A twelve-year-old roof can be finished. A twenty-eight-year-old one can be fine.
Steep-slope · single-family and small multifamily
Age is a prior, not a finding. It tells you when to start looking. What decides the roof is what can actually be observed about its condition — and how much of that observation is available without anyone getting on the roof.
Does my roof need replacing because of how old it is?
Age narrows the question; it never settles it. Ventilation, installation quality, slope, orientation and climate move a roof years in either direction, so a twelve-year-old roof can be finished while a twenty-eight-year-old one has life left. Weigh observed condition — surface, edges, the attic side, and repair history — against age. Treat age as a reason to look, never as a verdict.
The short versionSection link
Six things that are true before anyone looks at your roof. The middle two are the reason this page exists.
- What age is
- A prior, not a findingThe federal appraisal instruction that governs FHA-insured lending requires an effective age separately from the actual age, and states that it “may be greater than, less than, or equal to the actual age.”
- What the roof entry on that form asks for
- Material type and observed conditionNot a date. The instruction is to “enter the material type and rate the observed condition” and to “note any evidence of deterioration of roofing materials.”
- Where age does predict something
- Whether the shingles are still sealedAcross the 27 Florida roofs inside a 30-home Florida and Texas survey, every roof under six years old had under 1% of its shingle strips unsealed; 14 of 17 roofs older than six years had more than 1%.
- Where the same study says it does not
- The other three roofs, and every hip and ridgeThree of those seventeen older roofs stayed under 1%. Unsealing on hip and ridge shingles “showed no discernible relationship to roof age” at all.
- Why two slopes on one roof differ
- 623 thermal shocks a year against 39One roof at the IBHS aging farm in Richburg, South Carolina recorded a 25°F change within five minutes 623 times in a year on its south slope and 39 times on its north slope.
- What no page can settle
- The condition of the deckIt is under the covering. Nobody — including this page, a drone photograph, or a contractor standing in your driveway — knows what it looks like until something is opened.
This page says weigh condition over age. Here is when that is the wrong instruction.Section link
The position taken here is that observed condition outranks a date. That position has real limits, and in several situations following it will cost you money or leave you exposed.
Best when
- You are budgeting. A condition-led read tells you whether to save for two years or for ten; a birthday does not.
- You are buying a property and pricing the roof into an offer or a repair credit, where the seller’s “it’s only fifteen years old” is doing more work than it should.
- A contractor’s entire case for replacement is the age of the roof. That is the specific argument this page exists to make answerable.
- The roof is not leaking and nothing has failed. Condition evidence is at its most informative exactly when there is no emergency forcing the decision.
- You are deciding whether to pay for an independent assessment at all. Knowing which observations would change your mind is how you decide that.
Think twice if
- Water is coming in now. Diagnosis outranks the planning question entirely, and it belongs on the leak pages, not here.
- A lender, insurer, appraiser, or lease is asking for a remaining-life figure. Those parties want a named professional’s dated opinion on this specific roof, and a reader’s condition read will not satisfy them.
- Your policy ties roof coverage to the roof’s age — check whether yours does. Where a policy sets terms by a stated roof age, that clause is contractual and does not care how sound the roof looks. This page cannot tell you what any policy covers; read the document and take the question to your carrier or a licensed adviser.
- The covering is a discontinued product nobody can match. Then repairability, not condition, is deciding, and a technically repairable roof can still be practically unrepairable.
- There is a second layer over the original. A roof-over hides most of the evidence this page asks you to read, and the deck under it has not been seen since the day it was covered.
- A storm has just been through. Storm damage is assessed differently, on a different clock, and often with an adjuster involved.
- The house predates 1990 and reading the evidence would mean disturbing old felts, mastics, or insulation. EPA’s position is that “the only way to be sure whether a material contains asbestos is to have it tested by a qualified laboratory,” and that sampling should be done by a trained and accredited professional. Testing comes before disturbance.
What changes the answer
- Orientation. NPS advises that “particular attention should be given to any southern slope because year-round exposure to direct sun may cause it to break down first.”
- Colour. DOE’s Building America guidance notes that lighter-coloured asphalt shingles “tend to last longer as they absorb less heat than darker shingles and are less likely to be damaged by prolonged heat.”
- Installation quality. DOE lists high nails, low nails, overdriven, underdriven, angled, and inadequate nails as examples of improper nailing — one of the common errors in shingle installation, and invisible from the ground for a decade afterwards.
- Slope. DOE states that asphalt shingles “should only be installed on roof slopes of 2 in 12 or greater,” with double underlayment required from 2 in 12 up to 4 in 12. Shallower planes hold water and debris longer, and dry more slowly.
- What the attic is doing. NPS is blunt that “condensation on the underside of the roofing is undesirable and indicates improper ventilation,” and that “moisture will have an adverse effect on any roofing material.”
- Whether the covering sits on a second layer, and whether the deck was ever re-nailed or replaced under it.
- Repair history: how many, where, how close together, and whether the same detail keeps coming back.
- How long you will own it. A roof that has five uncertain years left is a different asset to someone selling next spring than to someone staying twenty years.
The evidence is real. It is just in eight different places, and only some of them are reachable.Section link
The reason age gets used as a proxy is not that it is good — it is that it is the one number everybody has. Condition evidence is better and more scattered, so it needs a map before it needs a checklist.
- The sun-facing slope. Where wear shows first. What you are looking for from the ground, with binoculars, is loss of the mineral surfacing — the granules that shield the asphalt from ultraviolet light — as patches and colour change rather than as an even fade. NPS puts it plainly: “particular attention should be given to any southern slope because year-round exposure to direct sun may cause it to break down first.”
- The shaded slope. The same roof, the same installation date, a materially different exposure. What appears here is usually moss, algae, or lichen rather than wear, because the surface stays wet longer. A shaded slope that still looks new is not evidence that the roof is young. It is evidence that the roof has two different clocks running on it.
