Lottie Williams was walking laps in a park in Turley, a suburb of Tulsa, Oklahoma, before dawn on January 22, 1997, when she saw a light cross the sky and took it for a shooting star.

A few minutes later something tapped her on the shoulder. She turned around and there was nothing behind her. On the ground was a piece of blackened fiberglass about five inches long.

It weighed roughly 16 grams, about as much as an empty soda can. She picked it up. She was not hurt at all.

That fragment made her the only person in the documented record ever struck by debris falling back from orbit.

What “only” is measured against

Guinness World Records lists Williams as the first person known to have been struck by orbital debris returning to Earth, and the record has stood since. The claim covers cases that were reported, investigated and confirmed. It does not claim that no human being has ever been brushed by a falling fragment.

The word “only” comes from the people who track this material. Armagh Observatory, writing on the twenty-fifth anniversary, called her still the only person to have been hit by space debris. Fox News reported in 2011 that The Aerospace Corporation’s own debris center listed her on its site as the only individual to be struck.

One entry, in the files of the organisations whose job is to keep the files.

The rocket

The light she saw was not a meteor.

The Guinness record entry names the object: an Aerojet AJ10-118K second stage from a Delta II rocket, coming down after months in a decaying orbit. That stage had launched an Air Force satellite from Vandenberg Air Force Base in California on April 24, 1996. The satellite was the Midcourse Space Experiment, a missile-defense sensor platform.

Nine months later the spent stage fell apart in the sky over Oklahoma.

Debris from a breakup does not land in one place. The main wreckage came down a couple of hundred miles away, in Texas, and NASA found the timing and location of the Tulsa strike consistent with that stage’s reentry and breakup. That match is why anyone could connect a five-inch scrap in a park to a rocket at all.

The fragment that could not be identified

Four years later, in 2001, Air Force experts examined a fragment from the event. Their conclusion was not a clean match.

The piece was too small and too badly deformed to be positively identified. What the analysts could say was that its composition was consistent with the insulator overwrap of the rocket engine’s thrust chamber, and that it had been heated to at least 1,200 degrees Celsius, the range recorded during an uncontrolled reentry.

Consistent with. Not confirmed as.

So the single documented case in the entire history of human beings and falling space hardware rests on a piece of material that could not be positively identified, examined four years after the fact.

The reentry is documented, and so is the timing. The fragment’s family is documented. Its exact identity is not.

The odds, and what they actually count

The figure most often attached to this story comes with three conditions.

The Aerospace Corporation, the nonprofit that studies reentering debris, estimates the chance that a particular person will be injured by falling space debris at less than one in one trillion. The Debrief quoted the estimate from it in 2024, and Armagh Observatory rounds it to one in a trillion.

That sentence carries less than it appears to.

It is a risk of being hit and injured, not simply hit. Williams was hit and not injured, so the event that actually happened is not the event the figure prices.

The figure is stated per person. It does not state a time window. One in a trillion per person per year and one in a trillion per person per lifetime are different claims, and the estimate as quoted does not say which it means.

Hit, injured, per person, no time window. That is the whole denominator on record.

Why the case sits in an insurance file at all

Falling objects are an ordinary named peril in property policies. The Insurance Information Institute’s standard list of covered perils carries “falling object” as item eleven on the HO-2, HO-3, HO-4 and HO-6 forms, and a rocket stage is a falling object.

There is also a treaty layer. The 1972 Convention on International Liability for Damage Caused by Space Objects puts liability for damage from a space object on the state that launched it. The only claim filed under it in the treaty’s published history came after Kosmos 954, a nuclear-powered Soviet satellite, came down over Canada in 1978.

Nobody had to test any of it in the Williams case, because there was no damage and no injury. A woman on a pre-dawn walk got tapped on the shoulder by a piece of a rocket and went on with her lap.

It is still the only one of its kind that anyone has written down.

The Peril Desk’s Risk Ledger prints the denominator on every odds figure it publishes, and this entry is the file where the denominator went missing.