earthquakes recorded near Los Angeles recently" as of l
gdelt_news	OK	226.3s	0	GDELT: 0 articles across 3 queries (lookback=60d). 'Los Angeles earthquake magnitude': error GDELT rate-limited after retries (429) \| 'Southern California earthquake USGS': error GDELT rate-limited after retries (429) \| 'San Andreas earthquake': error GDELT rate-limited after retries (429)
wikipedia	OK	0.2s	5	Fetched 5 Wikipedia entries (0 missing pages).
code_execution	OK	26.4s	0	## Key Findings Historical qualifying events (within 50 mi of downtown LA, M6.5+, haversine-verified): - Only 2 events qualify: 1971 San Fernando (M6.6, 26.2 mi) and 1994 Northridge (M6.7, 20.1 mi) - All other listed events fall outside 50 mi (Long Beach 33.1 mi but M6.4 <6.5; Kern County 7

3. Evidence Brief Sonnet 5 · 6753 chars

# Current state As of the most recent research (dated Sept 5, 2026), no M6.5+ earthquake has occurred within 50 miles of downtown Los Angeles since market creation; the market resolves YES only if such an event occurs before Dec 31, 2026, 11:59 PM ET (~117 days / ~3.8 months remaining). Recent Southern California seismicity (Brawley/Salton Sea and Coachella swarms) has been concentrated well outside the 50-mile LA radius and is not indicative of imminent LA-basin rupture. # Timeline of key events - 1933-03-10: M6.4 Long Beach earthquake — outside qualifying magnitude/possibly distance for this market (confirmed, historical). - 1952-07-21: M7.5 Kern County earthquake, ~130 km (~81 mi) from LA — outside 50-mile radius (confirmed, historical). - 1971-02-09: M6.6 San Fernando/Sylmar earthquake, ~26 mi from LA — qualifies (confirmed, historical). - 1994-01-17: M6.7 Northridge earthquake, ~20 mi from LA — qualifies, largest recorded LA-area event (confirmed, historical). - 2026-01 (Jan 2026): M4.9 earthquake swarm near Indio, Coachella Valley (250+ aftershocks) — outside LA radius (confirmed via news). - 2026-05-11: Swarm of 350+ earthquakes up to M4.7 near Brawley/Salton Sea — outside LA radius, recognized recurring pattern (confirmed via news). - 2026-08 (late Aug 2026): No M6.5+ or significant seismic activity reported near LA; ~18 M4.0+/yr baseline, most recent notable event M3.0 near Corona (confirmed via tracker). # Event Will a magnitude 6.5+ earthquake occur within 50 miles of Los Angeles before Dec 31, 2026 (per USGS)? # Outcomes to forecast - Yes - No # Kalshi market anchor No direct Kalshi YES price was returned by the kalshi_direct tool in this research pass (gap — see below). Using Polymarket as the best available cross-platform proxy: current YES price **7.2%**, down from a 30-day range high of 37% (7-day: -5.0pp, 30-day: -3.95pp), suggesting a market that priced in elevated risk earlier (likely near creation) and has since decayed toward baseline as no event occurred and time elapsed. Volume is thin (~$18K total, 140 data points), indicating low liquidity/conviction. This 7.2% appears materially higher than base-rate models justify for the remaining ~4-month window (see below), implying downside risk to the YES price if it doesn't reflect true short-window hazard. # Sub-question answers 1. **Polymarket price/trend/time remaining** — Current 7.2%, down 5pp (7d) and ~4pp (30d) from a high of 37%; ~117 days remain until close (Dec 31, 2026), per polymarket_direct and claude_news dating. 2. **Historical M6.5+ within 50mi since 1900** — Only 2 qualifying events: 1971 San Fernando (M6.6, 26 mi) and 1994 Northridge (M6.7, 20 mi); 1933 Long Beach (M6.4) and 1952 Kern County (~81 mi) don't qualify (code_execution, Wikipedia). Implied rate: λ≈2/125yr ≈0.016/yr. 3. **UCERF3 30-yr probability** — ~60% chance of M6.7+ within 100km of LA in 30 years (implied λ≈0.0305/yr); ~46% for M7.0+ (claude_news/UCERF3). Scaling down from 100km region to 50-mile circle yields a comparable ~0.015–0.021/yr effective rate. 4. **Has M6.5+ already occurred since market creation?