- The gutter and the downspout outlet. Where anything the roof sheds ends up. Accumulated granules at the outlet are the classic ground-level signal — and the most over-read one. Loose granules shed early in a roof’s life are largely surplus from manufacture and handling. What matters is whether they are still arriving years later, and whether the slope above it shows bare patches down to the mat. There is no published quantity that converts a gutter’s worth of granules into a number of years, and this page will not invent one.
- The ridge, the eaves, the rakes, and every penetration. The details. This is where installation shows rather than age — and it is where the field evidence is sharpest. In a survey of thirty single-family homes, the quantity of unsealed shingles in the field of the roof increased significantly with roof age, while unsealing on the hip and ridge caps “showed no discernible relationship to roof age.” NPS is equally direct about the other detail: “failure of the flashing system is usually a major cause of roof deterioration.” Flashing is where roofs actually leak, and flashing does not fail on a schedule.
- The underside of the deck. Visible from an open attic hatch without entering the space. NPS lists what registers there: water penetration “should be readily apparent, usually as a damp spot or stain,” and “condensation on the underside of the roofing is undesirable and indicates improper ventilation.” Daylight through the deck, dark staining around penetrations, and rows of protruding fastener tips all live here too. It is the most informative view of a roof available without going onto it — and it is the one most people never take. Reaching a ceiling hatch safely still means proper stepladder or attic-stair access; if that is not something you can do steadily, it is an observation to ask a professional for rather than to improvise.
- The eave, from inside. Insulation pushed up against the underside of the deck closes the intake path while the exhaust keeps pulling. It is a recognised reason a roof can age faster than its neighbours, and baffles are the cheap correction — see the ventilation guide for what is and is not required.
- The ceiling plane inside the rooms. Stains, bubbling paint, and rusted nail heads in drywall. The dashed arrow in the drawing is the important part: water travels along the underside of the deck and across the top of the ceiling before it drops, so a stain is a starting point for an investigation and not a location. A ceiling stain says water got in. It does not say the roof is old.
- The file. Not on the building at all. Permit records, the original invoice, the warranty registration, and — most usefully — every repair invoice since. Recurrence is evidence, and it is the class of evidence readers most often have sitting in a drawer and never look at.
Why age became the proxy in the first place
Because it is cheap and it is available. Everyone can say how old a roof is; almost nobody can say what condition it is in without buying an assessment or borrowing someone’s ladder. So a knowable weak signal displaces an unknowable strong one, and both homeowners and the people selling to them settle into the habit.
The federal appraisal system does not make that trade. The FHA appraisal instruction requires an effective age alongside the actual age, states that the effective age “may be greater than, less than, or equal to the actual age,” and requires the appraiser to “note any significant difference between the actual and effective ages.” For the roof specifically, the instruction is to “enter the material type and rate the observed condition” and to “note any evidence of deterioration of roofing materials.” The word observed is carrying weight there, and so is the fact that no field on that form asks how old the roof is.
That is a lending programme’s procedure, not a building code and not a diagnostic method. But it is a federal institution with money at stake declining to use age as the measure, which is a reasonable place to start an argument with a contractor who is.
Age is a real signal. It is just much weaker than the way it gets used.Section link
There is a defensible version of the age argument, and it is worth stating properly before taking it apart — because the same studies that establish it also establish its limits.
The best-measured age effect in residential roofing is not wear at all. It is adhesion. Asphalt shingles carry a factory sealant strip that bonds each course to the one below once the sun warms it, and until that strip is sealed a shingle is a nailed flap. A University of Florida study assessed thirty single-family homes in Florida and Texas. Twenty-seven of them, all in Florida, were surveyed by hand: shingle edges lifted, and the length and location of any lost adhesion recorded. The other three, near Houston, were surveyed by other engineers who did not quantify how much of each roof was unsealed — so the unsealing figures below are figures about those twenty-seven Florida roofs.
The age relationship in that data is genuine and it is one-directional: the quantity of unsealed shingles in the field of the roof rose significantly with age, and the study concluded that the shingles had been fully sealed to begin with, because adhesive residue was left on both faces when they let go. The authors also report that these findings are consistent with post-hurricane assessments finding that shingle roofs with six or more years of weathering were damaged at a fifty per cent higher rate than newer ones.
Now read the same dataset for what it says about individual buildings. Three of the seventeen roofs older than six years did not cross the one per cent line at all. The hip and ridge caps — a different location on the same roofs — showed no discernible relationship to age at all. And the study’s own conclusion about the coverings that had held is the sentence this page is built on: shingles with fully adhered sealant strips “provide excellent long-term resistance to uplift loads.” The roof that stayed sealed performed, regardless of how many birthdays it had.
| Finding | Source | What it supports | What it does not support |
|---|---|---|---|
| Roofs under six years: under 1% of shingle strips unsealed. Older than six: 14 of 17 roofs above 1%. | University of Florida survey — the 27 Florida roofs inside a 30-home Florida and Texas assessment | That sealant adhesion degrades with time, and that a roof past its first several years is more likely to have unsealed shingles than one that is not. | Any individual roof. Three of those seventeen older roofs did not cross the 1% line, and the quantified sample is twenty-seven houses in one hot, humid state. |
| Unsealing on hip and ridge caps showed no discernible relationship to roof age. | Same survey | That the detail work on a roof is governed by how it was built, not by how long it has stood. | A conclusion that hip and ridge caps are fine. They were frequently unsealed — just not as a function of age. |
| Post-storm failures at hips, ridges, eaves and rakes “appear to be independent of the age of the roof.” | Same dissertation, reviewing post-storm damage reports | That a large class of real roof failures is an installation and design question that an age-based argument cannot reach. | Any claim about how common those failures are relative to field failures, which the reports do not quantify consistently. |
| One roof: 623 rapid temperature swings a year on the south slope, 39 on the north. | IBHS aging farm, Richburg, South Carolina | That two planes of one roof, installed on the same day by the same crew, accumulate materially different thermal exposure — the mechanism behind slopes ageing at different rates. | A rate of deterioration. It is a count of thermal events at one instrumented site, not a service-life model. |
| Lighter-coloured shingles “tend to last longer as they absorb less heat than darker shingles.” | DOE Building America Solution Center | That a specification choice made once, at purchase, shifts where a roof sits inside its planning range. | Any quantity of years. The guidance states a direction and no magnitude. |
| Improper nailing — high, low, overdriven, underdriven, angled, inadequate — is a common installation error. | DOE Building America Solution Center | That installation quality is a first-order variable, and that it is set on day one and invisible afterwards. | Any way for a reader to check it. It is under the covering, and it is one of the strongest arguments for paying someone competent at the time of installation. |
Read this table one item at a time
Roofs under six years: under 1% of shingle strips unsealed. Older than six: 14 of 17 roofs above 1%.
- Source
- University of Florida survey — the 27 Florida roofs inside a 30-home Florida and Texas assessment
- What it supports
- That sealant adhesion degrades with time, and that a roof past its first several years is more likely to have unsealed shingles than one that is not.
- What it does not support
- Any individual roof. Three of those seventeen older roofs did not cross the 1% line, and the quantified sample is twenty-seven houses in one hot, humid state.
Unsealing on hip and ridge caps showed no discernible relationship to roof age.
- Source
- Same survey
- What it supports
- That the detail work on a roof is governed by how it was built, not by how long it has stood.
- What it does not support
- A conclusion that hip and ridge caps are fine. They were frequently unsealed — just not as a function of age.
Post-storm failures at hips, ridges, eaves and rakes “appear to be independent of the age of the roof.”
- Source
- Same dissertation, reviewing post-storm damage reports
- What it supports
- That a large class of real roof failures is an installation and design question that an age-based argument cannot reach.
- What it does not support
- Any claim about how common those failures are relative to field failures, which the reports do not quantify consistently.
One roof: 623 rapid temperature swings a year on the south slope, 39 on the north.
- Source
- IBHS aging farm, Richburg, South Carolina
- What it supports
- That two planes of one roof, installed on the same day by the same crew, accumulate materially different thermal exposure — the mechanism behind slopes ageing at different rates.
- What it does not support
- A rate of deterioration. It is a count of thermal events at one instrumented site, not a service-life model.
Lighter-coloured shingles “tend to last longer as they absorb less heat than darker shingles.”
- Source
- DOE Building America Solution Center
- What it supports
- That a specification choice made once, at purchase, shifts where a roof sits inside its planning range.
- What it does not support
- Any quantity of years. The guidance states a direction and no magnitude.
Improper nailing — high, low, overdriven, underdriven, angled, inadequate — is a common installation error.
- Source
- DOE Building America Solution Center
- What it supports
- That installation quality is a first-order variable, and that it is set on day one and invisible afterwards.
- What it does not support
- Any way for a reader to check it. It is under the covering, and it is one of the strongest arguments for paying someone competent at the time of installation.
Read the last column down the page. Every row supports a direction and none supports a date. That is what it means to say age is a weak predictor: not that it tells you nothing, but that it tells you which way to lean and never how far.
Put the observation first, and let age modify itSection link
This is the site’s own framework, not a measured model and not a score. It exists to do one thing: stop the conversation starting with a number of years.
The columns are deliberately coarse, and that is not laziness. A precise column would need a published service-life range to measure against, and for asphalt shingles — the covering most of this page is about — no primary source states one at all; the service life page sets out what is published, what is not, and why a warranty term is not a substitute for it. So place the roof roughly, in the terms you actually have. The rows are what can actually be observed, ordered from least to most alarming. Find the row first. The row is the finding; the column only changes what it means.
| What can actually be observed | Recently installed — within the first several years | Well into service — replacement is already a plausible budget line | Age unknown, or older than any range you have for this covering |
|---|---|---|---|
| Nothing visible from the ground, nothing inside, no repair history | Consistent with a roof doing its job. The useful action is to date the absence: photograph each plane and file the date, so next year’s look has a baseline. | The interesting case, and the one this page exists for. Kind conditions, not exemption — but the roof is telling you something the calendar is not. Move to an annual look and put one written assessment on the record. | Age is not the missing information. Get an assessment that states what was examined and what was not, and start the clock from that report rather than from a guess about the install date. |
| Distress at one detail only — one flashing, one penetration, one valley — with a clean field | Usually workmanship or detailing rather than wear. Repair it, and ask what caused it, because the same cause may well be present at the details you cannot see. | Still a repair question first. A sound field with one bad detail is not an end-of-life roof — but how much field life remains decides how much that repair is worth. | Repair the detail, and use the fact that someone is already up there as the occasion to have every plane assessed properly. |
| Field distress on one slope only, and it is the sun-facing or lowest-slope one | Premature for the age, so look for the cause: a blocked eave, a dark covering, an installation defect, or an exposure much harsher than the other planes. | The normal ordering. The sun-facing slope reaching the end first does not put the whole roof there — but it usually leads, and a slope-only replacement becomes a real option. | Plane-by-plane evidence is more informative than a date would be. Have each plane assessed and reported separately, and price the options that follow from that. |
| Field distress across more than one slope — bare patches to the mat, widespread curling or clawing, tabs lifting in wind | Something beyond age is operating: ventilation, a second layer, a product problem, or installation. Worth paying for an independent assessment before anyone quotes a replacement. | Age and condition agree. The question moves from whether to scope and timing — how much deck is at risk, what the ventilation needs, and whether the work waits for a season. | Condition has already answered it. Knowing the install date would not change the decision, only the paperwork. |
| Interior or attic-side evidence: recurring stains, staining on the underside of the deck, condensation, a soft or sagging area | This is never an age finding. Water is entering, or the assembly is condensing. Diagnosis outranks the replacement question entirely and comes first. | Diagnosis still comes first. A late-range roof with active water entry may be a repair, or it may be a deck problem the covering has been hiding for years. | Stop reading age into it. A soft or sagging area is a structural question, and it is the one row on this table with an urgency of its own. |
Read this table one item at a time
Nothing visible from the ground, nothing inside, no repair history
- Recently installed — within the first several years
- Consistent with a roof doing its job. The useful action is to date the absence: photograph each plane and file the date, so next year’s look has a baseline.
