Methodology
How we build a case
Two kinds of pages carry a verdict on this site: a case file on a car generation and a rating on a viral repair claim. Both run on the same rule — if we can't corroborate it, it doesn't go in. Here is the process behind each.
How a car case file is built
A case file covers one generation of one car — not a model's whole lifetime, because a 2006 Civic and a 2013 Civic share a name and almost nothing else. Everything in the file is corroborated against that exact generation:
- The problem list. Every failure mode we publish is checked against NHTSA complaints and recalls, CarComplaints owner reports, manufacturer technical service bulletins, and established owner forums — for that generation, sometimes for a specific engine or year range within it. Each major problem lists the sources it was checked against.
- Fewer beats padded. If a generation has three well-documented problems, the file lists three. We do not stretch a list to look thorough.
- Cost ranges, not scare numbers. Repair costs are 2024–2026 US independent-shop estimates, parts and labor, given as ranges — because a timing-chain job does not cost the same in rural Ohio and San Jose. When we are less certain, the range gets wider, not narrower.
What the reliability score means
Each case file carries a score from 1 to 10, where 10 means the generation is close to bulletproof. It is an editorial verdict, not a lab measurement: we weigh the documented failure modes by how severe they are, what they cost, how early they show up, and how much of the generation they touch. A car with one cheap, well-understood weak point scores higher than a car with a transmission that grenades at 90,000 miles — even if the second one has fewer total complaints.
The same evidence drives the best years and years to avoid on each file: model years inside a generation where a known problem was introduced or fixed.
Part numbers
Where a case file lists parts — filters, pads, plugs, belts — each number is checked for that generation and engine against at least two independent sources: manufacturer catalogs, filter-brand cross-reference lookups, and retailer fitment pages. A number we could not confirm from two directions is left out rather than guessed. Some outbound part links go to Amazon and can earn us a commission; the price you pay does not change, and the part list is settled before any link is attached.
How we rate viral car-repair claims
Every claim page is structured the same way: the source, the claim verbatim or summarized, our verdict on an 8-tier scale, the underlying chemistry or mechanics, what we'd actually recommend, and an FAQ. Here is how we land on a verdict.
The 8-tier rating scale
We use a wider scale than the typical "True / False / Mixed" trio because most viral car claims are not actually false — they are misleading in a specific direction. The wider scale forces us to be precise about how a claim fails.
True — The claim, as stated, holds up against our shop-floor experience and the cited sources. Rare in this genre.
Mostly True — The core claim is accurate but the framing (price, "they don't want you to know," ban claim) is exaggerated.
Mixed — Some genuine truth, some misrepresentation. The viewer would walk away with a mixed-accuracy mental model.
Misleading — The underlying chemistry, mechanics, or product category exists, but the video's framing creates a false expectation of result.
Outdated — Was once partly true (e.g., 1970s flooded-cell batteries). No longer applies to modern vehicles or parts.
Mostly False — The product or technique exists but the claim of permanence, completeness, or banned-status is fabricated.
False — The claim does not hold up at all. Either the product, the mechanism, or the asserted result is fabricated.
Dangerous — Reserved for claims where following the advice can damage your engine, void a warranty, cause a fire, or injure the person attempting it. We do not use this tier loosely.
Source hierarchy
When we're checking a specific factual claim — a price, a ban, a chemistry assertion, a manufacturer cover-up — we lean on the following sources in this order:
- OEM service information — factory service manuals, technical service bulletins (TSBs), recall notices. Authoritative for what a manufacturer specifies for a specific vehicle.
- SAE technical papers and ASTM/ISO test standards. Authoritative for what a chemistry or mechanical claim has been measured to do under controlled conditions.
- Trade publications and certifying bodies — Motor Magazine, Automotive Engineering, ASE, I-CAR, NIASE. Useful for industry consensus.
- Manufacturer technical bulletins from the product itself — Liqui Moly, BG, Bar's Leaks, POR-15. Treat with appropriate skepticism (they are marketing) but their listed compatibility and warnings are usually accurate.
- Direct shop experience and Diagnostic Network-style professional forums. Used as a sanity check on the above.
We don't use random forum threads, AI-generated content farm articles, or YouTube comments as primary sources.
Why we cite the channel that made the claim
Each claim page identifies the channel and links the source video. This is editorial fact-checking, not anonymous opinion. You should be able to watch the claim yourself, verify we're representing it fairly, and form your own view. Our rating is our judgment; the source citation is so you can audit it.
Conflicts of interest, stated up front
We participate in affiliate programs — primarily Amazon Associates — for products we recommend in our claim pages. Two firm rules:
- We never recommend the snake-oil product being debunked just because it pays a commission. (We frequently rate the underlying product category honestly — sometimes "yes, this kind of product exists and is fine if used correctly," sometimes "no.")
- A claim's verdict is set before the product list is built. We do not soften a "Dangerous" verdict because a related product happens to be high-commission.
Corrections
If we get a fact wrong, we want to know. Our disclaimer and corrections page documents the contact route and our correction practice. Material corrections are noted at the top of the affected page with a date stamp.