Chase the Nuisance Fault or Let It Go: Decision Tree

Why this matters

Not every oddity is worth chasing to the end, and not every oddity is safe to leave alone. Keep hunting a harmless intermittent blip and you burn hours and customer patience on a problem that was never going to hurt anything. Wave off a fault that looks minor but is masking a real, worsening condition and you get the callback where it finally failed for real, at a worse time and a bigger cost. This tree gives you the order to work through so the decision is deliberate, not a shrug or an obsession.

Start here: rule out a safety-relevant fault first

Before deciding whether a fault is worth chasing on cost-benefit grounds, check whether it involves a genuine hazard category: gas, a pressure vessel, stored energy, a structural or life-safety system, or anything electrical combined with water. If the fault sits in any of these categories, this cost-benefit tree does not apply. Treat it as a real fault until proven otherwise, and follow the safety-first response for that hazard before any further diagnosis. The rest of this article is for faults where the worst-case outcome is inconvenience or a repeat visit, not injury.

Step 1: Confirm you have actually done the basic checks

"Let it go" is a legitimate decision, but only after the checks in the companion reference article (telling a nuisance fault from a real one) have actually been done: reproducibility, logged trip data, physical evidence, timing, and device history. Do not let a customer's impatience or a busy schedule push you into skipping the basic legwork and calling it "probably nothing."

If the basics have been done and the evidence leans real, stop here and chase it as a real fault using the normal diagnostic path for that fault type.

Step 2: Weigh the cost of being wrong in each direction

If the evidence genuinely leans nuisance, weigh what each side of the decision costs if you are wrong:

• Cost of chasing further and it turns out to be nothing: more diagnostic time, more customer patience spent, a bigger ticket for a problem that would have resolved on its own.
• Cost of letting it go and it turns out to be real: a callback, a worse failure later, a customer who now believes you missed something, and in some cases a system that keeps degrading while nobody is watching.

These costs are rarely symmetric. A fault on a low-consequence system (a comfort feature, a convenience alert) tips toward letting it go with monitoring. A fault on a system where failure means real damage, real safety exposure, or a total loss of function tips toward chasing it further even when the immediate evidence is thin.

Step 3: Check whether the fault is getting more frequent

A single unexplained trip, months ago, that has not recurred is a different animal from a trip that is happening more often over time.

• If the frequency is flat or the fault has not recurred since the one occurrence, monitoring is usually the proportionate response.
• If the frequency is increasing, even if each individual trip still looks like a minor nuisance, that trend line is itself evidence of something worsening. An increasing trend overrides a "looks harmless" read on any single occurrence.

Step 4: If letting it go, make it a monitored decision, not a forgotten one

"Let it go" should never mean "stop thinking about it." The proportionate middle path:

• Document what you found, what you checked, and why you are not chasing it further today.
• Tell the customer plainly what to watch for and when to call back (a specific symptom, not "let us know if it happens again," which nobody remembers to act on).
• If your tools support it, set a reminder to follow up or ask about it on the next scheduled visit.
• If the equipment or your business process supports a data logger or a longer observation window, that is a better use of effort than repeated site visits chasing an intermittent that has not shown you enough yet.

Step 5: If chasing further, set a stopping point in advance

The other failure mode is the opposite one: chasing an ambiguous nuisance fault indefinitely because "it is probably something" and nobody wants to be the one who calls it done. Before you start a deeper diagnostic effort, decide up front what would make you stop:

• A specific number of hours or visits with no finding.
• A specific test that, if it comes back clean, ends the chase for now.
• A specific threshold of low consequence and low frequency below which continuing to chase costs more than it is worth.

Naming the stopping point before you start keeps the decision rational instead of driven by whoever is most stubborn in the moment.

Quick recap

1. Any hazard-category fault (gas, pressure, stored energy, electrical plus water) skips this tree entirely; treat it as real and act on safety first.
2. Confirm the basic real-vs-nuisance checks were actually done before treating this as ambiguous.
3. Weigh the asymmetric cost of being wrong in each direction, and weight it by the consequence of the system involved.
4. Rising frequency overrides a "looks minor" read on any single occurrence.
5. If you let it go, document, tell the customer what to watch for, and set a follow-up. If you chase it, set a stopping point before you start.

References

• Trade-standard practice for intermittent-fault triage and cost-benefit diagnostic decisions
• Manufacturer guidance on protective-device trip logging and monitoring features (general practice)
• See related: The Nuisance Fault vs the Real Fault, Telling Them Apart; Why Some Faults Are Safe to Tolerate and Some Aren't