NexusWatch · Yemen · 2026-10-07

🎯 Today's Call

The Ledger — 67 resolved today, 29 hit · OONI Brier 0.058 over 263, -2% vs base rate · FX Brier 0.271 over 2085, -6% vs base rate · 939 open, next resolves 2026-10-08

67 calls settled against external ground truth: 29 hit, 38 miss. We called Colombia's currency depreciation at 31% (base rate 25%) — it hit. Bangladesh missed at 30% (base rate 24%). The aggregate rate is 43% on a set where base rates averaged 25%, meaning our adjustments added 18 percentage points of skill above the prior. Still open: Morocco at 35% (base rate 26%) for ≥1.4% depreciation by 2026-10-20, our highest divergence today.

📊 Top Signal

Iran and Russia are running the highest censorship measurements we've seen in the same 7-day window — 51,147 confirmed blocks in Iran, 48,832 in Russia, both measured by OONI through 2026-10-06. Iran's deviation jumped +18 today, the largest single-day move in the monitored set, driven entirely by the censorship component. This is not narrative — OONI records every connection attempt that gets a block page or timeout, and the volume is an order of magnitude above Turkey (2,181) and India (3,501) in the same period.

Why it matters: Censorship spikes of this scale precede crackdowns more reliably than any wire report. We've seen this pattern before political arrests, protest suppression, and information blackouts. Both countries are already at structural risk levels above 80 (Iran: 85, Russia: 67), meaning the baseline is already severe — this is escalation on top of that. If you're operating in either country, planning travel, or managing supply chains with exposure there, the next 72 hours are higher-risk than last week.

🌍 The Board

Movers Iran 85+18 (▲+18) — OONI recorded 51,147 confirmed blocking measurements in the last 7 days, latest 2026-10-06. Sudan 90+6 (▲+6) — driver not identified in today's data. Argentina 29+8 (▲+5) — driver not identified in today's data. South Sudan 89+3 (▲+3) — driver not identified in today's data. Afghanistan 88+0 (▼-3) — driver not identified in today's data. Cuba 41+7 (▼-5) — driver not identified in today's data.

Crises Iran → Censorship component at 12, the highest live signal in the set. OONI's measurement density suggests pre-emptive infrastructure lockdown rather than reactive blocking. Watch for: further deviation if Russia's trajectory converges with Iran's, which would signal coordinated posture rather than parallel domestic triggers.

Markets Crude oil ETF (USO) up 0.64% to $144.91, natural gas ETF (UNG) up 1.80% to $10.74. Energy sector ETF (XLE) gained 0.47%. The move is modest but directionally consistent with risk premium on Hormuz chokepoint exposure — Iran's censorship spike puts ~20% of global oil transits adjacent to an 85-level country with an active live deviation. Dollar index ETF (UUP) dropped 0.31%, which typically signals risk-on positioning, but the energy move suggests selective hedging rather than broad appetite. Gold ETF (GLD) up 0.72%, another counter-signal. What's priced in: baseline Iran risk. What's not: a multi-week censorship campaign converging with Russia's.

What would change our mind: OONI measurement density in Iran dropping below 5,000 confirmed blocks in any 24-hour period, sustained for 48 hours.

🙊 What We're Not Saying

• Argentina's +5 deviation move — no driver identified in today's data. No headline, no OONI spike, no sanctions event. We're not guessing.
• Sudan's +6 and South Sudan's +3 — both unattributed. The conflict and governance baselines are already catastrophic (Sudan: 90, South Sudan: 89), so these are incremental moves on top of standing severity, but we don't have the causal event.
• Cuba's -5 drop and Georgia's -4 drop — likely decay from stale disaster-feed data aging out of the 24-hour window, not de-escalation. We're not framing decay as improvement.
• The WHO feed lists five duplicate "Ebola disease caused by Bundibugyo virus - Democratic Republic of the Congo" entries — data formatting issue, not five separate outbreaks. Treating as one active notice until the feed resolves.

🧨 The Long Fuse

Iran's 7-day censorship trajectory has held between 12 and 18 since 2026-09-30, with no sign of decay. Russia's has held a similar band. If both countries sustain measurements above 40,000 weekly through 2026-10-14, we're looking at structural infrastructure enforcement rather than a reactive spike — the kind of posture that precedes extended blackout capability, not just protest suppression. Forward condition: if OONI records another week at this density, the probability of a coordinated information blackout in either country within 30 days moves materially higher.