A precision-guided weapon can adjust its flight instead of following an entirely unguided path. That is the starting point, not a promise that every weapon will arrive exactly where intended. Its launch platform, navigation, sensors and operating conditions all shape the result.
Two weapons may share a bomb design but reach the air in different ways. Two advertisements may both claim exceptional accuracy while measuring different things. Understanding those differences is more useful than treating the longest range or smallest number as a league table.
The launch platform changes the starting conditions
An air-launched weapon begins with the aircraft's altitude and forward motion. Some are powered missiles; others glide after release. A ground-launched weapon must obtain the energy needed to leave the launcher and continue its flight, usually through a rocket motor. The labels describe launch origin, not a complete account of propulsion or guidance.
The Ground-Launched Small Diameter Bomb, or GLSDB, illustrates the distinction. Saab describes it as combining the Small Diameter Bomb design with a rocket-based launch system. The ground-launched version is a specific configuration, not simply an airborne bomb dropped from a different height.
Aircraft integration, launcher compatibility and the conditions under which a weapon is tested therefore matter alongside its name. Shared components do not automatically mean identical performance. For the broader distinction between flight path and propulsion, see our guide to ballistic, cruise and hypersonic missile labels.
Guidance is a combination of jobs
Navigation estimates where the weapon is. Guidance uses information to direct its flight. An inertial navigation system uses onboard sensors; satellite navigation, such as GPS, supplies another source of position information. These methods can work together.
A seeker provides information about what the weapon is approaching. Imaging infrared uses an infrared image; radar uses radio-frequency measurements; semi-active laser guidance detects reflected laser illumination supplied by another source. Command guidance instead receives instructions from outside the weapon. These are categories, and a particular munition need not contain every one of them.
Air-launched StormBreaker is a useful example of combination rather than a single label. In a November 2023 release, RTX described imaging infrared, millimeter-wave radar and semi-active laser alongside GPS and inertial navigation. That describes the available guidance architecture. It does not establish that every mode performs equally well in every environment.
Likewise, the presence of an inertial system or anti-jamming feature is not proof of immunity to interference. Our explanation of GNSS jamming and spoofing covers the distinction between disrupted signals and misleading ones.
Read a range figure with its conditions attached
Saab's GLSDB page publishes a range of 150 kilometers. RTX's November 2023 StormBreaker release stated that the air-launched weapon could travel more than 40 miles against moving land and maritime targets. Those are manufacturer statements about named systems, not results from a common independent comparison.
They also use different units. More than 40 statute miles is more than roughly 64 kilometers; nautical miles would mean something else. A conversion makes units comparable, but cannot make release conditions, profiles or configurations identical.
Neither headline figure provides a complete performance envelope. Release conditions, flight profile and the specific configuration belong in a meaningful comparison. Where a source leaves those conditions unstated, readers should not fill them in or assume the quoted figure applies to every launch.
Even a familiar product name can conceal a different development stage. In September 2025, RTX reported a ground-launched StormBreaker prototype test conducted that spring with a rocket motor. The announcement did not publish a numerical horizontal range. It therefore cannot support transferring the airborne figure to the ground-launched prototype, or calling the prototype operational on the strength of that test.
Accuracy is a distribution, not a fence
Circular error probable, usually shortened to CEP, is a statistical radius associated with 50 percent of the relevant impacts or position errors under the stated definition. The National Research Council's explanation makes the 50 percent distinction explicit. The 2017 NATO glossary similarly describes CEP as an accuracy indicator.
A CEP value does not mean that every round lands inside that circle. Nor does it specify the largest possible miss. Comparisons also need to establish the reference point, test population and conditions behind the statistic.
Saab's product page claims accuracy within one meter for GLSDB, but the wording alone does not supply a full statistical definition, sample size or test report. It should remain an attributed manufacturer claim. It cannot automatically be relabeled as a one-meter CEP or compared directly with another system's measured CEP.
A hit, a functioning weapon and an effect are different results
Accuracy asks where a functioning weapon arrives relative to its intended point. Reliability asks whether it functions as required. Effect asks what happened after arrival. These measures overlap in a mission, but they are not interchangeable.
A close impact does not establish the desired effect. A successful test does not reveal the failure rate of an entire inventory. A short video may show a real explosion without identifying the weapon, demonstrating the impact distance or establishing the later damage.
That is why reading battlefield claims requires more than watching the final seconds. Look for the exact variant, the source of the claim, the test conditions and the measured outcome. A useful performance account keeps those details together. The word precision is the beginning of the explanation, not its conclusion.