** — No; confirmed no qualifying event through late Aug 2026 (claude_news, earthquaketracker.org). 5. **Elevated seismicity/swarms near LA faults?** — No advisories on Newport-Inglewood, Puente Hills, Sierra Madre, or San Andreas-near-LA; all 2025-2026 swarms (Brawley, Coachella/Indio) are in Imperial Valley, geographically distinct and not a recognized LA precursor (claude_news). 6. **Related Kalshi markets implied rates** — CA M8.0+ by 2028: 10%; CA M8.0+ by 2035: 33%; Japan M8.0+ by 2030: 37%. These are for larger magnitudes/broader regions/longer horizons and not directly comparable, but the 2028 CA-M8.0 price (10% over ~2 more years) is roughly consistent with a low-single-digit-percent annualized severe-quake rate, supportive of this market's low base-rate expectation (kalshi_related). # Key facts (high-confidence, factual) 1. [code_execution/Wikipedia] Only 2 historical M6.5+ events within 50 mi of downtown LA since 1900: 1971 San Fernando, 1994 Northridge. 2. [claude_news] No M6.5+ event within 50 mi of LA has occurred in 2025-2026 to date. 3. [claude_news] UCERF3: ~60% chance M6.7+ within 100km of LA in 30 years. 4. [polymarket_direct] Polymarket YES = 7.2%, down from 37% high; low volume (~$18K). 5. [kalshi_related] CA M8.0+ by 2028 priced at 10%; by 2035 at 33%. # Cross-market signals - Kalshi related: CA-wide M8.0+ markets (10%/2028, 33%/2035) imply low annualized severe-quake odds, broadly consistent with a low-probability outcome here despite different magnitude/region thresholds. - Polymarket: Same-event proxy at 7.2%, declining, thin volume — likely overpriced relative to base-rate models for remaining window. - Sportsbook implied: N/A (no sportsbook coverage of earthquake events). # Analyst opinions and speculation - code_execution triangulation: base-rate (Poisson, historical) and UCERF3-scaled models converge; both were computed assuming ~15 months remaining (likely miscalculated from Sept 2025 baseline) — should be recalculated for the true ~4-month remaining window, yielding a much lower probability (~0.5-1%) than the reported 12-18 month figures (~1.5-3.2%). - No credible source suggests imminent LA-basin rupture; all 2026 swarms are in a distinct, unrelated fault system (Brawley Seismic Zone). # Directional lean per outcome - **Yes**: Supported only by long-run UCERF3 hazard (~60%/30yr) and historical precedent (2 events/125 yrs); no supporting evidence of imminent activity; short remaining window (~4 months) sharply limits probability. - **No**: Strongly favored — no qualifying event to date, no elevated LA-fault activity, short remaining window, base-rate models (recalculated for ~4 months) imply ~0.5-1% Yes probability, far below Polymarket's 7.2%. # Gaps / unknowns - No direct Kalshi YES price was retrieved for this specific ticker (tool did not return it) — must rely on Polymarket proxy. - Exact market creation date/duration on Kalshi unclear, complicating precise time-remaining calibration. - code_execution's headline probability figures (1.5-3.2%) appear to use an incorrect ~15-month window rather than the actual ~4-month remaining window; true point estimate likely lower (~0.5-1%). # Calibration anchors - Polymarket YES (proxy anchor): 7.2%, trending down. - Historical base rate: 2 qualifying events in ~125 years (λ≈0.016/yr) → ~0.5-0.7% over remaining ~4 months. - UCERF3-scaled: ~0.02-0.03/yr region-wide, scaled to 50-mile circle → ~0.5-1% over remaining ~4 months. - Reasonable calibrated estimate: **~1-3%** Yes (below current 7.2% cross-market proxy), reflecting recalculated short remaining window and absence of precursor activity.

