On February 28, the United States and Israel launched Operation Epic Fury against Iran. Within days, the Strait of Hormuz — through which 20 million barrels of oil pass daily — went from 138 ship transits per day to five. Oil doubled from $70 to $119 intraday before settling near $99. The IEA coordinated its largest-ever emergency release: 400 million barrels from 32 nations.
This is not a report about the war. This is a report about what the war forced — the simultaneous repositioning across every major asset class that the CFTC's Commitment of Traders data exposes like an X-ray of broken bones.
The CFTC publishes what speculators (hedge funds, CTAs, macro funds) are actually doing with real money every Tuesday. The last three weeks — spanning the Iran shock — reveal a repositioning cascade that is internally contradictory, historically extreme, and not yet finished.
| Asset | Net Position (Feb 24) | Net Position (Mar 10) | Change | Signal |
|---|---|---|---|---|
| S&P 500 | -477,391 | -358,096 | +119,295 (covering) | SQUEEZE FAILING |
| 10Y Treasury | -1,987,780 | -1,878,928 | +108,852 (covering) | CROWDED EXIT |
| WTI Crude | -23,384 | -28,145 | -4,761 (adding shorts) | CONTRARIAN BET |
| Gold | +95,974 | +98,399 | +2,425 (holding) | CONVICTION |
| Japanese Yen | -26,317 | -49,219 | -22,902 (adding shorts) | DOUBLING DOWN |
| Euro | +36,797 | +5,231 | -31,566 (liquidating) | EXODUS |
The S&P squeeze that failed. Speculators covered 119,295 contracts — roughly $34 billion of mechanical buying pressure — in three weeks. By the latest CFTC data (Mar 13), net shorts had compressed further to -134,500. That's 343,000 contracts of short covering from the Feb 24 extreme. And SPY fell 3.6% during this period ($687 → $662). This is the market equivalent of running up the down escalator. The shorts were the last involuntary buyers. When they finish covering, the escalator wins.
Who is selling into it? Three candidates: (1) Pension funds and sovereign wealth funds mechanically rebalancing out of overweight US equities after the 82% rally since Oct 2023. This is structural, not discretionary. (2) Passive outflows as retail sentiment hits GFC levels — AAII bearish readings above 50% for three consecutive weeks. (3) Corporate insiders, who sold $12.4B in February, the highest monthly total since November 2024. All three are "natural sellers" — they sell because selling is their function, not because they have a view.
Crude shorts into a 40% rally. This is the most contrarian positioning in the COT data. Oil went from $70 to $99 and specs added 4,761 contracts of shorts. Their thesis: the war ends, Hormuz reopens, 400M barrels of SPR hit the market simultaneously, and crude crashes back to $65-70. This is a binary bet on ceasefire. Polymarket prices ceasefire by March 31 at just 14%. If the shorts are wrong, oil goes to $120 — priced at 42% probability — and the forced covering adds another $5-10 to the spike.
Gold holds while everything moves. Spec net longs barely changed (+2,425 contracts) while every other asset class saw violent repositioning. Gold is the one market where positioning and price agree: specs are long, price is up 37% in 12 months, and neither is flinching. This is either the last safe haven or the most complacent trade in the book. GLD did drop 1.3% on Friday ($460.84) — the first meaningful down day in weeks. Watch for gold specs to finally crack if the dollar rally (UUP +0.76%) accelerates.
Yen shorts nearly doubled. Specs went from -26,317 to -49,219 in three weeks — shorting the traditional safe-haven currency during a war. This only makes sense if you believe the dollar's war premium overwhelms yen's safe-haven bid, which requires the US to remain the clear military victor. The bet is that capital flows TO the US (flight to the hegemon, not flight to safety) — a subtle but critical distinction.
Euro longs destroyed. The most dramatic move in the table. Spec euro longs collapsed from +50,204 (Feb 10) to +5,231 (Mar 10) — a 90% liquidation in five weeks. Europe's exposure to Iranian oil and Middle Eastern gas is structurally higher than the US. The Euro positioning unwind is the market pricing in European energy vulnerability, which rhymes with the 2022 Ukraine playbook but is more severe (Hormuz dwarfs Nord Stream in barrel-equivalents).
Inversion Theory asks: who is forced to respond, and what card do they play? The Iran shock triggered a sequence of forced responses, each consuming optionality from a shrinking deck:
March 17-18. The Fed's first meeting since Operation Epic Fury. Three things matter more than the rate decision (which is hold, at 92% certainty):
The dot plot. Current median: one 25bp cut for 2026. If the median shifts to zero cuts, the market reprices aggressively — SPY tests $640-650 and TLT breaks below $85. If it shifts to two cuts, that's the dovish surprise that rescues the short covering into an actual rally. The spread between these two outcomes is roughly 4% of S&P equity value, decided by the median of 19 dots.
