How UAP Evidence Is Evaluated

A strange light is easy to record.

A reliable conclusion is harder.

The difference between the two is evidence quality.

UAP investigation is not simply watching a video and deciding what it looks like.

It is reconstructing an event from incomplete measurements.

Start with provenance

Before asking what an object is, investigators need to know where the evidence came from.

Important questions include who recorded it, whether the file was copied or edited, whether the original is available, when and where it was recorded, what platform carried the sensor, what mode the sensor was using, and whether metadata survived.

A viral clip may contain visually dramatic information while losing the data needed for physical reconstruction.

Eyewitness reports matter

Human reports can identify where to investigate.

Experienced pilots, radar operators, military personnel, scientists, and ordinary observers can all notice something that deserves attention.

But experience does not eliminate the limits of perception.

A witness may accurately report: “It appeared to accelerate.”

That is not identical to: “The object physically accelerated at a measured rate.”

The first is an observation.

The second is a physical claim that requires geometry and measurement.

Distance changes everything

An object can look fast because it is close.

It can look slow because it is far away.

A small nearby object and a large distant object can occupy a similar number of pixels.

Without distance, size and speed can become impossible to determine.

Platform motion matters

Many famous UAP videos were recorded from moving aircraft.

That creates a moving frame of reference.

The apparent motion in the image can come from the object, observer, camera rotation, aircraft banking, sensor tracking, zoom, or combinations of them.

Investigators therefore reconstruct the observer's motion as well as the target's apparent motion.

Line of sight

A camera records direction.

It does not always directly reveal three-dimensional position.

If the sensor knows its own position, orientation, range to target, and pointing angle, an analyst can estimate where the target is.

If some variables are missing, the analyst may need to model a range of possible solutions instead of one precise trajectory.

That range is part of the result.

Sensor mode

Infrared, visible-light cameras, radar, radiofrequency sensors, and human vision measure different things.

Infrared imagery records differences in detected radiation.

It does not automatically reveal the true shape, color, distance, or composition of an object.

Radar can provide range and velocity information but can also contain clutter or tracking artifacts.

No sensor is context-free.

Multi-sensor evidence

Multiple independent measurements can strengthen a case.

A visual observation supported by radar, infrared, and telemetry can provide more constraints than one sensor alone.

But quantity does not automatically equal quality.

Two sensors can share the same geometry problem.

Several observers can share the same mistaken assumption about distance.

Corroboration is strongest when different sources add genuinely independent information.

Environmental context

Investigators compare the observation with the environment.

Useful data include wind, cloud layers, temperature, illumination, Sun and Moon position, stars and planets, atmospheric effects, terrain and sea state.

An object's behavior relative to the wind can be especially important when evaluating balloons or lighter-than-air objects.

Known traffic

Analysts can compare a report with commercial aircraft tracks, military activity when available, unmanned aircraft, rocket launches, satellites, satellite flares, and balloons.

The objective is not to force every case into one of these categories.

It is to test ordinary explanations before claiming an extraordinary one.

The Go Fast example

The well-known Go Fast video appeared to show a small object moving extremely quickly near the ocean surface.

AARO reconstructed the geometry using information visible in the sensor display.

Its analysis placed the object around 13,000 feet rather than near the water.

After accounting for platform motion and wind, AARO assessed that the object did not demonstrate anomalous speed.

The apparent speed was strongly influenced by motion parallax.

AARO still did not definitively identify the object.

One question was answered: Did it demonstrate extraordinary speed?

Another remained open: What exactly was it?

Competing hypotheses

A rigorous analysis does not begin with one favored explanation.

It compares alternatives such as balloon, bird, satellite, aircraft, UAS, atmospheric effect, sensor artifact, or unknown physical object.

A hypothesis becomes stronger when it predicts the observed data better than alternatives.

Confidence is separate from identification

Investigators can sometimes make a high-confidence statement about one property while remaining uncertain about identity.

For example, high confidence that an object moved with the wind, high confidence that apparent speed was not anomalous, but insufficient data to identify the object itself.

Good analysis preserves those distinctions.

Active archive

AARO uses an active archive for cases that lack enough information for comprehensive analysis.

This prevents a weak dataset from being converted into a strong conclusion.

The case can be reopened if better information later appears.

What high-quality UAP data would look like

NASA's independent study emphasized better data collection.

A stronger observation would preserve calibrated sensor information, exact time, precise location, platform position and orientation, range, original uncompressed data, multiple independent sensors, environmental conditions, and chain of custody.

The more variables are known, the fewer explanations remain possible.

The goal is not to debunk

The goal is to identify what the evidence can support.

Sometimes the result is ordinary.

Sometimes it remains unresolved.

A good investigation is not measured by how exciting the conclusion is.

It is measured by how much uncertainty it removes without inventing certainty that the data do not contain.