Explainers
How Are Missing People Actually Found? The Science of the Search
When someone vanishes, the search is not guesswork — or it shouldn’t be. There is a real science to finding the missing, and it is far stranger and more rigorous than television suggests. The same branch of mathematics that located a sunken nuclear submarine and the flight recorders of a jet lost in the open Atlantic is used to decide where to look for a missing hiker. Decades of data on how lost people actually behave can draw a map of where they’re likely to be. And when the worst has happened, a quiet alliance of geology, botany, and trained dogs can find a hidden grave. Here is how the search really works — and what actually helps in the first hours, when it matters most.
The math that finds the lost: Bayesian search theory
In 1968, the submarine USS Scorpion was lost with all hands somewhere in hundreds of square miles of Atlantic seabed. The Navy found it by doing something counterintuitive: instead of searching the single most likely spot, they asked dozens of experts to wager on many competing scenarios, turned those bets into a probability map of the whole area, and searched where the expected payoff — the chance the wreck was there, times the chance they’d actually spot it if it was — was highest. Each empty patch updated the map and made the remaining patches more likely. They found the Scorpion. The same method, Bayesian search theory, later recovered the black boxes of Air France Flight 447 from the deep ocean in 2011 after earlier searches had failed, and it runs today inside the U.S. Coast Guard’s search-and-rescue planning software.
The core idea is humane as well as powerful: you don’t need to know where someone is to find them. You need to honestly weigh every possibility, look where the odds are best, and let each cleared area sharpen the map. It is the opposite of tunnel vision.
The data on how lost people behave
A searcher facing a forest doesn’t start from zero. For decades, Robert Koester and colleagues built the International Search and Rescue Incident Database (ISRID) — more than 145,000 real searches — and discovered that lost people are not random. They behave in patterns that depend on who they are. A person with dementia tends to travel differently than a lost child, who behaves differently than a despondent adult or an injured hiker.
From that data come hard numbers: for each category of missing person, how far they’re typically found from the place they were last seen, how long they keep moving, whether they go uphill or down, and whether they end up near roads, streams, or drainages. Searchers turn those statistics into distance-ring probability maps — concentric zones around the last-known point, weighted by likelihood. It is why a good search looks organized rather than frantic: it is being guided by the behavior of thousands of people who were lost before.
When the search is for a hidden grave
Some searches, tragically, are recoveries. Here a remarkable, quiet discipline takes over — pioneered by a Colorado group called NecroSearch that brought together geologists, botanists, entomologists, and dog handlers to find clandestine graves. No single method is reliable alone, so they are layered:
- Remote sensing — aerial and satellite imagery, LiDAR, and multispectral scans that reveal ground disturbance invisible from eye level.
- Forensic botany — a disturbed grave changes the soil, and the plants growing above it change too; vegetation can betray what’s beneath.
- Geophysics — ground-penetrating radar, magnetometers, and soil-resistivity surveys that sense a disturbance without a single shovel of dirt being turned.
- Cadaver dogs — still among the most sensitive detectors ever fielded.
Increasingly these are combined as a funnel: geographic analysis narrows the region, LiDAR narrows it further, and near-surface geophysics pinpoints where to look — each step shrinking the ground and sparing families a fruitless dig.
The database that reunites the missing with the unidentified
There is a heartbreaking gap in America: at any moment up to 100,000 people are actively missing, while 11,000+ sets of unidentified human remains sit in medical examiners’ offices. Sometimes the missing person and the unidentified remains are the same human being — in two different filing cabinets that never spoke to each other. The federal NamUs system exists to close that gap, matching long-term missing-person cases to unidentified remains using DNA, forensic genetic genealogy, dental records, anthropology, and fingerprints. It has helped resolve more than 46,000 cases. For a family, that match is everything: not the answer they prayed for, but the truth, and a place to grieve.
How we think about it
This is the lens we bring to missing-persons cases: a search is about places and probabilities, never about pointing a finger at a person we could be wrong about. The science above aims effort at the ground most likely to hold an answer. When suspicion about a living or deceased person arises, that belongs with investigators who have subpoenas and a duty of care — not in a headline. We help by understanding the search, honoring the missing, and pointing toward what actually works.
What actually helps in the first 48 hours
If someone you love goes missing, a few things genuinely matter:
- Report immediately. There is no 24-hour waiting rule; that’s a myth. The first hours are the most valuable in any search.
- Preserve the “last-known” details. Last location, what they were wearing, their phone, their car, their state of mind, who they were meeting. This is exactly what searchers turn into a probability map.
- Share recent photos and identifiers with police right away.
- Ask about NamUs for a long-term case, and about whether the case qualifies for DNA or genetic-genealogy work.
- Lean on advocates. The National Center for Missing & Exploited Children (1-800-843-5678) helps with missing children; local and national missing-persons organizations can guide families through the system.
A missing person is a person likely in trouble, and a family living a nightmare. The point of all this science is simple: to shorten that nightmare, and to bring people home — or bring them the truth.
Frequently asked questions
How do search teams decide where to look for a missing person? Modern searches combine probability mapping (Bayesian search theory, the same method used to find the USS Scorpion and Air France 447) with lost-person-behavior data — decades of statistics on how far different categories of missing people travel and where they’re typically found — to build weighted search zones rather than guessing.
Is there really a 24-hour wait before reporting someone missing? No. That is a myth. You can and should report a missing person to police immediately; the earliest hours are the most important to a successful search.
How are hidden graves found? Through a multidisciplinary approach — remote sensing and LiDAR, forensic botany, ground-penetrating radar and other geophysics, and cadaver dogs — layered so each method narrows the search area before any excavation.
What is NamUs? The U.S. National Missing and Unidentified Persons System: a database that matches missing-person cases to unidentified remains using DNA, genetic genealogy, dental records, and more. It has helped resolve over 46,000 cases.
Every missing person is someone’s whole world. If you’re searching for someone, we’re sorry — and we hope the people with the maps and the dogs and the databases bring them home. For support resources, see our support page.
Sources
- Bayesian search theory (overview) and the search for Air France Flight 447 (Stone et al.)
- Koester, Lost Person Behavior / the ISRID database
- Geoforensic methods for detecting clandestine graves (Journal of Forensic Science and Medicine, 2024)
- NamUs — National Missing and Unidentified Persons System (NIJ)