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

78% conf

Yes	3%
No	97%

Only two M6.5+ events within 50 miles of LA have occurred since 1900 (1971 San Fernando, 1994 Northridge), implying λ≈0.016/yr, and UCERF3 hazard scaled to the 50-mile circle gives a similar ~0.02/yr rate; over the remaining ~117 days this yields roughly 0.5–1% Yes. No qualifying event has occurred to date and all 2026 swarms (Brawley, Coachella) are in the geographically distinct Imperial Valley with no LA-fault advisories. No direct Kalshi price was retrieved; the Polymarket proxy at 7.2% (thin volume, trending down from 37%) is typical longshot overpricing in low-liquidity tail markets. I shade slightly above the pure base rate to ~3% to account for model uncertainty and fat-tailed seismic clustering, but the evidence overwhelmingly favors No.

gpt-5.6-sol

76% conf

Yes	2%
No	98%

The available market anchor is the thinly traded Polymarket proxy at 7.2% YES; no direct Kalshi price was retrieved. Historical data show only two qualifying earthquakes since 1900, implying roughly a 0.5% chance over the remaining 117 days, while UCERF3-scaled estimates point toward approximately 0.5–1%. No qualifying event has occurred yet, and recent Southern California swarms were outside the 50-mile radius with no recognized connection to imminent LA-fault rupture. I retain a modest tail above the base-rate estimates for model and geographic uncertainty but forecast well below the proxy price.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasts converge almost exactly (3.0% and 2.5%) using the same Poisson/UCERF3 base-rate calculation, suggesting herding rather than independent triangulation; neither explores why the Polymarket proxy might be more informative than a naive base rate (e.g., prior swarms triggering Coulomb stress transfer onto LA-adjacent faults) rather than simply dismissing it as "thin-liquidity overpricing."

Neither forecaster flags the missing direct Kalshi YES price as a major calibration risk — the brief explicitly notes "No direct Kalshi YES price was returned... gap" — yet both proceed as if the Polymarket 7.2% is a reliable enough proxy to anchor against, without discounting for cross-platform basis risk (different resolution criteria, liquidity, or user base).

The historical count of "2 qualifying events in 125 years" likely undercounts pre-1932 seismicity due to incomplete instrumental catalogs (the network only became robust after ~1932), meaning true λ could be higher than 0.016/yr; neither forecast adjusts for catalog-completeness bias in the base rate.

Neither discusses tail/clustering risk from the ongoing 2026 Brawley/Coachella swarms (350+ events, up to M4.7) — while outside the 50-mile radius, historical Southern California seismicity research shows remote triggering of San Andreas-adjacent faults is a documented (if rare) phenomenon, which both dismiss too quickly as "geographically distinct" without quantifying residual risk.

Both forecasts use nearly identical confidence levels (0.76-0.78) despite acknowledging significant model uncertainty (the brief's own analyst flagged a miscalculated 15-month vs 4-month window error), suggesting insufficient penalty for the underlying data/methodology issues in the source models they're relying on.

Neither considers that USGS magnitude revisions or delayed aftershock cataloging near the Dec 31 close date could create resolution ambiguity (e.g., a borderline M6.4-6.6 event needing recalibration), a scenario that could matter given the close proximity of two historical events (1971 M6.6, 1994 M6.7) to the 6.5 threshold.

6. Aggregator Fable 5.1 · final calibrated forecast

2% (77% confidence)

Yes: 2% No: 98%

Both forecasters and the brief converge on a base rate of roughly 0.5–1% for a M6.5+ within 50 miles of LA over the remaining ~117 days, derived from two qualifying events since 1900 and UCERF3 hazard scaled to the region; no qualifying event has occurred and no LA-fault advisories or precursor activity exist. The critique's points about catalog incompleteness, remote triggering from Brawley/Coachella swarms, and borderline-magnitude resolution ambiguity are valid reasons to hold a tail above the pure Poisson rate, but they are already reflected in shading to ~2.5% rather than ~0.7%. The Polymarket 7.2% proxy is thin, decaying, and typical of longshot overpricing in tail-risk markets, so it should not pull the estimate much higher. Final call: Yes ≈ 2.5%, No ≈ 97.5%.

Pipeline Timing

Total pipeline time: 365.0s

Per-tool research timings shown in the Research section above.