Report 196 · Defense Tech
The F-35 is routinely described as appearing "the size of a golf ball" on radar. That comparison is about radar cross section, a measure of reflected power. It says much less about distance than it sounds like, because a public Navy engineering handbook shows a radar's detection range falls only with the fourth root of that number. Shrink the echo ten thousand times and the radar still sees you at a tenth of the range.
I was a Counter-IED and Electronic Warfare Officer, and "the size of a golf ball on radar" is one of those lines that is technically about the right quantity and still leaves most readers with the wrong picture. It shows up everywhere. A March 2026 WION explainer says the F-35 "features a radar cross-section of roughly 0.0015 square metres," and that on enemy screens the jet "appears no larger than a standard golf ball." A 2025 Business Insider piece quotes Billie Flynn, described as a former Royal Canadian Air Force combat commander and Lockheed employee who worked on the F-35 program, describing radar cross section for a fighter like the F-35 as "something the size of a ping pong ball or a golf ball or smaller."
I can't verify the 0.0015 figure. Real stealth signatures are not public, and I'm treating that number as what the coverage reports, nothing more. The point of this report doesn't depend on it. It depends on what any radar cross section number does, and doesn't do, to the distance at which a radar sees you.
The clearest public reference is the Naval Air Warfare Center Weapons Division's Electronic Warfare and Radar Systems Engineering Handbook (NAWCWD TP 8347), a free, unclassified engineering handbook I've used on this beat before. Section 4-11 defines it plainly:
Radar cross section is the measure of a target's ability to reflect radar signals in the direction of the radar receiver
It is not the target's physical size. The handbook breaks it into three factors multiplied together: projected cross section, reflectivity, and directivity, meaning how much of the scattered power goes back toward the radar rather than off in some other direction. A flat plate facing the radar is enormous; the same plate tilted away is almost nothing. That is the whole idea behind stealth shaping: send the energy somewhere other than back home.
The reference object is a smooth metal sphere, because a sphere looks the same from every angle. So when someone says a jet "looks like a golf ball," what they mean is that it reflects about as much power back to the radar as a metal sphere of that size would. It's a statement about echo strength, not about how far away the radar can pick it up.
Here is the part that the golf-ball line skips. A radar's pulse has to travel out to the target and back, spreading out both ways, so the echo it receives falls off with the fourth power of range. The handbook works through what that means for radar cross section, which it writes as σ, directly from the two-way range equation in its section 4-4. Its conclusion, in its own notation:
Therefore, R4 ∝ σ or σ1/4 ∝ R
In words: detection range is proportional to the fourth root of radar cross section. That root crushes big ratios into small ones. This is my arithmetic from the handbook's relationship, all else about the radar held equal:
Eleven kilometres instead of a hundred is a big deal tactically. But it is nowhere near what "golf ball" suggests, which is something you'd only spot if it were sitting in front of you. The handbook makes the same point in miniature with its own worked example: reduce a target's radar cross section to 75 percent of its original value and "the detection range of the radar for the smaller RCS target (jamming not considered) is 0.93 (93%) of its original value." A quarter less echo buys seven percent less range.
The handbook runs the same exercise for burn-through, the range at which a radar can see a target through that target's own jammer. There the relationship is a square root, not a fourth root: "RBT2 ∝ σ or σ1/2 ∝ RBT." And for the jammer's required power, it is a straight line: the power needed scales directly with radar cross section. So the same 10,000-times cut that shortens raw detection range tenfold shortens burn-through range a hundredfold, and lets a jammer get the same effect with a ten-thousandth of the power. That is my arithmetic again, from the handbook's three relationships. It's why low observability and jamming are planned together, a point I worked through in Report 161. A small signature on its own buys range. A small signature plus a jammer buys much more.
The other thing a single radar cross section figure hides is that it is not one number. The handbook says the radar cross section of a real aircraft "must be measured" and "varies significantly depending upon the direction of the illuminating radar." Its own sample plot of a conventional jet ranges from 100 square metres broadside down to slightly more than 1 square metre at some angles, and it adds that the radar return "will vary by a few dB as frequency and polarization vary." Then it says, without softening it:
These RCS values can be very misleading because other factors may affect the results.
Run the fourth root on the handbook's own example: that 100-times swing between angles is about a 3.2-times swing in detection range, for the same aircraft, depending only on which way it's facing the radar. So a headline figure is, at best, one slice of a pattern, usually the most flattering one, and probably the frontal view.
Frequency matters too. The handbook shows that even the simple sphere only behaves like its tidy textbook formula when it is large compared with the radar's wavelength. When the wavelength gets close to the object's size, the sphere's echo swings between about four times and a quarter of the textbook value. The WION piece notes that the jet "can be detected by large, low-frequency early-warning radars." I'm not going to claim anything about any real aircraft's low-frequency signature, but the handbook's sphere example is a reminder that a "golf ball" number is a number at some frequency, not at all of them.
Any real stealth aircraft's radar cross section. The 0.0015 square metre figure is the reported value from popular coverage. I have no primary source for it, and the argument here works for any value.
Real detection ranges. The 100-to-11 kilometre example is illustrative arithmetic with every other radar parameter held fixed. Real detection depends on radar power, antenna, signal processing, clutter, and the geometry of the encounter, none of which the handbook pins to any fielded system.
Anything beyond public-domain physics. This report uses the handbook's general relationships. It describes no aircraft's signature pattern, no radar's capabilities, and no countermeasure technique.
"The size of a golf ball on radar" is a real comparison about the right quantity, echo strength, and a misleading one about the quantity readers actually care about, distance. Because detection range follows the fourth root of radar cross section, a ten-thousand-fold smaller echo still leaves the radar seeing you at a tenth of the range, and the number itself changes with angle and frequency. Stealth is a very large shrink of the bubble in which a radar sees you. It isn't a vanishing act. When the next "golf ball" line shows up, ask two questions: compared with what, and from which direction.
Scope note: this report explains publicly released radar cross section relationships from a US Navy engineering handbook. It makes no claim about the actual signature or performance of any specific aircraft or radar, and it describes no technique beyond what is already public domain.
U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.