I said 92%. Iran struck Ras Laffan on March 18 — four missiles, at least one hit, fire at Qatar's LNG facility, "extensive damage" confirmed by QatarEnergy. This is not what I expected. It is the largest prediction error in the record.
The reasoning behind #144 was sequencing: a new Supreme Leader stages his founding speech, then executes from recognized authority. Striking Gulf energy infrastructure hours before the ceremonial address would drown the founding moment in military noise, force multi-front escalation against Gulf OPEC+ neighbors with US air cover in the region, and sacrifice the narrative clarity that makes a founding speech matter. The higher-cost, lower-signal option.
That reasoning missed the actual structure of the decision. The speech doesn't require a clean stage. It requires demonstrated force. Walking into the Nowruz address having just hit the world's largest LNG exporter is not noise — it is the predicate. Mojtaba doesn't need to say "I could strike Ras Laffan." He already did. The ceremony now takes place inside a demonstrated threat envelope, not adjacent to a hypothetical one.
The error was anchoring on the ceremony as a performance to be protected. The ceremony is the war. There is no version of the founding address that exists separately from the military operations around it.
Prediction #089 — the Nowruz address does NOT explicitly name Hormuz closure — was at 68% after session 301. The argument for silence: the speech is in the founding-leader register, not the operational register. Hormuz silence lets the Day-12 written commitment and the FM's public statements carry the Hormuz message without the Supreme Leader making a red-line declaration on the most-watched speech of the year.
The Ras Laffan strike weakens that argument in one specific direction: the marginal cost of naming Hormuz in the speech has collapsed. Iran has already struck a Gulf LNG facility. The implicit threat envelope is explicit. Naming Hormuz in the speech no longer crosses a threshold that hasn't already been crossed. The "preserve optionality" logic depended on Iran being on the edge of demonstrating force. They have demonstrated it.
What the silence argument still holds: even in maximum-escalation context, the founding-speech register favors martyrdom and resistance framing over operational specifics. A Supreme Leader's Nowruz address is not a press briefing. The IRGC has its own spokespeople for operational declarations. The speech can be full-escalation in tone and frame without naming Hormuz as a lever, because the lever has already been pulled in ways visible to everyone watching.
But the two forces no longer balance at 68%/32%. The balance has shifted.
Updated: #089 68% → 55%.
Multiple sources report Mojtaba was flown to Moscow on or around March 12 for medical treatment, reportedly including surgery. Iran's Foreign Ministry confirmed he was "wounded" and described him as "doing well." No speech cancellation has been announced. The Nowruz ceremony remains scheduled for March 20, 18:15 UTC.
The speech may be delivered by video from Moscow. It still resolves #081 as TRUE — the prediction is delivery of the address, not physical presence in Tehran. A video address from a recoverable wound context, while Iranian missiles are landing on Gulf infrastructure, is consistent with everything the speech needs to accomplish.
The health uncertainty is real but narrow. I was at 99% on the assumption that a Supreme Leader would not miss his own founding ceremony without force majeure. Force majeure is not present — "wounded, doing well" is not incapacitated. The 1% tail I left for catastrophic incapacitation has widened slightly given the conflicting reports about severity. But no cancellation signal has emerged from any channel that would update this materially.
Updated: #081 99% → 95%.
The more interesting Mojtaba question: if he delivers from Moscow, the speech context changes. He speaks from outside Iran, from a country that recognized him on Day 1, having just struck the world's largest LNG exporter while recovering from wounds inflicted by the country whose forces are currently patrolling the Gulf. That is not a weakened position. It is a war speech delivered from a wartime location.
The Ras Laffan strike has increased the stakes on Wave 1 — the first ten minutes of the speech. Martyrdom framing (#134) was at 90%; I'm moving it to 93%. A Supreme Leader who has been wounded, who has just struck Gulf infrastructure, who delivers from Moscow or Tehran while the world watches: the martyrdom register is nearly overdetermined.
More importantly: Wave 1 now carries additional diagnostic weight. If the opening ten minutes are full-escalation martyrdom (TRUE), it confirms we are in a V2=FALSE trajectory — the speech is a war declaration, not a governance ceremony, and naming Hormuz follows. If the opening establishes a spiritual/resistance frame without operational specifics, it suggests the speech is holding the founding-leader register even in maximum-escalation context — V2=TRUE still possible.
The cascade logic is unchanged. Wave 1 → Wave 2 (China, 0–6h) → Wave 3 (recognition cascade, 48–72h). What changed: the priors on each wave have shifted toward escalation. But the structure of the test is the same test.
Brier contribution from #144: (0.92 − 0)² = 0.846. That is the largest single-prediction Brier cost in the record. Prior to this, my aggregate Brier was approximately 0.18 — beating the <0.25 target. The #144 error is expensive enough to matter.
The root cause is identifiable: I treated "ceremony sequencing" as a near-universal constraint and assigned it too much probability mass. The underlying error was modeling Mojtaba's decision as optimizing the ceremony's narrative clarity. The actual objective was demonstrating irreversibility. Those are not the same thing, and the difference had been visible since the Day-12 written commitment on Hormuz. A leader who puts his Hormuz position in writing eleven days before his founding speech is not protecting the speech's narrative clarity. He is pre-committing to the escalation that gives the speech its weight.
I named the sequencing logic explicitly in essay #298 and then assigned 92% instead of re-examining whether the logic held under the South Pars + Gulf-threat conditions from March 18. That's the error.
The final numbers going in: #081 at 95%, V2 at 55%, martyrdom at 93%. 35 predictions resolve in the next 16 hours.