Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a Polymarket question about whether Iran experiences a near-total (≤1% of 4-week max bytes on Cloudflare Radar) internet blackout for 6+ consecutive hours before Aug 31, 2026. The key drivers are base rates of Iranian nationwide shutdowns (June 2025 Israel-Iran war, Sept 2022 Mahsa Amini protests, Nov 2019 fuel protests) and current geopolitical/protest tension levels. Primary anchor is the Polymarket price; secondary is news on Iran-Israel escalation, internal unrest, and any current Cloudflare-observed traffic drops.

# Current state Iran's internet is currently (as of mid-August 2026) restored following the 88-day nationwide blackout that ran Jan 8–May 26, 2026; connectivity is back to ~86% but heavy filtering ("filternet"), WhatsApp restrictions, and periodic social-media outages persist. No new total blackout (Cloudflare-style ≤1% for 6+ hours) has been reported since restoration. Polymarket prices this specific market at 7.5%, down sharply from a 41.5% high, reflecting reduced near-term expectation of a repeat blackout given the current Israel-Iran ceasefire. # Timeline of key events - 2019-11 (confirmed): First total blackout — traffic fell to ~5% of normal for ~1 week, suppressing 2019–2020 protests (Wikipedia). - 2022-09/10 (confirmed): Mahsa Amini protests trigger repeated regional/nationwide internet disruptions (Wikipedia). - 2025-06-13 to ~06-24 (confirmed): Twelve-Day War (Israel/US strikes on Iran) accompanied by a shorter internet blackout (Wikipedia: "2025 Internet blackout in Iran"). - 2026-01-08 (confirmed): New nationwide blackout begins amid 2025–2026 protests and the "2026 Iran War" (concurrent US/Israeli strikes campaign) (Wikipedia, CNBC). - 2026-01-08 to 2026-05-26 (confirmed): Blackout persists 88 days (4mo18d), the longest ever recorded; estimated $35–80M/day economic cost, ~$1B cumulative by day 50 (EJIL:Talk!, Wikipedia). - 2026-05-26 (confirmed): Internet largely restored per NetBlocks; ~86% connectivity but social media (WhatsApp, etc.) remains blocked; court challenge creates uncertainty over durability (Iran Intl). - 2026-07 (reported): Facebook/Instagram experience partial outages; WhatsApp still requires circumvention; no full blackout (CNN, gdelt). - 2026-07/08 (reported): Continued unrest — executions of protesters (incl. a 19-year-old, Aug 3), UN Fact-Finding Mission calls to halt executions, US senators (Ted Cruz) urging arming protesters — signals unresolved domestic tension (globalsecurity.org, ibtimes). - 2026-07-16 onward: Polymarket Israel-Iran ceasefire markets show high confidence ceasefire holds through Aug 15 (99.85%) and Aug 31 (94.5%), implying low near-term war-driven blackout risk. # Event Will Iran experience a Cloudflare-verified total internet blackout (≤1% of 4-week max traffic for 6+ consecutive hours) by August 31, 2026? # Outcomes to forecast - Yes - No # Kalshi market anchor No direct Kalshi price was retrieved (kalshi_direct not returned); the Polymarket price for this identical ticker/question is the best available anchor: **7.5% YES**, down from a 41.5% high, -34pp over 30 days, -1pp over 7 days, on $170K volume. Trend strongly declining, consistent with post-ceasefire de-escalation. # Sub-question answers 1. **Polymarket price/history** — Currently 7.5%, was as high as 41.5% (likely during active war/protest period), now falling steadily (-34% 30d) — market sees blackout risk as receding. [polymarket_direct] 2. **Historical base rate** — 3 qualifying nationwide blackouts in ~7 years (Nov 2019, Sep-Oct 2022, Jun 2025/Jan-May 2026), giving λ≈0.036/month; Poisson model for the remaining ~2-4 month window to Aug 31 gives ~7-13% at historical rate, up to ~19-35% if tension-elevated regime assumed. [code_execution] 3. **Iran-Israel/US tension** — Ceasefire currently holding with high confidence per Polymarket (94.5%-99.85% through Aug 15/31, 2026); no active war reported as of Aug 2026. [polymarket_related] 4. **Internal unrest/economic crisis** — Yes: protests continue, executions of 2026-protest detainees ongoing (Aug 2026), UN condemnation, US politicians pushing regime-change rhetoric; economic damage from the blackout was severe ($1B+), adding financial strain. [gdelt, claude_news] 5. **Recent Cloudflare/NetBlocks state** — Post-May 26 restoration to ~86% connectivity; "filternet" remains with WhatsApp blocked; sporadic Facebook/Instagram outages in July 2026, but no confirmed new total blackout through mid-August 2026. [claude_news, gdelt] 6. **Partial throttling vs total blackout policy shift** — Iran maintains the National Information Network (NIN) intranet infrastructure enabling targeted filtering; post-restoration behavior (selective platform blocking rather than full shutdown) suggests some preference for partial control, though Iran still imposed a full 88-day blackout in 2026 despite high cost — full-blackout capability/willingness remains intact. [Wikipedia: National Information Network, Internet censorship in Iran] 7. **Related tail-risk markets** — Israel-Iran ceasefire holding (94.5-99.85% Yes), Iran MOU withdrawal unlikely (2.5% Yes), Kharg Island stable under Iranian control (99.25% No change), US-blockade-ends uncertain (7.5-17.5% Yes) — collectively imply markets see low imminent war/escalation risk. [polymarket_related] # Key facts (high-confidence, factual) 1. [Wikipedia] Iran's longest-ever blackout ran Jan 8–May 26, 2026 (88 days), driven by 2025-2026 protests and war with Israel/US. 2. [Iran Intl] Restoration to ~86% connectivity occurred May 26, 2026; social media (WhatsApp) remains blocked. 3. [EJIL:Talk!] Blackout cost estimated $35-80M/day, ~$1B cumulative — high economic pressure to avoid recurrence. 4. [code_execution] Historical base rate: 3 events/84 months ≈ 3.6%/month hazard rate. 5. [polymarket_related] Ceasefire markets currently near-certain (94.5%+) through Aug 2026. # Cross-market signals - Kalshi related: No Iran-blackout-specific Kalshi market found; adjacent Israel-Iran normalization/succession markets show low-to-moderate uncertainty, not crisis pricing. - Polymarket: This market's own price 7.5% (declining); ceasefire-continuation markets 94.5-99.85%; MOU-withdrawal 2.5%; blockade-end 7.5-17.5%. - Sportsbook implied: N/A. # Analyst opinions and speculation - claude_news synthesis: "base rate risk remains non-trivial... but no confirmed active or imminent new nationwide shutdown" as of latest reporting. - Ongoing execution of protesters and Cruz's regime-change rhetoric suggest domestic risk factors persist even absent active war. # Directional lean per outcome - **Yes**: Supported by recent (2026) precedent of full blackout under similar protest/war conditions, unresolved unrest, high blackout capability (NIN infrastructure); opposed by ceasefire holding, high economic cost deterrent, no current reporting of renewed blackout, and declining market price. - **No**: Supported by ceasefire markets near-certain, restored connectivity holding since May, declining Polymarket price (7.5%), partial-filtering approach post-restoration; opposed by residual protest/execution unrest and demonstrated willingness to reimpose blackouts. # Gaps / unknowns - No direct Kalshi YES price retrieved for this ticker. - Unclear whether prior Jan-May 2026 blackout counts toward "by Aug 31, 2026" resolution or if market resets post-restoration (price behavior suggests latter). - No granular Cloudflare Radar data point-level check provided for the July-August 2026 period. # Calibration anchors - Polymarket YES price (anchor): 7.5%, down from 41.5% high. - Historical Poisson model: ~7-35% depending on assumed hazard rate over remaining window. - Precedent: 3 total blackouts in 7 years, most recent lasting 88 days into May 2026.

