Put credit spreads, measured

What a put credit spread is

Also called a bull put spread. You sell a put and buy a lower-strike put with the same expiry. You collect the difference in premium, and your worst case is the distance between the strikes minus that credit. The long put caps the loss, so the broker holds only the width of the spread, not the full strike.

That changes the arithmetic. A cash-secured put ties up the whole strike to earn its premium; a spread earns less premium on far less capital, which is why its return on the money at risk is much higher, and why it is easy to oversize.

What ten years of quotes say

We replayed about 1,000 entries on QQQ and SPY from 2016 through 2026: a 25-delta short put, a long put 5% below, 30 to 45 days to expiry, shorts priced at the bid and longs at the ask. The sample covers three bear markets: late 2018, early 2020 and 2022.

On QQQ the average spread earned about +6% on the dollars at risk, with a 95% interval of +2% to +9%, and 89% of spreads won. It was positive in nine of eleven years. The two losing years were 2018 and 2022. SPY was positive too, a little weaker. By comparison the plain short put earned about +0.4% a trade, on roughly twenty times the capital.

Timing did not help. Standing aside below the 200-day moving average, or on a volatility signal, never beat simply holding the spread across those years; the stand-down is insurance with a known cost, not an edge.

The part that hurts: sizing

A spread's loss is a gap, not a slow drift. When it loses, it loses most of the width at once, and a stop-loss did not help in our tests because the price jumps through it. The only lever that works is how much of the account is at risk.

In a bootstrap of 129 real months, risking 10% of equity per spread never produced a 20% drawdown within a year. At 15% the odds were about 3%, at 20% about 15%, and at 30% about 38%. Most of the damage people attribute to the strategy is the size they ran it at.

Diversify the factor, not the ticker. Spreads on QQQ, TQQQ and a tech stock are the same bet. In our tests the one underlying that improved a QQQ spread book was gold, with a correlation near 0.2, and even that is a lead, not a finding.