Powell's language on oil. Does he call it "transitory" (1970s echo, market rallies on hope) or "supply shock with persistent components" (market sells on honesty)? Powell is navigating between a president who has publicly threatened to fire him (Polymarket: Powell out from Fed Board by May 30 at 64%) and an inflation mandate that $100 oil makes impossible to meet.
The summary of economic projections (SEP). January GDP nowcast was 2.1%. February's -92K jobs and $100 oil likely push 2026 GDP forecast below 1.5%. If the Fed officially projects sub-1% growth with above-target inflation, that is the S-word — stagflation — in the official record for the first time since the 1970s.
Where prediction markets disagree with positioning — those are the signals.
Crude specs are SHORT while prediction markets price oil to $120 at 42%. This is the widest positioning-vs-probability divergence in the dataset. The specs are betting on ceasefire (14% probability) or SPR relief (which takes 120 days to fully deliver). The prediction market is pricing the more likely scenario: a protracted conflict with Hormuz staying shut. Someone is going to be violently wrong.
Similarly: recession probability at 33.5% with the Fed on hold, -92K jobs, and $100 oil. This feels too low. The 2008 playbook: recession was officially declared in December 2008, but it started in December 2007. The NBER lags by 12 months. We may already be in a recession that hasn't been named yet.
Inversion Theory is strongest when identifying what can't persist rather than what will happen:
1. Crude shorts at $99 can't persist if Hormuz stays shut. At 5 transits/day vs 138 normal, the physical supply deficit is ~17M bbl/day. The 400M barrel SPR release covers 23 days of deficit. After that, it's demand destruction or $150+ oil. Specs will be forced to cover or go bankrupt.
2. S&P short covering without a rally can't persist. Either the natural sellers exhaust (pension rebalancing has a finite target weight), in which case the next wave of covering produces the rally, or the natural sellers accelerate (recession + earnings downgrades), in which case SPY tests $620-640 regardless of positioning. The current equilibrium — shorts covering while prices fall — is a transitional state, not an equilibrium.
3. The Fed holding at 3.50-3.75% can't persist with -92K jobs AND $100 oil simultaneously. One of those two pressures will resolve. If oil drops (ceasefire), the Fed can cut. If jobs keep deteriorating, the Fed will be forced to cut regardless of oil — prioritizing employment over inflation as it did in 2001. The sequencing of which pressure breaks first determines whether we get recovery or stagflation.
4. Euro positioning at +5K can't go much lower. From +50K to +5K, the liquidation is 90% complete. The last 10% is either forced (stop-loss cascades) or held by conviction longs who won't sell. Either way, the euro's downside from positioning pressure is largely exhausted. If ceasefire happens, the euro snap-back could be violent — thin positioning amplifies price moves.
The framework says extremes produce opposites through forced responses. Let me try to disprove this:
What if the short covering succeeded? If you zoom out to S&P specs at -134K (latest data) vs -477K (Feb 24), the covering IS producing stabilization — SPY hasn't crashed to $620 despite a 40% oil spike, a -92K jobs print, and a war. The shorts were a CUSHION, not failed buyers. Without 343K contracts of covering ($34B of buying), SPY might be at $630. The squeeze didn't produce a rally, but it prevented a crash. That's not failure — that's shock absorption.
What if crude shorts are right? Wars in the Middle East have historically produced 3-6 month oil spikes followed by sharp reversals as supply reroutes and demand adjusts. The 1990 Gulf War: oil spiked from $17 to $41, then crashed back to $20 within 6 months. The specs might simply be reading history correctly while the prediction market overweights recency.
What if the forced response (SPR release) actually works? The 2022 SPR release of 180M barrels successfully brought oil from $120 to $75 over 6 months. The current release is 2.2x larger. If it delivers on the 120-day timeline, oil could be back at $70 by July — exactly what crude shorts are positioned for.
Verdict: the framework illuminates the forced-response chain but doesn't predict the resolution. The short covering data is more ambiguous than it first appears — it may be shock absorption, not failed bullishness. The strongest signal remains the crude positioning divergence, which will resolve violently in one direction.
The COT data is the clearest X-ray available of a market under geopolitical stress. Six asset classes repositioning simultaneously, each telling a different story but all responding to the same shock. The war forced the SPR card (irreplaceable below 400M barrels). The oil spike trapped the Fed. The short covering cushioned but didn't rescue equities. The euro longs got destroyed. The yen shorts doubled their bet on American hegemony. And gold — the oldest money — barely flinched.
Watch FOMC on March 18 for the next forced response. The dot plot is the market's next X-ray.