When maintenance becomes the problem

During WWII, a study of RAF bombers overturned everything we thought we knew about preventive maintenance.

Most maintenance thinking rests on a single, comfortable assumption: machines fail because they get old, so replacing worn parts before they break must make equipment safer. It feels obvious. It is also, in many cases, wrong.

In the 1940s, the British biologist Conrad Hal Waddington was asked to improve the availability of Royal Air Force B-24 bombers. The aircraft were being pulled from service for scheduled overhauls on a fixed calendar. Waddington studied the failure records and found something no one expected: the planes broke down more often in the period immediately after maintenance, and then grew more reliable the longer they were left alone.

The maintenance meant to prevent failure was, in fact, introducing it.

The intervention itself — disassembling, adjusting and reassembling healthy machines — was injecting new faults. When Waddington's team extended the maintenance intervals and stopped touching equipment that showed no sign of trouble, the bombers flew roughly 60% more before service, with fewer breakdowns, not more. The phenomenon has been known ever since as the Waddington Effect.

~60% more flight time after intervals were extended

1940s — when the paradox was first documented

Fewer breakdowns, not more, with less intervention

Why intervention introduces failure

The reasons are as true for a cement mill or a pump today as they were for a wartime bomber. Every time a healthy machine is opened up, it is exposed to fresh risk: bolts torqued incorrectly, seals disturbed, contamination introduced, connectors left loose, new parts that fail early in their life. Reliability engineers call this "infant mortality" — the elevated failure rate that follows any invasive action on otherwise sound equipment.

Crucially, most failures are not driven by age at all. Decades of reliability studies have shown that only a minority of components wear out on a predictable, time-based schedule. The rest fail randomly, from causes calendar-based overhauls can neither predict nor prevent — which means replacing them "just in case" adds cost and risk while removing very little danger.

The answer isn't less care — it's better-timed care

None of this argues for neglecting machines. It argues for matching the maintenance strategy to how each asset actually fails. Some low-criticality parts are best simply run until they break. Others need scheduled attention. But for critical, expensive or hard-to-replace equipment, the smartest approach is neither run-to-failure nor calendar-based overhaul — it is to intervene only when the machine tells you something is genuinely changing.

That is exactly what condition monitoring makes possible. Instead of guessing at intervals, you watch the real health of the asset continuously and act on evidence, not the calendar. You leave healthy machines alone — avoiding the Waddington trap — and you move fast the moment a true fault begins to develop.

How Fizix turns the paradox into an advantage

Fizix smart multi-sensor devices watch each machine around the clock across four signal domains — 3-axis vibration, acoustic emission, magnetic flux and temperature — and run failure-related analytics on the device itself. Our digital-twin platform, Mate, turns that stream into a clear health picture: when nothing is developing, you do nothing; when a real fault appears, you get an early, specific warning and time to plan.

The result is maintenance that is both lighter and safer — fewer unnecessary strip-downs, fewer intervention-induced failures, and repairs made only when, and only where, the data says they are needed.

Leave healthy machines alone. Act the moment the data changes. That is the modern answer to Waddington's paradox.