Week 1: The Experiment Begins

Forecast date: March 16, 2026
Target date: March 23, 2026 (7-day horizon)
Current price: Brent $99.47 | WTI $98.33

The Forecast

Point estimate: $113.55
68% confidence interval: $91.75 – $135.35
95% confidence interval: $69.95 – $157.15

Naive baseline comparison: $97.28 (current price as prediction)

Scenario Probabilities

Why These Probabilities?

Escalation weighted at 40% based on:

1. Strait of Hormuz partially closed (March 7) — 20% of global oil supply disrupted
2. UK active participant — RAF engaged Iranian drone (March 8), PM authorized US use of UK bases
3. Price velocity exceeding forecasts — Brent jumped $80 → $108 in 6 days (March 2-8)
4. Institutional forecasters lagging reality — Goldman Sachs revised forecasts 4× in 6 days, each time playing catch-up

The institutions are underestimating speed and severity. Markets moved faster than their worst-case scenarios.

What Happened This Weekend

Escalation signals:

• Tehran's Mehrabad Airport struck (March 15 evening) — residents report "bunker bomb" level strikes
• Iran drone attacks on Citibank buildings in Dubai/Bahrain (March 13) — financial targets now in play
• Guardian headline (March 15): "Oil could pass 2008 record of $147.50 a barrel" — Kharg Island damage compounding Strait disruption

Deescalation signals:

• No vessel attacks in Strait of Hormuz for 48 hours (March 13-15) — tactical pause or genuine pullback?
• Trump calling for China, France, Japan, South Korea, UK to send warships — diplomatic pressure building

Current price: $99.47 (up from $97.28 on March 13)

Markets are climbing but not exploding. Volatility decreasing (7.4% daily range on March 12 → 3.99% on March 16). Either:

1. My escalation scenario is overstated, OR
2. Markets are waiting for clearer signals before pricing in full disruption

Belief Checkpoint

What I'm watching this week:

• Does the 48-hour Strait pause extend or break?
• Do airport strikes escalate into broader infrastructure targeting?
• Do institutional forecasts continue lagging reality?

Where I could be wrong:

• Escalation scenario might be too high if diplomatic pressure works
• Stalemate scenario underweighted if this becomes a grinding, months-long conflict at current intensity
• Black swan scenario too low if Saudi facilities or major production hubs get hit

The goal: Update beliefs based on evidence, not ego. If I'm wrong, say it clearly and explain why.

Methodology Reminder

Scenario-weighted forecast:
Point estimate = Σ (probability × target price)

Confidence intervals:
Derived from weighted volatility assumptions (normal distribution)

Naive baseline:
Current price as prediction (no model). This is what I'm trying to beat.

Scoring (starts Week 2):
Mean Absolute Error (MAE) vs naive baseline. Directional accuracy. Confidence interval calibration.

Next Updates

Wednesday (March 18): Social drop — key insight from market data
Thursday (March 19): Mid-week check — any major events shifting probabilities?
Friday (March 21): Weekly retrospective — what I learned, what I missed

Target date: Sunday, March 23 — we'll see if $113.55 was anywhere close to reality.

Why This Experiment Matters

Forecasting under fire is different from forecasting in peacetime.

When events move fast and stakes are high, the temptation is to:

• Anchor to institutional forecasts (safety in consensus)
• Freeze in uncertainty (paralysis feels safer than being wrong)
• Revise history ("I always thought...")

This experiment forces me to:

• Lock forecasts publicly (no retroactive edits)
• Update beliefs systematically (document what changed and why)
• Score accuracy ruthlessly (did I beat a naive baseline?)

Eight weeks. Locked probabilities every Monday. Public scoring every Friday.

Let's see what I learn about thinking under pressure.

Forecast locked March 13, 2026. No edits after publishing.