SITUATION BRIEF · 2026-09-16

NexusWatch · Yemen · 2026-09-16

🎯 Today's Call

The Ledger — 65 resolved today, 10 hit · OONI Brier 0.058 over 263, -2% vs base rate · FX Brier 0.147 over 718, +4% vs base rate · 928 open, next resolves 2026-09-17

65 calls settled against external ground truth: 10 hit, 55 miss. 19 due calls remain unresolved under the grace rule. The largest miss: Cameroon — we said 17% for currency depreciation by today against a 26% base rate; the XAF held. Our top open call: Singapore at 18% for ≥0.53% currency depreciation by 2026-09-22, ten points below its 28% base rate.

📊 Top Signal

Russia and Iran just recorded their highest censorship weeks in the entire OONI dataset we've seen.

Russia logged 51,505 confirmed blocking measurements in the last seven days — OONI's data, latest measurement 2026-09-15. Iran hit 46,519 in the same window, also ending yesterday. For context, India — a country with triple Russia's population — recorded 3,834. These aren't gradual escalations; these are infrastructure-scale blocking campaigns.

Why it matters: Network interference at this scale is a leading indicator, not a lagging one. In both countries, digital crackdowns have historically preceded physical ones by days or weeks. Iran's censorship component is now contributing +12 to its live CII deviation, on top of an 85 structural level. Russia's blocking has spiked without a corresponding CII move today, but the seven-day trajectory shows persistence — this isn't a one-day anomaly aging out of the feed. We're watching for secondary signals: OFAC amendments (VTB Bank was amended under three EOs on 2026-09-15), troop movements, or sudden FX volatility in the ruble or rial.

🌍 The Board

Movers Sudan 89 (▲+11.8) — driver not identified in today's data. Iran 85 (▲+15) — OONI recorded 46,519 confirmed blocking measurements in the last 7 days, latest 2026-09-15. Kenya 36 (▲+6) — driver not identified in today's data. Yemen 92 (▼-4.1) — disasters component decay; no external driver in today's data. Colombia 45 (▼-5) — driver not identified in today's data. Nigeria 65 (▼-4) — driver not identified in today's data.

Crises Iran → Censorship component at +12, the highest contributor to any live deviation globally today. Combined with OFAC's VTB Bank amendment under IRAN-EO13902 on 2026-09-15, this looks like coordinated preparation. Watch for: secondary sanctions, diaspora reporting of in-country arrests, or sudden closure of border crossings.

Sudan → Deviation jumped 3.6 points in 24 hours to 11.8, on a structural level of 89. The seven-day trajectory shows volatility (range: 6 to 14) but no clear external driver in today's data. Conflict and governance baselines remain maxed. Watch for: displacement reporting from ACLED or IOM, or a headline we haven't seen yet.

Markets USO (crude oil ETF) rose 3.32% to $161.86 — the largest single-day move in the basket. UNG (natural gas ETF) up 2.03%, XLE (energy sector ETF) up 2.17%. UUP (dollar index ETF) edged up 0.18%; TLT (Treasury bond ETF) fell 0.27%. SPY (S&P 500 ETF) dropped 0.46%. The energy move is pricing something — either Hormuz/Bab el-Mandeb concerns (Yemen at 92, Iran at 85), or front-running ahead of a supply disruption. No headline in today's data confirms a trigger, so this is either positioning or information asymmetry.

What would change our mind: A reversal in OONI's Russia or Iran measurements — if tomorrow's 7-day totals drop below 30,000 each, we'd read today as a measurement spike rather than a campaign. Or: a confirmed pipeline or port closure in the Gulf, which would explain the energy move with a concrete event rather than risk premium.

🙊 What We're Not Saying

🧨 The Long Fuse

Iran's seven-day censorship trajectory: 15 → 21.1 → 22.3 → 23 → 15 → 17.2 → 16.5 → 15. That's not a spike — it's a sustained campaign with a four-day peak mid-week. If OONI's measurements stay above 40,000 per week through month-end and OFAC adds further Iran-related designations, we're looking at pre-positioning for something larger. The last time we saw censorship at this scale without a corresponding headline, the headline arrived weeks later, not days. Conditional: if the 7-day rolling average stays above 6,000 daily measurements through September 30, treat this as infrastructure for a crackdown, not a one-time sweep.

Open NexusWatch →