Escalation-by-attrition dominates today's outlook: an active North Korean missile cycle and Iran's counter-espionage crackdown run alongside Armenia's near-complete break from Russia's CSTO, while China's property slump stays entrenched — but a direct Russia-NATO clash remains a low-probability tail despite Polymarket noise.

Military is my failing sector (Brier 0.450, over by 35pp) and the military agent is disqualified (Brier 0.45), so I apply heavy discipline and lean on the base rate rather than escalation rhetoric. But the base rate here is exceptionally strong and fresh: fact-check confirms NK fired 10+ SRBMs on Aug 20 (yesterday) and a launch on Aug 6 — Pyongyang is in an active launch cycle, not a dormant period. During active cycles, 'another launch within 30 days' historically resolves true well above 0.80. I compress toward 50 for my sector bias, landing at 0.72 (down from an unadjusted ~0.82). This respects the lesson from forecast memory (anchor on the specific event-type base rate, incorporate countervailing factors). What would make me wrong: a diplomatic overture/summit signal prompting a pause, a technical or resource constraint after a heavy volley, or a deliberate quiet period around a domestic political date.