Index options tend to be priced for bigger moves than the index then makes: hedgers pay a premium for protection, and the seller who takes the other side collects it. This version sells out-of-the-money calls and puts on three index ETFs, QQQ, IWM and EEM, when implied volatility is high, and spreads the exposure across tech, small caps and emerging markets so no single index decides the outcome.
volatility risk premium / theta
Multi-index iron condor (QQQ, IWM, EEM)
The idea
How it was tested
Pre-registered on 2026-08-31. The signal is implied volatility rank above 60. On each symbol, one position at a time: a short call at 102% and a short put at 98% of spot, 30-45 days to expiry. Exit at 50% of max profit, after 25 days, or at a −50% stop. The spec lists only the two short legs.
Where it died
The first simulator run, on 2026-08-31, showed a 100% win rate. The cause was a sign bug: for short positions the stop had been computed at the same level as the target. After the fix the same simulator reported 131 trades, a 38.2% win rate and a 0% annual return, and the strategy was archived that day.
Neither number meant anything. The simulator had no option prices to work with, because the lab holds no historical option chains. On 2026-10-17 the lab-wide review recorded the strategy as never run, and the verdict changed from a measured failure to void.
The record
Thesis
Sell out-of-the-money calls and puts on three index ETFs when implied volatility rank is above 60, betting that implied volatility overstates the realized move. Spreading across tech, small caps and emerging markets lowers single-index exposure.
Specification
- signal
- IV rank > 60
- structure
- short call at 102% and short put at 98% of spot, 30-45 days to expiry, one position per symbol
- exit
- 50% of max profit, 25 days, or −50% stop
Gates
Pre-registrationpass
Was the hypothesis, its parameters and its pass/fail rule written down before any data was looked at?
Gate 7 — Jensen alphanot evaluated
After regressing out market exposure, is there alpha left?
Never measured on real option prices — the lab has no historical option chains.
Why it died
Never validly tested. A simulator run on 2026-08-31 first showed a 100% win rate — the stop for short positions had been computed at the same level as the target. After the fix it reported 131 trades, a 38.2% win rate and a 0% annual return, and the strategy was archived. But that simulator had no option prices to work with, and the lab's later review records the strategy as never run: it is unverifiable, not refuted.
What it taught
A 100% win rate is a bug report, not a result. And a short-options backtest without option prices measures the simulator's assumptions, whatever sign the number comes out with.
Verdict history
- 2026-08-31Pre-registered; simulator run with a sign bug (100% win rate), then fixed: 131 trades, 38.2% win, 0% annual — archived
- 2026-10-17Declared unverifiable: no historical option chains Void