- Well into service — replacement is already a plausible budget line
- The interesting case, and the one this page exists for. Kind conditions, not exemption — but the roof is telling you something the calendar is not. Move to an annual look and put one written assessment on the record.
- Age unknown, or older than any range you have for this covering
- Age is not the missing information. Get an assessment that states what was examined and what was not, and start the clock from that report rather than from a guess about the install date.
Distress at one detail only — one flashing, one penetration, one valley — with a clean field
- Recently installed — within the first several years
- Usually workmanship or detailing rather than wear. Repair it, and ask what caused it, because the same cause may well be present at the details you cannot see.
- Well into service — replacement is already a plausible budget line
- Still a repair question first. A sound field with one bad detail is not an end-of-life roof — but how much field life remains decides how much that repair is worth.
- Age unknown, or older than any range you have for this covering
- Repair the detail, and use the fact that someone is already up there as the occasion to have every plane assessed properly.
Field distress on one slope only, and it is the sun-facing or lowest-slope one
- Recently installed — within the first several years
- Premature for the age, so look for the cause: a blocked eave, a dark covering, an installation defect, or an exposure much harsher than the other planes.
- Well into service — replacement is already a plausible budget line
- The normal ordering. The sun-facing slope reaching the end first does not put the whole roof there — but it usually leads, and a slope-only replacement becomes a real option.
- Age unknown, or older than any range you have for this covering
- Plane-by-plane evidence is more informative than a date would be. Have each plane assessed and reported separately, and price the options that follow from that.
Field distress across more than one slope — bare patches to the mat, widespread curling or clawing, tabs lifting in wind
- Recently installed — within the first several years
- Something beyond age is operating: ventilation, a second layer, a product problem, or installation. Worth paying for an independent assessment before anyone quotes a replacement.
- Well into service — replacement is already a plausible budget line
- Age and condition agree. The question moves from whether to scope and timing — how much deck is at risk, what the ventilation needs, and whether the work waits for a season.
- Age unknown, or older than any range you have for this covering
- Condition has already answered it. Knowing the install date would not change the decision, only the paperwork.
Interior or attic-side evidence: recurring stains, staining on the underside of the deck, condensation, a soft or sagging area
- Recently installed — within the first several years
- This is never an age finding. Water is entering, or the assembly is condensing. Diagnosis outranks the replacement question entirely and comes first.
- Well into service — replacement is already a plausible budget line
- Diagnosis still comes first. A late-range roof with active water entry may be a repair, or it may be a deck problem the covering has been hiding for years.
- Age unknown, or older than any range you have for this covering
- Stop reading age into it. A soft or sagging area is a structural question, and it is the one row on this table with an urgency of its own.
This framework is original to Understanding Roofing. It is not derived from a dataset, it produces no score and no number of years, and it cannot see the deck. It applies to steep-slope residential coverings; low-slope assemblies are assessed differently. Treat it as a way to organise a conversation with a professional, not as a substitute for one.
Repair history is condition evidence, and it is the only kind you already ownSection link
Everything else on this page needs someone to look at the building. This does not. It needs someone to look in a folder.
A single repair is an event. Three repairs on one slope in five years is a pattern, and a pattern is a finding about the roof rather than about the weather. NPS puts the economics without attaching a figure to it: “maintenance costs tend to multiply once trouble starts. As the cost of labor escalates, repeated repairs could soon equal the cost of a new roof.” The brief adds the instruction that follows from that — problems “should be evaluated at the outset of any project to determine if the roof can be effectively patched, or if it should be completely replaced.”
What to pull out of the folder, in order of how much it tells you:
- Every roofing invoice, with dates and locations. Lay them out chronologically and mark each one on a rough sketch of the roof. Clustering is the signal. Three invoices at one chimney is a flashing problem being treated as three separate leaks.
- The interval between them. Repairs getting closer together is a reasonable signal that repair is buying less time each round — no dataset converts an interval into a decision, but it is arithmetic you can do at a kitchen table and take to a professional.
- What each repair said the cause was. Invoices that describe a location but never a cause are a sign that nobody diagnosed anything. A leak repaired where the stain was rather than where the water entered will come back, and why the stain is rarely under the entry point explains the mechanism.
- Whether any of it is still under warranty. A workmanship warranty on a repair that has failed again is leverage, and it expires quietly.
- The permit record. Often the only independent evidence of when the covering actually went on, and sometimes the only evidence of whether a second layer was added. Your local building department holds it.
- The original invoice or specification. Product, colour, and whether a tear-off or a roof-over was done. That last item changes what every other piece of evidence on this page means.
None of this requires a ladder, a drone, or a contractor. It is the cheapest condition evidence available and it is the piece most consistently skipped — including by professionals, who generally are not handed the folder.
What condition evidence cannot reach, and where this page stopsSection link
A page that argues for observation over age owes you a clear statement of what observation does not get to. This is that statement.
Three things nobody can observe from where you are standing
The deck. It is under the covering and under the underlayment. From the attic hatch you can see its underside where the attic is open, which is genuinely useful and is not the same thing as seeing the deck. Its condition is established at tear-off, which is why a written allowance and a unit price belong in the contract before the first shingle comes off — the case is made at deck damage and allowances.
Whether the shingles are still sealed. This is the best-evidenced age-related property in the whole subject, and it is checked by lifting shingle edges by hand on the roof. No photograph resolves it. If someone tells you a roof is or is not sealed, the follow-up question is how they know.
What is under a second layer. A roof-over conceals the original covering, the original deck, and any damage that existed when it was installed. It also changes what the surface above it is telling you, because the layer beneath affects how the top layer sits and how heat moves through the assembly.
And one thing this page deliberately will not do
It will not give you a threshold. There is no granule count, no percentage of curled shingles, and no number of repairs that converts into a verdict, because no published dataset supports one and the figures that circulate on roofing websites are assertions rather than measurements. What this page offers instead is an ordering: the observation first, the age second, and a qualified person’s dated opinion before money moves. If you want that opinion to be worth what you pay for it, know which kind of inspection you are buying before you book it.
What changes this on a real buildingSection link
Seven things that move a specific roof away from what its age suggests.
- Climate
The Insurance Institute for Business & Home Safety, which runs an outdoor roof aging farm at Richburg, South Carolina, is direct about the load: “roofs experience relentless exposure to ultraviolet radiation from the sun every day. The ultraviolet wavelength is known to degrade asphalt.” “Sunshine can drive roof temperatures over 200°F,” and “precipitation can drop roof temperatures as much as 60°F in 5 minutes.” Asphalt is thermoplastic: it softens hot and hardens cold, and it is the cycling rather than any single temperature that does the work. IBHS also notes that clay, concrete, and slate “change little with time” — so the whole age-versus-condition argument on this page is sharpest for asphalt and softest for stone.
Climate exposure is site-specific and orientation-specific. A figure measured at one research site in South Carolina describes what that instrument recorded there; it is not a coefficient for your address.- Moisture and ventilation
What the attic is doing to the underside of the deck is often a larger determinant than the weather on top of it. NPS: condensation on the underside “is undesirable and indicates improper ventilation,” and “moisture will have an adverse effect on any roofing material; a good roof stays dry inside and out.” A familiar version of it is intake buried under blown insulation at the eave while the ridge keeps exhausting — a condition that is usually cheap to fix, invisible from the street, and capable of taking years off a covering that was installed correctly. No source quantifies how many years, and this page does not.
There is no universal ventilation ratio and this page publishes none. Required net free area, how it is measured, and whether an assembly must be vented at all depend on the adopted code edition, local amendments, climate zone, and assembly type. Vented attics and correctly designed unvented assemblies are both legitimate — see the ventilation guide.- Slope and drainage
DOE’s Building America guidance states that asphalt shingles “should only be installed on roof slopes of 2 in 12 or greater,” and that from 2 in 12 up to 4 in 12 “a double underlayment application is required.” A low-slope plane sheds more slowly, holds debris longer, and gives wind-driven water more opportunity to work back under a course. Two planes of the same age on the same house can be at genuinely different points in their service life for this reason alone — see how pitch is measured and why it gates material choice.
- Maintenance
A roof that has been cleared, had its flashings looked at, and had small failures caught early is not the same asset as an identical roof that has been ignored, and the difference compounds. NRCA’s position, published alongside its homeowner guide, is that “a manufacturer’s warranty does not replace the need for periodic inspections and necessary maintenance and repairs.” NPS makes the same point about consequences: “maintenance costs tend to multiply once trouble starts.”
- Structural weight
The covering, the flashings, and the deck run on separate clocks, and the deck’s clock is the one you cannot read. A covering that looks tired above a sound deck is a straightforward replacement. A covering that looks acceptable above a deck that has been damp for years is a much larger job that will only declare itself at tear-off. This is why a deck allowance and a unit price belong in a contract before work starts rather than in a change order afterwards.
Nothing here is a structural determination. Sagging, deflection, soft spots, and questions about whether framing or sheathing can carry a load are answered by a licensed design professional or a qualified contractor who has looked at the building, working to the code the local authority has adopted.- Code and jurisdiction
Nothing on this page is law. There is no nationwide building code for site-built construction in the United States. Whether a roof may be repaired, recovered, or must be torn off; how many layers are permitted; what triggers an ice barrier or a ventilation requirement — all of that is set by whichever model code your state or local government adopted, as that government amended it, on its own effective date. The FHA appraisal instructions cited here are the procedure of a federal mortgage insurance programme, not a building requirement anywhere.
Confirm the adopted edition, its amendments, its effective date, and what is actually required on a permit with your authority having jurisdiction. A contractor’s summary of the rule is not the rule.- Access and site conditions
If the house predates 1990, the felts, mastics, and old insulation involved in reading this evidence may contain asbestos. EPA’s guidance is that “the only way to be sure whether a material contains asbestos is to have it tested by a qualified laboratory,” that testing is warranted where material is damaged or where planned work would disturb it, and that samples “should be taken by a properly trained and accredited asbestos professional.” An age is not determinative either way — which is, in a different register, exactly the argument this whole page is making.
The warranty is the one document that really does run on ageSection link
Which is precisely why it is a poor guide to whether the roof is finished.
- A warranty term is a remedy schedule, not a lifespan
The number on the wrapper describes what a manufacturer will do about a manufacturing defect and for how long. It is not a prediction about your building, and it was written by a company that has never seen it. Why a “30-year shingle” is a warranty term is the page that works through the arithmetic.
- Proration is an age clock, and it is the only one that is exact
Once a prorated period begins, the remedy falls on a schedule tied to the installation date regardless of how the roof is actually performing. A well-ventilated roof in a kind exposure and a cooking roof over a sealed-off soffit are worth the same percentage on the same day.
- Wear is normally excluded; defects are what is covered
The distinction that matters when you are reading condition evidence is that ordinary weathering is generally what a warranty excludes, not what it pays for. Evidence that a covering is aging is usually evidence that no claim exists. Evidence of a defect — a pattern, a batch, a failure inconsistent with the exposure — is a different conversation, and how roofing warranties actually work sets out what to read in the document.
- Prior repairs carry their own, shorter warranties
A workmanship warranty on a repair is a contract with the contractor who did it, on that contractor’s terms, for that contractor’s stated period. When a repair recurs, whether it is still covered is often the most valuable fact in the file — and it is a fact about a document, not about the roof.
Repairability
Repairability decides more replacement dates than condition does, and it is almost never discussed in those terms. A roof is repairable when a matching covering can still be obtained, when the detail that failed can be opened and closed again without disturbing half a slope, and when someone who works on that material is reachable. Fail any of those and a technically sound roof becomes a practical replacement.
NPS states the underlying economics without a figure attached: “maintenance costs tend to multiply once trouble starts. As the cost of labor escalates, repeated repairs could soon equal the cost of a new roof.” That is the sentence to hold in mind when a third repair is proposed at the same detail — and partial against full replacement is where the scope question gets worked through properly.
A warranty is a contract between a reader and whoever wrote it. What it covers, what voids it, whether it transfers, and how it is enforced are set by that document and by the law where the reader lives. Read the actual warranty for the product and the installer in front of you — not a summary of one, including this one.
Questions to ask an installerSection link
Every one of these forces a claim about your building rather than a claim about roofs in general. A contractor who has actually looked can answer them in specifics; one who is arguing from the calendar cannot.
Which slopes did you actually get onto or see, and which did you not?
A roof has planes with different exposures and different conditions. An assessment that describes “the roof” as one object either did not look plane by plane or is not reporting plane by plane. Both are worth knowing.
How does the sun-facing slope compare with the shaded one, and can you show me?
This is the single fastest test of whether someone looked. The two slopes of one roof are the built-in control group: same installer, same date, same product, different exposure. If they look the same, that is informative. If they do not, the difference tells you what exposure has been doing.
Are the shingles still sealed, and how did you check that?
Sealant adhesion is the property with the clearest published relationship to age, and it is checked by hand on the roof — not from a photograph and not from a drone. An answer that describes the method is worth far more than an answer that describes the conclusion.
What did you see on the underside of the deck, and how did you get to it?
Staining, daylight, condensation, and protruding fastener tips are the highest-value observations available on a roof that is not leaking yet. “There was no access” is a legitimate answer and should appear in writing. Silence is not.
What is your basis for saying this roof is at the end of its life — the condition you observed, or its age?
Ask it plainly, and expect a real answer — most contractors have one. Someone who can point to specific observed distress on specific planes has made a case you can weigh. A case that reduces to a number of years has not been made yet, which is a reason to ask for the findings rather than a reason to distrust the person.
If you were repairing this instead of replacing it, what exactly would you repair?
Forcing the repair scope to be described makes the replacement case testable. If the repair list is one flashing, replacement is a choice. If the repair list is most of the roof, it is not.
What did you not examine, and what would change your answer if you could see it?
Every assessment has boundaries. The FHA appraisal instruction says of its own procedure that “an inspection performed in accordance with these guidelines is visual and is not technically exhaustive.” A professional who states their limits the same way is more credible, not less.
Require these in writing
- A plane-by-plane condition description — each slope named, its orientation, and what was observed on it.
- The date and method of the assessment, and dated photographs of every finding relied on.
- Whether shingle sealant adhesion was checked, on which planes, and by what method.
- What was observed from inside the attic, or a written statement that there was no safe access and why.
- The condition of every flashing location individually — valleys, penetrations, headwalls, sidewalls, chimney, skylights.
- The repair option in full, priced and scoped, alongside the replacement option — not as an afterthought.
- A deck replacement allowance with a stated unit price, agreed before work starts.
- An explicit written list of what was not examined and why.
Misconceptions and failure modesSection link
Common misconceptions
Common belief
My roof is twenty years old, so it is due.
What is actually true
Twenty is a reason to look, and looking is cheap relative to replacing. It is not a finding. The same twenty years applied to a north-facing 8:12 slope with open soffit intake and to a south-facing 3:12 slope over a sealed-up eave produce two roofs in genuinely different condition, and the difference is observable.
Common belief
It is only twelve years old, so whatever is wrong must be a defect.
What is actually true
Possibly — and a pattern of failure inconsistent with the exposure is worth pursuing with the manufacturer. But installation and ventilation explain a great many young roofs in poor condition, and neither is something a manufacturer covers. DOE lists high, low, overdriven, underdriven, angled, and inadequate nailing as common shingle installation errors; none of them is visible from the ground, and all of them shorten a roof.
Common belief
If it is not leaking, it is fine.
What is actually true
A leak is a late signal, not an early one. The evidence that arrives before a leak — unsealed tabs lifting in wind, granule loss to the mat, staining on the underside of the deck — is what gives you a choice about when the work happens rather than having the weather choose for you.
Common belief
Granules in the gutter mean the roof is done.
What is actually true
They mean the roof is shedding, which every asphalt roof does from the day it is installed. What converts that into a service-life signal is the combination — still arriving years later, plus visible bare patches on the slope above, plus the covering’s age. No published figure turns a quantity of granules into a number of remaining years, and any page that gives you one has made it up.
Common belief
A drone survey can tell me how many years are left.
What is actually true
Aerial photography documents the surface well and is a genuinely useful record. It cannot test whether a shingle is still sealed, it cannot see the underside of the deck, and it cannot see the deck at all. Photographs establish what a plane looked like on a date. Remaining life is an opinion someone qualified signs.
Common belief
The roof looks the same as the neighbours’ and theirs was replaced, so mine is due.
What is actually true
Two houses built the same year by the same builder can have different orientations, different attic conditions, different crews on the day, and different colours. The neighbour’s replacement date is evidence about the neighbour’s roof and about their contractor’s sales process. It is not evidence about yours.
How it actually fails
- Sealant strips releasing while everything else looks intact
- The factory adhesive that bonds each course to the one below loses adhesion over time. The University of Florida survey found that partially unsealed shingles were “found on 8 of 10 three-tab shingle roofs and 11 of 17 laminate shingle roofs for a total of 19 of 27 surveyed roofs,” that the quantity increased significantly with roof age, and — importantly — that the shingles had been fully sealed first, because adhesive residue remained on both faces.What you can see: Almost nothing from the ground. Tabs that lift and drop in wind but sit flat when it is calm; a roof that starts losing pieces in storms it used to survive. This is the failure mode that most justifies paying someone to check by hand.
- Heat and ultraviolet embrittlement, then curling and clawing
- Repeated softening and hardening plus ultraviolet exposure stiffen the asphalt and shrink the mat. IBHS names granule loss, blistering, unsealing, curling, exposed fasteners, buckling, and staining as the distresses its aging farm tracks.What you can see: Curled edges or clawed centres across a whole plane rather than in one spot, appearing on the sun-facing slope first. Visible from the ground with binoculars, and the clearest ground-level end-of-life signal there is.
- Flashing failure on a covering that is nowhere near finished
- NPS: “failure of the flashing system is usually a major cause of roof deterioration,” from poor workmanship, thermal stress, or corrosion of the flashing or its fasteners. None of those mechanisms is keyed to the age of the covering around them.What you can see: Interior staining that recurs in the same place after every driving rain from one direction; a repair history clustered at one chimney, valley, or skylight. See flashing failures by location.
- The deck degrading under a covering that still looks acceptable
- Slow wetting from condensation or a long-tolerated small leak softens the sheathing without ever producing a dramatic symptom above it. The covering is doing its job; what is under it is not what it was.What you can see: Staining or punky decking visible from the attic hatch; a roof plane that reads as slightly uneven or wavy from the street. This one is frequently invisible until tear-off — which is the argument for a written allowance.
- The repair that keeps coming back
- A symptom is treated at the place the water appeared rather than at the place it entered, or a detail is repaired without correcting the reason it failed. The invoice history is the diagnostic record, and nobody reads it.What you can see: Three invoices in five years for the same slope. Repairs at increasing intervals of urgency. A stain that returns in the same spot after each repair has “held” for one season.
Sources and further readingSection link
Understanding Roofing / Published
Scope and limitations
- It cannot tell you how many years your roof has left.
- That is a dated opinion about one building, signed by someone qualified who has been on it — not something a page or a photograph produces.
- It cannot see your deck.
- Nothing can, until a section of covering is removed.
- Every statement here about deck condition is about how to recognise the possibility, not how to establish the fact.
- It publishes no service-life range for any material.
- The published figures, their sources, and the disagreements between them live on the roof service life page, and asphalt shingles have no primary-source range at all.
- It publishes no cost figure, and no threshold at which repair spending becomes replacement spending.
- That arithmetic depends on the scope of both options for one building.
- It publishes no ventilation ratio and no required net free area.
- Those depend on the adopted code edition, local amendments, climate zone, and whether the assembly is vented at all.
- The measured evidence it leans on hardest is narrow.
- The unsealing figures come from twenty-seven Florida roofs inside a thirty-home Florida and Texas assessment; the thermal-cycling counts come from one research site in South Carolina.
- Neither is a national sample, and neither describes your building.
- Almost everything here is about asphalt shingles, because that is where the measured aging evidence is.
- IBHS notes that clay, concrete, and slate change little with time, and the argument on this page is correspondingly weaker for them and for metal.
- It cannot tell you what your insurance policy does when a roof passes a stated age, whether a warranty claim exists, or what your jurisdiction requires.
- Those are contractual and legal questions for your carrier, the warranty document, and your authority having jurisdiction.
FHA Single Family Housing Appraisal Report and Data Delivery Guide
U.S. Department of Housing and Urban Development, Office of Single Family Housing / Issued 20 May 2024, effective 19 August 2024
That the appraisal report requires an effective age entered separately from the actual age; that “the effective age may be greater than, less than, or equal to the actual age”; that the appraiser must “note any significant difference between the actual and effective ages and explain in the ‘condition of property’ comments section”; that for Roof Surface the instruction is to “enter the material type and rate the observed condition” and to “note any evidence of deterioration of roofing materials (missing tiles, shingles, flashing)”; that the appraiser must “report the observation and condition of the attic and the roof structure”; and that “an inspection performed in accordance with these guidelines is visual and is not technically exhaustive.”
The procedure of a federal mortgage insurance programme, not a building code and not a diagnostic standard. It governs how an appraiser completes a form for FHA; it establishes nothing about what any roof requires, and it is cited here only as evidence that a federal institution with money at stake records observed condition rather than age. This was the current version of the guide when it was checked on the date above. Two things around it are moving: Mortgagee Letter 2025-18 (27 June 2025) rescinded the remaining-economic-life reporting requirement without disturbing any wording quoted here, and FHA is adopting the modernised Uniform Appraisal Dataset 3.6, under which the 1004/70 report form this guide is written around is replaced by a single restructured appraisal report. The effective-age instruction itself survives that change — it is Handbook 4000.1 policy, not a quirk of one form — but the field-by-field wording quoted here should be re-verified against whatever guide governs the new report.
Natural Weathering and Hazard Exposure
Insurance Institute for Business & Home Safety (IBHS), Richburg, South Carolina research facility
That “roofs experience relentless exposure to ultraviolet radiation from the sun every day. The ultraviolet wavelength is known to degrade asphalt”; that “sunshine can drive roof temperatures over 200°F” and “precipitation can drop roof temperatures as much as 60°F in 5 minutes”; that “in one year, the number of times a south facing roof slope in Richburg, SC observed a temperature change of at least 25°F in 5 minutes ranges from 74 times to 623 times”; that “comparing north and south slopes, one roof hut on the Richburg, SC farm observed a 25°F temperature change 623 times in a year on the south slope but only 39 times on the north slope”; that “once sealed, a typical asphalt shingle roof is best able to withstand the wind, and its resistance immediately starts to decline from there. After 7-10 years, the sealant weakens enough to significantly reduce the shingle’s ability to resist these pressures”; the aging distresses it tracks — granule loss and blistering, unsealing and curling, exposed fasteners, buckling, and staining; and that clay, concrete, and slate “change little with time.”
