After a volatility spike, fear fades faster than option prices assume. Sell an at-the-money call and an at-the-money put while implied volatility is at the top of its range, and collect both the decay, which is fastest at the money, and the fall in implied volatility. The other side is whoever bought protection in the spike.
volatility risk premium / theta
ATM straddle selling
The idea
How it was tested
Pre-registered on 2026-08-31. The signal is implied volatility rank above 75. Sell an at-the-money call and an at-the-money put with the same expiry, 30-45 days out. Exit at 50% of max profit, after 5 days, or at a −30% stop. One position at a time.
Where it died
A simulator run on 2026-08-31 reported 1,107 trades, a 48.0% win rate and a 0% annual return, and the strategy was archived that day. The same simulator had just been caught giving short-options strategies a 100% win rate because the stop was inverted. Fixed or not, it had no historical option prices to work from.
On 2026-10-17 a lab-wide review concluded that no option strategy here could be verified without historical chains. The archive stood, but the verdict changed from a measured failure to void.
The record
Thesis
When implied volatility rank spikes above 75, sell at-the-money calls and puts: after a spike, volatility should fall faster than at-the-money option prices imply, and ATM options decay fastest.
Specification
- signal
- IV rank > 75
- structure
- short ATM call and short ATM put, same expiry, 30-45 days
- exit
- 50% of max profit, 5 days, or −30% stop; one position at a time
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
Archived on 2026-08-31 after a simulator run reported 1107 trades, a 48.0% win rate and a 0% annual return. The same simulator had just been found giving short-options strategies a 100% win rate because of an inverted stop, and it had no historical option prices: the lab holds no option chains, so no number about this strategy rests on real premiums. The result is void rather than a measured failure.
What it taught
The verdict was right for the wrong reason: archiving on a number that cannot be trusted is not the same as knowing. "Unverifiable with the data we have" is a legitimate, and more honest, final state.
Verdict history
- 2026-08-31Pre-registered; simulator run: 1107 trades, 48.0% win, 0% annual — archived
- 2026-10-17Lab-wide finding: option strategies unverifiable without historical chains Void