The consensus trade for a military conflict is buy bonds.

Within 48 hours of the February 28 US-Israeli strikes on Iran, the 10-year Treasury yield was rising.

Here is what the data actually shows:

Before the first strike, large speculators in WTI crude oil already held a 33-week high net-long of 172,000 contracts. The smart money was not reacting to news. It had already placed the bet.

The bond selloff had three simultaneous drivers — not one. Inflation repricing was only a fifth of the Treasury move. The rest was term premium and something most coverage has entirely missed: petrodollar recycling collapse. Gulf revenues were stranded by the Strait closure. The structural buyer that historically cushions bond selloffs did not show up. That is why this was more violent than 2022.

Andurand made 20% in two weeks. Caxton lost $1.3 billion. The same trade — UK gilts — was the source of both outcomes depending on which side you held.

And then there is the TACO trade. Reuters confirmed via LSEG exchange data that $500M+ in oil futures changed hands in a single minute, 15 minutes before Trump's Truth Social post announcing a five-day strike pause. A Columbia Law study of 93,000 Polymarket markets found a 69.9% win rate across 210,718 suspicious wallet-market pairs — 60 standard deviations above chance. One account placed its first trade 71 minutes before the February 28 strikes at a 17% implied probability. Profit: $553,000.

Every claim in the full piece is sourced inline — CFTC filings, Nomura research, the Columbia Law academic paper, Reuters exchange data, court records.

Full piece here: