The Myth That Runs Through the Whole System

It’s not just what pilots are taught. It’s what pilots quietly work around, what mechanics tolerate, what investigators assume, and what certification lets slide.

“Running out of fuel is the stupidest thing you can do.”

Every pilot agrees with that sentence. That’s exactly the problem.

Say it to the pilot who always tops off the tanks, never plans a leg over two hours, and gets twitchy the moment the needle dips anywhere near half — someone whose idea of a wild Saturday is a thorough preflight, who won’t cross an active runway without hearing the clearance repeated back a third time — and they’ll nod along. Obviously that line isn’t about them. It’s about someone else. Someone careless. Someone who rarely visibly confirms fuel in the tanks.

But “stupidest thing you can do” is a sentence about behavior, and behavior isn’t where this problem lives. It lives in an instrument that was never required to tell that same careful pilot the truth for most of the flight. You can do everything right, including the visual check, and still be reading a gauge you were told very early on was intentionally designed to be untrustworthy until it read empty.

That’s the part the stigma hides. “Only a stupid pilot runs out of fuel” isn’t just wrong about the pilots it blames — it’s wrong about what it lets off the hook. The belief that a fuel gauge is only required to be accurate at empty is a myth that runs through every layer of the system that’s supposed to catch a bad gauge before it fails to warn somebody about a sudden loss of power. Once you see how far it reaches, “trust the gauge less” stops looking like a solution and starts looking like the symptom.

Where the myth lives

Start with the FAA’s own Pilot’s Handbook of Aeronautical Knowledge, which states that certification rules require gauge accuracy “only when they read empty.” That’s not an accurate summary of 14 CFR §23.1337, which requires the indicator to show usable fuel during flight — a functional requirement across the whole flight, not a single checkpoint. The zero-fuel calibration point is additional, not a ceiling.

But the Handbook is just the most visible place the myth shows up. Trace it and it touches almost everyone in the chain.

Mechanics

 A wandering fuel gauge rarely grounds an airplane the way a clearly failed component like a magneto that requires repair or replacement, because the accepted shop standard for fuel quantity “good enough” has drifted from its basic requirement. Consider what is required for an airworthy fuel quantity instrument in an ordinary Piper Cherokee: remove the sender, build a jig, set up reference heights, read the gauge at each point, and confirm it is within tolerance. That’s not a rogue standard — it’s in the Piper Maintenance Manual. When I talk to Piper mechanics, I ask how many performed the mandatory Service Bulletin covering wing spar corrosion inspection behind the fuel tanks. Every hand goes up. Then I point to the same Service Bulletin’s fuel-system items — gaskets, perishable fuel lines, vent lines, and removal of the fuel quantity senders themselves for inspection — and ask who built the jig and performed the fuel quantity calibration check sitting right there in the maintenance manual. No hands. Not one. It’s worth noting there are several Piper-mandatory service bulletins touching fuel systems — this is far from the only one — yet the fuel quantity piece is the part that consistently gets skipped.

Here’s the real contrast. Fuel quantity indication rarely shows up in the Service Difficulty Reporting system at all. To the FAA, that silence reads as evidence the fleet is fine — no SDRs means no problem in the field, as far as the agency can tell. But that conclusion only holds if a broken gauge and a “compliant” one would look different in the data, and under the myth’s own logic they don’t. A sender that’s stuck, dead, or wandering meaninglessly still technically satisfies “accurate at empty” if it isn’t obviously reading full with the tank dry. Nobody files a difficulty report on a gauge that was never expected to do more than that in the first place. The absence of complaints isn’t proof of a healthy fleet. It’s proof the standard is too low to ever generate one.

Pilots

Look at what pilots buy: painted fuel sticks, dipsticks cut for specific tank profiles, and totalizers wired into the panel. An entire aftermarket exists because pilots assume the certified instrument will not do its job, so they need something else. None of these substitutes is the fuel gauge §91.205(b)(9) requires. None is an approved equivalent. They are workarounds, bought in volume by pilots who quietly stopped expecting the installed gauge to function. Markets that large do not grow around imaginary problems.

Here is the tell. Modern GPS moving maps will happily draw a range ring from the totalizer — how far the airplane can fly on the fuel it thinks remains. Pilots plan real diversions around that ring without a second thought. Yet it depends on an assumed starting quantity and fuel measured through a flowmeter: highly accurate only if the starting value is known, entered correctly, and there are no leaks. FAA guidance says so directly: AC 23-17C states that a fuel totalizer “does not sense the quantity of fuel in the tank and it should not be used as a fuel quantity indicator,” and that “fuel quantity indicators should be used as the primary fuel-remaining instruments.” That’s not a fringe opinion. That’s the agency’s own written policy, sitting unread in an engineering and manufacturing advisory circular while pilots plan fuel stops around a screen that was never supposed to be trusted that much. Meanwhile, the one instrument connected to the fuel quantity onboard gets only a sideways glance. The number that could be closest to the truth is the one nobody was taught to believe.

Training has become the default risk-control tool for what is, underneath it all, a very human problem. Fuel gets a pamphlet or video instead of a standard to bring fuel quantity systems to an airworthy state.

Investigators

Remembering that only stupid pilots run out of fuel therefore Fuel exhaustion is one of the few probable-cause findings that reads like a character judgment instead of a systems failure.  Nobody’s reputation suffers from “broken wrist pin”; that is just bad luck. “Ran the tanks dry” sounds like wanton carelessness, especially when the pilot is no longer there or willing to say what if anything the fuel gauge indicated.  When an engine stoppage’s exact cause is not cleanly instrumented, there is an incentive to write “powerplant failure” instead of the harsher finding of “fuel exhaustion.” Nobody must act in bad faith for that pattern to shape the accident record.

Rossier’s August 2026 Sport Aviation article puts a number on the blind spot: roughly half of light-aircraft engine failures trace back to fuel. Half. That is not a rounding error hiding inside “powerplant failure.” It is a category wearing a mechanical costume.

Here is how far procedure can outrun the obvious. An inspector walks a crash site with no fuel staining anywhere — not a drop, nothing but teaspoons of fuel in the tank, no fire. The engine still gets pulled, mounted on a stand, and run, because procedure does not skip steps just because reality is standing there with its arms crossed. There is nothing wrong with rigor. But there is something bleakly funny about a system this careful about confirming the obvious while not checking the instrument that should have warned the pilot fuel was running out.

If the engine runs cleanly, the case writes itself: bad pilot judgment, or powerplant failure for unknown causes. It’s a safe assumption that if a pilot was directed not to trust the fuel gauge, the workarounds relying on visual confirmation, fuel burn, and time in flight simply failed. You would think investigators would check whether the gauge itself gave a warning,  most don’t.  And herein lies the issue: if you were to use the NTSB reports to fix this accident cause, you’d end up reinforcing the workarounds through more rigorous pilot training, with nothing about fixing an instrument that should have warned the pilot about the impending powerplant failure.  And the reality is that is exactly what we get.

Here’s a detail worth sitting with. FAA guidance in AC 23-17C states that “ground and cruise attitude(s) are the minimum attitudes required for operation within the three percent tolerance” — meaning ordinary cruise flight sits squarely inside the range an airworthy fuel gauge should be guaranteed to be accurate in. That’s the one phase of flight where the certified indicator should be at its most trustworthy by actual regulation. And yet fuel exhaustion in cruise is the most prevalent of all fuel related accidents. Even where the gauge should do everything right, the surrounding culture had already trained the pilot not to trust it. An instrument pilots are told lies constantly cannot suddenly be expected to be believed the one time it tells the truth.