A research organisation funded by property insurers, describing its own outdoor aging farm. The temperature counts are instrument readings at one South Carolina site and are not a national figure or a coefficient for any address. The 7-to-10-year sealant statement is presented on the page without a citation to a specific study, so it is reported here as IBHS’s stated position rather than as a measured result, and it is not used on this page as a threshold.
The Wind Resistance of Asphalt Roofing Shingles
Craig Robert Dixon, doctoral dissertation, University of Florida (advisers David O. Prevatt, Forrest J. Masters, Kurtis R. Gurley) / December 2013
The field assessment of 30 single-family homes in Florida and Texas, twenty-seven of them surveyed and quantified in Florida; that partially unsealed shingles were “found on 8 of 10 three-tab shingle roofs and 11 of 17 laminate shingle roofs for a total of 19 of 27 surveyed roofs (70%)”; that “the percentage of unsealed shingles for all roofs less than six years old is less than 1%, while 14 of 17 roofs older than six years had more than 1% of their shingles partially or fully unsealed”; that “a statistically significant difference in the mean total quantity of unsealed shingles was established between roofs with ages 0-6 and those 7-13 and 14-20”; that “the quantity of unsealed hip and ridge shingles showed no discernible relationship to roof age”; that unsealed shingles had adhesive residue on both faces, “which indicates that the shingles were initially fully sealed”; that “new shingles generally appear to remain fully sealed for the first 4-5 years of service life”; that hip, ridge, eave and rake failures reported after storms “appear to be independent of the age of the roof and more closely tied to the design and installation of these edge conditions”; and that “the major conclusion is that shingles with fully adhered sealant strips provide excellent long-term resistance to uplift loads.”
University building-science research, examined and cited directly rather than through any secondary account of it. The quantified unsealing figures come from twenty-seven Florida roofs — the three Houston roofs in the thirty-home assessment were surveyed by other engineers who did not quantify them — so the sample is one hot, humid, hurricane-exposed state and is not national. Installation ages for twenty-three of the twenty-seven Florida roofs came from the homeowner or roofing permit records; four were unknown. It is a study of wind resistance, not of service life, and it says nothing about any individual building. Its Chapter 4 was submitted for publication in the Journal of Wind Engineering and Industrial Aerodynamics; that journal article is paywalled and was not read for this page.
Preservation Brief 4: Roofing for Historic Buildings
National Park Service, Technical Preservation Services (Sarah M. Sweetser) / February 1978
That “particular attention should be given to any southern slope because year-round exposure to direct sun may cause it to break down first”; that “failure of the flashing system is usually a major cause of roof deterioration” from poor workmanship, thermal stress, or deterioration of the flashing metal or its fasteners; that water penetration “should be readily apparent, usually as a damp spot or stain”; that “condensation on the underside of the roofing is undesirable and indicates improper ventilation”; that “moisture will have an adverse effect on any roofing material; a good roof stays dry inside and out”; that “periodic checking of the underside of the roof from the attic after a storm or winter freezing may give early warning of any leaks”; that problems “should be evaluated at the outset of any project to determine if the roof can be effectively patched, or if it should be completely replaced”; and that “maintenance costs tend to multiply once trouble starts. As the cost of labor escalates, repeated repairs could soon equal the cost of a new roof.”
Written in 1978 for historic buildings and their traditional materials. Its guidance is addressed to owners working with architects and preservation contractors, and its examples are slate, tile, wood, and sheet metal rather than modern asphalt shingles. It is not a code, not a diagnostic standard, and its inspection descriptions assume professional involvement.
Asphalt Shingle Roofs
U.S. Department of Energy, Building America Solution Center (Pacific Northwest National Laboratory)
That “lighter-colored asphalt shingles tend to last longer as they absorb less heat than darker shingles and are less likely to be damaged by prolonged heat”; that “one of the common errors in shingle installation is improper nailing” and that “high nails, low nails, overdriven nails, underdriven nails, angled nails, and inadequate nails are all examples of improper nailing”; that “roofing nails should be driven in straight and flush, not overdriven, underdriven, or angled”; and that “asphalt shingles should only be installed on roof slopes of 2 in 12 or greater” with “a double underlayment application” required from 2 in 12 up to 4 in 12.
Best-practice guidance for builders, not adopted law anywhere. Its code references are to specific model code editions that many jurisdictions have superseded, and its slope statement is a general practice rule — the instructions for a specific product govern that product.
Guidance for homeowners
National Roofing Contractors Association, Professional Roofing (Mark S. Graham) / October 2020
That “a manufacturer’s warranty does not replace the need for periodic inspections and necessary maintenance and repairs to a newly installed asphalt shingle roof system,” and that NRCA encourages homeowners who are not comfortable performing inspections and routine maintenance themselves to hire professional roofing contractors to do it.
An article in a trade association’s own magazine announcing its homeowner guide, not the guide itself. NRCA is a contractors’ association; it is a competent technical publisher and an interested party in how often homeowners engage contractors. The underlying 30-page guide is sold rather than published free and was not read for this page.
Residential Fall Protection — Guidance
U.S. Occupational Safety and Health Administration
That “falls are the leading cause of death for workers engaged in residential construction,” and that “workers engaged in residential construction six (6) feet or more above lower levels must be protected by conventional fall protection (in other words, guardrail systems, safety net systems, or personal fall arrest systems).”
An employer obligation for construction work. It is not homeowner guidance and a homeowner is not subject to it; it is cited here as evidence of how seriously the exposure is treated by people who do this for a living.
How do I know if I have asbestos in my home (floor tile, ceiling tile, shingles, siding, etc.)?
U.S. Environmental Protection Agency
That “the only way to be sure whether a material contains asbestos is to have it tested by a qualified laboratory,” that EPA recommends testing where material is damaged — fraying or crumbling — or where a planned renovation would disturb the suspect material, and that “samples should be taken by a properly trained and accredited asbestos professional (inspector).”
General consumer guidance. It does not establish that any particular material or building age contains asbestos, and it is not the regulatory requirement for any specific work; federal, state, and local rules govern testing, handling, and disposal.