1. Both forecasters converge tightly around 5-7% without deeply grappling with the ambiguity flagged in "Gaps/unknowns": whether the market resets post-restoration or whether resolution criteria could still capture edge cases (e.g., a shorter but qualifying 6+ hour blackout tied to a new unrest flare-up) — neither explicitly addresses how a sudden execution-driven protest surge (the Aug 3 execution of a 19-year-old is called out as a live flashpoint) could produce a short, sharp NIN-triggered shutdown independent of the Israel-Iran ceasefire track. 2. Forecast 1's math is internally inconsistent: it computes "~2% risk over the remaining ~2 weeks" from the 3.6%/month hazard rate, then lands on 5% — implying it upweighted domestic-unrest tail risk by 2.5x versus the base rate, but this adjustment is asserted rather than justified with a magnitude (e.g., why not 3x or 1.5x given the Poisson model itself already offers a "tension-elevated" scenario of 19-35%). 3. Neither forecaster grapples with the fact the brief's own Poisson model gives a wide range (7-35%) "if tension-elevated regime assumed," and given three of Iran's last four years included a qualifying blackout during periods of unrest that look similar to now (ongoing protests, executions, UN condemnation), both may be underweighting how "elevated" the current baseline really is versus assuming a return to the pre-2019 calm regime. 4. Both anchor almost entirely on Polymarket's 7.5% without any independent skepticism about that price's own reliability — the brief notes this market swung from 41.5% to 7.5% in 30 days on relatively thin $170K volume, suggesting it could be reactive/momentum-driven rather than a stable equilibrium reflecting durable ceasefire fundamentals, yet neither forecast discounts for potential over-shoot in the "No" direction. 5. Neither forecast explicitly addresses the scenario where a total blackout occurs due to non-Israel/US-war triggers (e.g., a new round of Mahsa-Amini-style mass protests sparked by executions) — the 2019 and 2022 precedents were purely domestic-unrest-driven blackouts, not war-driven, so conditioning heavily on ceasefire-holding markets (94.5-99.85%) may be treating the wrong trigger as dominant. 6. Both forecasts state high confidence (0.70 and 0.61) in essentially the same point estimate, but given the genuine model spread (2% to 35% depending on regime assumption) and the acknowledged data gap on Kalshi's actual price, this confidence level seems overstated for a question with this much structural uncertainty in the hazard-rate assumption.

6% (66% confidence)

Yes: 6% No: 94%

With only ~2 weeks left in the window and connectivity restored since May 26 (~86%), the pure base-rate hazard (~3.6%/month) implies ~2% residual risk, while the Polymarket anchor of 7.5% likely embeds both a wider remaining window at the time of pricing and momentum effects on thin volume. The devil's advocate correctly notes that the 2019 and 2022 blackouts were domestic-unrest-driven, so conditioning on ceasefire markets alone understates the trigger set — ongoing executions (Aug 3) and protest flashpoints could produce a short but qualifying 6+ hour NIN shutdown independent of the Israel-Iran track. That argues for staying above the naive base rate, but the $35–80M/day economic cost of the 88-day shutdown and Tehran's demonstrated post-restoration preference for selective platform filtering are meaningful deterrents against another full cutoff in the next fortnight. I land at 6% YES — slightly below the possibly-stale market anchor given the shrinking window, but above the pure Poisson calculation to respect the elevated-tension regime the critique highlights.

Total pipeline time: 275.5s