Certification

Even FAA Certification compliance is not immune.  Some engineers in the compliance space operate for decades on the belief that only the zero-fuel point genuinely matters, the paperwork shows it.   There are more than a dozen aircraft and modifications approved with no ability to measure all the fuel in the aircraft, which is a very clear requirement.    That engineering focused AC 23-17C is explicit that fuel quantity indicators are also governed by §23.1301, which “requires the installed indicators function as designed and not create a hazard in their operation” and “precludes indicators that read higher than the actual fuel level, since this would constitute a hazard.” That’s a real, separate regulatory backstop against exactly the failure mode this whole piece is about — an over-reading gauge — and it rarely gets mentioned alongside §23.1337 reference the Pilot Handbook alludes to.

As a supplier in this space, I’ve personally seen fuel quantity certification documentation that cited the Pilot’s Handbook as its basis for accuracy requirements — not the regulation, the Pilot’s Handbook. A student pilot training document, standing in for 14 CFR §23.1337 in a compliance record. That’s the myth showing up in the paperwork that’s supposed to prove an aircraft or aircraft modification meets the regulations.

Here’s the part worth sitting with. The manufacturers themselves clearly believe fuel quantity accuracy is a real, ongoing concern. Piper carries multiple mandatory service bulletins touching fuel systems, including one requiring periodic removal, inspection, and calibration verification of the fuel senders themselves. Cessna has its own: a mandatory service bulletin demanding an annual fuel gauge accuracy check across its single-engine fleet. Two competing manufacturers, independently, both decided this needed a recurring, mandatory maintenance action. And yet in that same span of decades, essentially one FAA Airworthiness Directive exists addressing fuel quantity indication accuracy across the entire piston GA fleet. The people building the airplanes think this is worth mandating and we see it in their requirements documents.   The agency regulating them has, in practice, treated it as a pilot training footnote instead.

Somewhere else, they already fixed this

It’s worth knowing this isn’t a universal problem. Australia’s civil aviation maintenance order, CAO 100.5, requires periodic accuracy testing of fuel quantity gauges on aircraft that rely on float-type sensing or lack a self-test function — checked “at all major graduations,” not just at empty, against a defined tolerance (no more than roughly 5% of nominal tank capacity), at intervals not exceeding 48 months. Where the error exceeds that tolerance, a corrective placard is mandatory. Modern self-testing digital systems are carved out of the recurring check, on the reasoning that continuous self-diagnosis already does the job — a fail-evident exemption, not a loophole.

Australia’s separate pilot guidance, AC 91-15, is just as direct, and for the exact reason this article has been making: “Placing sole reliance on a fuel quantity gauge to assess fuel quantity and not cross-checking fuel quantity information from a second source, exposes the pilot in command to the risk of being unable to determine actual fuel remaining should the fuel quantity gauge indication become faulty.”

No “only accurate at empty.” No myth to untangle. Just a straightforward instruction to cross-check, because the regulator assumes the gauge is supposed to be trustworthy across the range in the first place.

Two English-speaking, FAA-adjacent aviation authorities, looking at the same instrument, reached opposite cultures. One mandates full-range testing and tells pilots plainly not to trust a single source. The other lets a Handbook line about “only at empty” stand unexamined for decades.  The difference isn’t the technology. It’s the standard.

Worth being precise about what “the same instrument” actually means here: Australia’s general aviation fleet is overwhelmingly American-built — Cessnas, Pipers, Beechcraft, Cirrus, the same type-certificated airframes with the same fuel senders and the same gauges flying in the United States. This isn’t a story about different hardware getting different results. It’s the identical instrument, manufactured by the identical companies, subject to two entirely different regulatory expectations depending on which side of the Pacific the airplane happens to be registered.

Belief, not fact

Say it out loud and the absurdity announces itself: the FAA’s own training manual says a required flight instrument is accurate at a flight-test-derived point and nowhere else. In this context, “empty” means the most adverse flight condition in which the engine first shows a fuel-related interruption. That matters for certification, but it does not give the pilot useful information during any part of a normal flight. It is what happens when one sentence is pulled out of a regulation that was meant to be read in whole.

