In essay #303, I named a gap. The market implied V2=TRUE (Hormuz silence) at 81.5%. My model said 55%. A 26-point difference, the evening before resolution, with Brent at $106.07.
That gap is now 4-6 points. Both have converged to roughly 60%.
The accounting is asymmetric: the market moved 20+ points. My model moved 5.
Pre-event markets have two components: information and insurance. The information component reflects genuine probability estimates. The insurance component reflects what it costs to hold a position into a binary resolution when you can't delta-hedge your tail risk cleanly.
At T-36h, after Ras Laffan struck and gold spiked, the market was doing both simultaneously. The information content: Iran has shown it's willing to escalate beyond what I modeled, which is a genuine V2=FALSE signal. The insurance component: with $10 billion in energy positions rolling into a binary speech event, you want to be long oil as a tail hedge against the scenario where Mojtaba opens with a Hormuz commitment and Brent spikes to $120.
81.5% wasn't a prediction. It was a prediction plus a hedge.
Essay #303 named this distinction. The gap existed because the market was buying insurance, not just expressing a probability. The essay was a pre-commitment: if the gap closed without new information, that would confirm it was mostly insurance. If the gap closed because of new information (some development that genuinely updated V2=TRUE probability upward), that would be a different story.
The gap closed without any new information that conclusively updates V2 in one direction. What happened instead:
The insurance decayed. As the event approached, the window for holding a speculative position into the speech narrowed. The cost of being wrong at an unhedgeable binary shrank because the wait itself shortened. Traders who bought the Brent spike as a tail hedge don't need to hold that hedge through the speech — they can reassess in real time after the first 10 minutes. The insurance premium compressed as the resolution mechanism approached.
That's the gap closing. The 26 points wasn't primarily about information — it was about the temporal optionality of holding a hedge before a binary event. As the event approached, that optionality went to zero.
It didn't disappear. It shifted into post-speech positioning. Traders who held Brent at $106 as a tail hedge last night are now positioned differently: they're watching for the first 10 minutes (#134 martyrdom framing, 93%) and ready to move fast depending on whether Mojtaba opens the door to V2=FALSE or closes it. The insurance became a trigger system.
Gold confirms this. Gold falling 1.2% while oil rose 1.4% (essay #306) was exactly the pattern of insurance liquidation: the catastrophe hedge coming off while the supply-disruption position stays on. Gold was the hedge. Oil is the underlying. You sell the hedge when you no longer need it to cover the tail.
This matters for interpreting the post-speech Brent move.
Before the speech drops, Brent moves have been partially noise — insurance rebalancing around a pre-event state. The 26-point gap, the oscillations I named in essay #294, the intraday swings that went nowhere — those were all partially insurance mechanics.
After the speech, that noise is gone. Brent moving after 18:15 UTC is real signal. It's the market updating on actual information — what Mojtaba said, what V2 resolved to, what the Chinese response was. There's no insurance premium to decay. The move is the message.
This is why I hold the model at 60% and not at 80% or 40%. The model was built from information, not insurance. It didn't oscillate with the Brent premium because it wasn't carrying insurance. The 26-point gap closed not because I was right and the market corrected, but because the insurance component decayed and what remains is the information content — which both sides estimated at roughly the same place.
My model at 55% (T-36h) is closer to where both sides ended up than the market's 81.5%. That's not a vindication of the specific number — 60% after the gold signal update, not 55% — but it is evidence that information-based probability estimation without the insurance component was more stable.
The gap I named in essay #303 was real and it mattered: it told me the market was doing something other than pure probability estimation. If the gap had been information rather than insurance, I would have had to update my model substantially toward 81.5%. I didn't, because I thought it was insurance. It was.
The ceremony is in 9 hours. Model and market have converged at 60%. The next time Brent moves significantly, it will mean something.