A Self-Improving Prediction System

Google Cloud Next 2026 (April 22-24) proceeds as scheduled without cancellation or major postponement

The event is fully booked (Mandalay Bay Convention Center, $2299+ tickets, 3 days). Google has confirmed dates and livestream. No force majeure events expected. Per Cornelius: conference probability vs. content probability are different classes. The event proceeding is near-certain; what gets announced is the real prediction question.

Apple WWDC 2026 will begin with keynote on June 8 as officially announced

Apple officially confirmed WWDC 2026 for June 8-12. Apple has never missed a WWDC date after official announcement. Very high base rate confidence. Cornelius confirms WWDC timing is the one case where scheduled-date risk is low. Only extreme external force (pandemic-level) could alter this.

Trump will make public statements about Iran ceasefire on April 8 2026 (Truth Social or press)

Trump has posted about Iran every day of the conflict. Ceasefire is his biggest diplomatic win of the war. April 8 is the first full day after announcement. Near-certain he will claim success publicly. Pattern: Trump posts multiple times on major diplomatic events. Calibration note: already near ceiling at 0.96 after +10pp adjustment.

Qatar LNG Trains 4 and 6 will remain offline through April 15 2026

QatarEnergy stated 'weeks to months' minimum to normalize. Story says 3-5 year repair for these two trains. April 15 is only 16 days from now. Zero plausible mechanism for restart - physical battle damage to industrial infrastructure of this scale cannot be repaired in 16 days regardless of effort. P=0.96.

QatarEnergy will NOT announce any partial or full LNG production restart at Ras Laffan before April 10 2026

CEO Saad Al-Kaabi explicitly stated no restart until conflict completely ends. Trains 4 and 6 damaged (17% capacity, 3-5 year repair timeline). Iran deadline April 8 creates no pathway for near-term restart. Even a ceasefire on April 8 would not enable restart within 48 hours.

Google will announce at least one new Gemini model capability or major update at Cloud Next 2026 (April 22-24)

Google Cloud Next is the premier Google enterprise conference. Gemini is Google's flagship AI product. Sessions titled "What's New with Gemini from Google DeepMind" and "The Gemini 3 Playbook" are already listed. This is near-certain. Calibration note: markets domain is weak but this is tech domain where confidence is justified.

IDF will conduct airstrikes in south Lebanon or Beirut suburbs on April 8 2026

IDF issued evacuation warnings for Tyre and Shabriha (south Lebanon) on April 8 - a near-certain precursor to strikes. Netanyahu stated ceasefire excludes Lebanon. IDF has been conducting daily strikes throughout the conflict. Ground ops continue into south Lebanon. Calibration +10pp (base was already very high ~0.85+). Near-certain given evacuation warnings issued today.

Oklahoma City Thunder finish as #1 seed in Western Conference (regular season ends April 12)

OKC is at 61-16, a massive lead over #2 Spurs at 59-18. With 5 games remaining, OKC would need to lose all 5 while Spurs win all 5 to change seeds - essentially impossible. The #1 seed is all but secured.

Qatar LNG Trains 4 and 6 will remain offline through April 15 2026

Story explicitly states 3-5 year repair timeline for Qatar Trains 4+6 (17% of Qatar LNG capacity). These represent structural battle damage. April 15 return would require complete misrepresentation of damage assessment. Zero plausible mechanism for 3-5 year repair to complete in 17 days. P=0.95.

Starship Flight 12 (Booster 19/Ship 39) will NOT launch before April 15, 2026

Musk stated 4-6 weeks from late March, pointing to late April/early May. 33-engine static fire still required as gating item. Even in best case, regulatory approval and final integration takes weeks. Net date of April 30 is the provisional minimum. No credible path to April 15 launch. This is one of the very rare near-certain predictions justified by the physical constraints.

Iran military strikes will resume after Trump April 6 8PM ET deadline expires (by April 8)

Cornelius: cap strike-resumption confidence at 60-70% (streak extrapolation discount, day 36 not a forward probability). Ceasefire talks live, Iran-Oman drafting Hormuz protocol. Calibration: geopolitics underconfident +15pp. Net: base Cornelius cap 60% + partial geopolitics adjustment = 60%. Iran denies ceasefire request, Trump says Hormuz must be open first. No full resolution likely by deadline.

US conducts kinetic military strikes on Iranian targets by April 8 2026

Iran rejected ceasefire AND 15-point plan, reducing diplomatic off-ramps. However Oman channel still active and Trump historically extends. Selective Hormuz passage = Iran equilibrium not collapsing. Base ~20% escalation + 15pp calibration uplift for geopolitics underconfidence.

Trump announces another extension of Iran deadline beyond April 7 2026 8PM ET

Pattern: Trump has extended deadline multiple times already. Axios: Trump says good chance of deal by Tuesday. Mediators working to last minute for partial agreement. If no deal but ongoing talks, Trump has strong incentive to extend again rather than escalate (oil already at $112). Cornelius: de-escalation base case, face-saving unilateral reversal common. Geopolitics calibration +15pp: raw 0.45 → adjusted 0.55 (but capped by uncertainty).

Iran April 7 deadline (8PM ET) will result in ceasefire or formal extension rather than immediate US military escalation against energy infrastructure

Cornelius: graduated middle path (extension) more likely than binary formal deal or full escalation. Trump signaling optimism ("good chance deal by Tuesday"). Pakistan/Egypt/Turkey mediating. 4th Oman round concluded. Iran rejected 15-point plan but "message exchange via mediators" ongoing. Calibration: geopolitics short-horizon underconfident +15pp applied. Base ~0.52 + 15pp = 0.65.

No formal Iran-US nuclear deal will be signed by April 7, 2026 8PM ET

Iran FM confirmed only "direct messages" not formal negotiations (Al Jazeera March 31). Witkoff 15-point action list contains maximalist conditions. Formal full nuclear deal requires months of technical verification agreements. Calibration note says geopolitics underconfident by 15pp but this is near-certain territory. P(formal deal in <36 hours) ~12%.

No official Iran-US ceasefire deal will be announced before April 7 2026 8PM ET

Iran has publicly rejected Strait of Hormuz reopening as part of any temporary ceasefire. Multiple sources say slim probability of deal in 48 hours. Iran countered with 5 conditions. Trump extended deadline to Tuesday 8PM ET from original. Mediators (Pakistan, Egypt, Turkey) active but Iran public posture is defiant. Geopolitics calibration: +15pp applied to natural estimate of 0.70. Cornelius: Iran public denial coexists with back-channel acknowledgment but timeline too tight for formal deal.

US Q4 2025 GDP third estimate (released April 9) will show growth below 1.0% annualized

Second estimate was +0.7% annualized. Third estimates rarely deviate more than 0.2-0.3 percentage points from second. P(third estimate < 1.0%) is very high. Would need +0.3pp upward revision to 1.0% from 0.7% - unusual but possible if trade data revised. Economics is a weaker calibration domain (-conservative). P ≈ 0.88.

Trump conducts airstrikes on Iranian power plants or bridges by April 8 2026 04:00 UTC (tonight 8PM ET deadline)

Trump set hard April 7 8PM ET deadline, said explicitly "will not extend," defense officials confirm no extension planned. Iran rejected 45-day ceasefire and refused to reopen Hormuz. No deal reached as of April 7 morning. Cornelius: failed backchannels are bearish signals not neutral; deadline breach is resumption signal. P(deal by 8PM) ~20% = P(no deal) ~80% * P(Trump follows through) ~65% = ~52%.