Follow the myth to its own logical end and it collapses on itself. If a gauge only had to be right at empty, a simple binary low-fuel warning light — on or off, nothing in between — would satisfy the requirement just as well as a full graduated instrument, since both would only need to be correct at the one moment that supposedly counts. Nobody would accept that. A light isn’t a gauge, and 14 CFR §91.205(b)(9) requires a fuel gauge indicating the quantity of fuel in each tank, not a threshold indicator that only speaks up once. §91.213(d) closes the loop further: the fuel gauge cannot even be placarded inoperative and deferred the way some other equipment legally can.  It has to work, continuously, as installed. The myth’s own logic would excuse a device the regulation doesn’t even recognize as adequate — which is as good a sign as any that the logic was broken from the start.

No one would say this about other aircraft instruments. Imagine claiming that an altimeter is accurate only at a single flight test point near sea level and that the rest is the pilot’s problem. Said plainly, it sounds like a joke. And yet this fuel-gauge myth gets repeated, generation after generation, down to the CFI teaching a seventeen-year-old on a first cross-country or even introductory flight — not because anyone checked it, but because it is aircraft tribal knowledge.

That is not how aviation training is supposed to work. Aviation runs on things you can verify. This one line runs on faith in the unseen — accepted because it has always been accepted. That is exactly how a wrong reading survives for so long inside a profession built on precision.

I don’t say that as a guess. This exact passage has been raised directly, more than once, with FAA and aviation alphabet-organization personnel from working engineers up to people much further up the chain — and it hasn’t moved. Other sections of the Pilot’s Handbook get revised. This one sits outside the process entirely. That’s not what a documentation error looks like. That’s what an unquestioned belief looks like.

The fuel system is the most human-touched system on the airplane

Compare fuel to ignition, because avgas combustion needs both equally. A perfect spark does nothing without fuel, and a full tank does nothing without a spark. Neither is the junior partner in making power. Yet a bad magneto grounds the airplane — full stop, no argument. Only fools and madmen fly with a known-dead magneto. A fuel gauge that has wandered for months routinely flies on.  I have personally flown aircraft with fuel gauges that resembled windshield wipers more than aircraft instrumentation and thought nothing of it. Same combustion process, same two ingredients, wildly different tolerance for a known fault.

A bad magneto cannot hide. It shows up as an RPM drop on runup — a direct functional check every pilot performs before takeoff, while there is still nothing but taxiway ahead. Fuel is checked too. Every flight manual expects the pilot to confirm quantity, caps, vents, drains, and leaks during the walkaround. But that check happens on the ground, before the variables begin: fuel burn, tank selection, switching discipline, route changes, uncoordinated flight, leaks, venting problems, or a gauge that wanders once the airplane is moving. Ignition gets a functional cross-check before takeoff, and redundancy keeps the airplane in powered flight if one side fails. Fuel gets a visual preflight check, then depends on the pilot repeatedly doing the right thing in flight, the aircraft not finding a way to deliver fuel to the atmosphere, and an instrument the culture has trained pilots not to trust. That is a big difference.

Every instrument owes you a warning

There is a design principle already built into the certification standards that does not get enough attention: an instrument’s job is not just to display a number; it is to warn you. Section 23.1309(b) requires warning information to alert the crew to unsafe system operating conditions. That is not fuel-specific. It is the general expectation for every instrument on the panel.

If we wanted fuel to behave more like spark on our aircraft, the answer is not mysterious. It already exists: compare an accurate fuel-level measurement with fuel burn. One source tells the pilot what is physically onboard; the other estimates what should remain based on flow and time. When they agree, the pilot gains confidence. When they diverge, the airplane has identified a warning condition before it becomes an engine problem. It actually insures that the fuel quantity system maintains its accuracy through the life of the aircraft.  Frankly  it is the cross check that allows digital fuel quantity systems to have a pass in Australia. That is the fuel-system equivalent of dual ignition: not blind faith in one source, but meaningful redundancy and a functional cross-check.

The fix isn’t more distrust. It’s less myth.

None of this is a license to skip the preflight check, ignore burn rate, or stop watching the clock. Good airmanship never goes out of style. But “only a stupid pilot runs out of fuel” was never the whole story — it’s the version that let the training material, the shop floor, the accident record, and the certification process all quietly agree not to look too closely at the instrument itself. The regulation always asked for more than that. It’s time the rest of the aviation world caught up to it — and, per Australia, some of it already has.

Scott Philiben
Founder CiES Inc

Sources

– 14 CFR §23.1337(b), (b)(1) — Fuel quantity indicator

– 14 CFR §23.1309(a)(1), (b) — Equipment, systems, and installations

– 14 CFR §23.959(a) — Unusable fuel supply

– 14 CFR §23.969 — “Fuel tank expansion space

– 14 CFR §23.1301 — Function and installation

– 14 CFR §91.205(b)(9) — Required instruments and equipment, VFR day

– 14 CFR §91.213(d) — Inoperative instruments and equipment

– AC 23-17C, “Systems and Equipment Guide for Certification of Part 23 Airplanes and Airships,” Federal Aviation Administration, February 23, 2004 (source of the TSO-C55a three-percent-of-full-scale tolerance figure, the ground/cruise-attitude basis for that tolerance, the §23.1301 hazard/over-reading guidance, and the fuel totalizer guidance)

– FAA-H-8083-25C, Pilot’s Handbook of Aeronautical Knowledge, Chapter 7 (verified primary-source text of the “only accurate at empty” fuel gauge passage; correction confirmed incorporated into a revision work copy, August 2026, per FAA Testing Standards Section correspondence)

– FAA Airworthiness Directive, glass-panel EFIS fuel quantity indication error, various affected STC installations (87 FR 8377, 2022)

– 1998 FAA civil penalty action against a flight school for operating with a known-defective fuel gauge, as reported by AOPA Pilot Protection Services (Kathy Yodice)

– Robert N. Rossier, “Fuel System Considerations,” Sport Aviation (source of the approximately 50% figure cited for light aircraft engine failures related to fuel issues; sourcing methodology for this figure has not been independently confirmed)

– Piper Service Bulletin No. 1006A (corrosion inspection of main spar behind fuel tanks, including fuel quantity sender removal and inspection) and Piper Mandatory Service Bulletin No. 1266 (Dec. 16, 2014) with corresponding FAA Airworthiness Directive 2015-06414 (fuel starvation warning placard)

– Cessna Mandatory Service Bulletin SEB99-18R1 (1999), requiring an annual fuel quantity gauge accuracy check on affected single-engine models — cited from a secondary forum reference; primary bulletin text not independently pulled for this piece

– Civil Aviation Order 100.5 (General requirements in respect of maintenance of Australian aircraft) 2011, Appendix 1, Clause 6, Civil Aviation Safety Authority (Australia) — fuel quantity gauge accuracy testing requirements

– AC 91-15 v1.1, “Guidelines for Aircraft Fuel Requirements,” Civil Aviation Safety Authority (Australia), September 2021 — pilot cross-checking guidance

– The claim that essentially one FAA airworthiness directive addresses fuel quantity indication accuracy across the piston GA fleet reflects the author’s search of Federal Register and FAA AD records during research for this piece, not an exhaustive or agency-confirmed count

– The claim that fuel exhaustion in cruise flight is the most prevalent phase for fuel-related accidents is based on the author’s own review of NTSB CAROL accident data filtered to fuel-related defining events.

– The characterization of investigator classification incentives is presented as a plausible institutional dynamic, not a documented or proven finding