For homeowners, property owners, and buyers in hail country

A Class 4 roof is a laboratory result, not a hail-proof roof.

Steep-slope and residential low-slope roofs · single-family and small multifamily

The rating is real, the test is well documented, and the label still does not mean what the sales sheet implies. Here is what each test actually does, and where the reader's own judgment has to take over.

30-second answer

Does an impact-resistant roof stop hail damage, and how should I think about hail where I live?

Hail damage is decided by what a storm does to a specific assembly, not by a label. UL 2218 Class 4 means a two-inch steel ball, dropped in a laboratory onto new material, left no crack on the back. Real hailstones are softer, irregular, and land on aged roofs. Choose impact resistance to improve the odds and possibly the premium, never as immunity.

Stop — immediate hazard

During hail, get inside and stay away from glass. After hail, stay off the roof.

Hailstones large enough to damage a roof are also large enough to injure a person. Shelter indoors on the side of the building away from the wind, keep clear of windows and skylights, and do not go outside to move a vehicle while the storm is on.

Afterwards, the temptation is to climb up and look. Do not. A roof that has just been hailed is covered in loose granules on a slope, and the damage this page is about is not visible from up there anyway — the diagnostic marks are on the underside of the shingle, in the metal at the edges, and in the pattern across elevations. Everything a homeowner usefully contributes can be done from the ground, from a window, or from a photograph.

If a power line is down or arcing, treat it as live, keep everyone well away from it and from anything it is touching — fences, trees, puddles, and vehicles included — and call your electric utility or 911. If a skylight, window, or ceiling has been breached and glass or water is coming into a room, close the room off. If there is smoke, arcing, or a burning smell, leave the building and call 911.

Learning paths and saved lessons

At a glance

The short versionSection link

The first four figures below are laboratory constants, the next two are definitions the trade and the weather service use, and the last two are insurance structures. None of them is a prediction about a specific building, a specific storm, or a specific claim.

What Class 4 means: A 2.00 in projectile, on new material, in a laboratoryUL 2218 uses steel balls of 1.25, 1.50, 1.75 and 2.00 in for Classes 1 to 4. FM 4473 uses freezer ice balls of the same four diameters. Both are graded on new specimens.
What counts as a UL 2218 failure: A crack or tear visible on the BACK of the specimenIBHS: crushed or dislodged granules, dents, and openings visible only on the top of the shingle are not failures under UL 2218 performance criteria.
FM 4473 Class 4 energy: 26.81 ft-lb (36.4 J) from a 2 in, 62.9 g ice ballFM Approvals' published target kinetic energy for the largest class. Classes 1 to 3 are 3.72, 7.77 and 14.95 ft-lb (5.0, 10.4 and 20.3 J).
What the tested standard says about age: Weathering can significantly lower hail resistanceFM 4473, in its own scope: exposure of roof coverings to the elements over an extended period of time has a potential to significantly lower the hail resistance of the roof materials, and the standard is intended for testing new roof covers.
Severe hail, as the NWS defines it: 1 in diameter — quarter size — or largerSPC: significant severe hail is 2 in or larger. The 1 in threshold replaced a 3/4 in threshold on 5 January 2010, which puts a step in the long-term report record.
Functional vs cosmetic, in trade terms: Water-shedding reduced, or service life shortenedNRCA's published guidance: a roof system is not functionally damaged by hail unless its water-shedding capability is diminished or its expected service life is reduced. Granule loss alone is not treated as functional damage.
Percentage deductible: A share of the dwelling limit, not a flat dollar figureMontana's insurance regulator's illustration: a 5% hail deductible on a $325,000 dwelling limit is $16,250 before the insurer pays anything.
Actual cash value on an older roof: Can settle at zero after the deductibleTexas Department of Insurance worked example: a $10,000 roof, $4,000 deductible, actual cash value coverage — a 20-year-old roof settles at $0.

Tradeoffs

This page's advice — buy impact resistance on the odds, not the promise — and where that advice is wrongSection link

The position here is that an impact-rated covering is worth money in a hail-frequent place because it shifts the distribution of outcomes, and that it is worth nothing as a guarantee. There are buildings and situations where that framing does not hold.

Best when

You are replacing the covering anyway, so the impact-rated version is a marginal upgrade cost rather than a whole new project.
Your insurer offers a documented premium credit for a classified roof covering, and you have read what the credit is attached to.
The building sits where large hail is not rare — not merely where hail is reported, which is a different map.
The alternative is a thin, light, basic three-tab covering, where the published damage threshold is lowest.
You intend to hold the building long enough for a reduced-damage outcome in one storm to matter.
The whole exterior is being considered together — an impact-rated covering over soft aluminium fascia, vents, gutters, and a skylight is only part of what hail marks.

Think twice if

The pitch is that the roof will not be damaged. That claim is not supported by any of the three test standards on this page, and IBHS's own testing of Class 4-rated products found failures below Class 4 impacts.
The upgrade is being sold as a way to get a claim paid. Nothing about a classification decides coverage, and in at least one state the same classification is what makes a cosmetic-damage exclusion attachable to the policy.
The building is in a place where hail is reported often but rarely large. Frequent small hail and rare large hail call for different spending.
The roof is a low-slope membrane or a tile roof. UL 2218 grades prepared roof coverings and FM 4473's stated scope is new rigid roofing; the failure modes and the repair economics are different, and so is the conversation.
The covering is being chosen for hail alone. Wind uplift, fire classification, ventilation, weight, and repairability all still apply, and a covering chosen on one axis usually loses on another.
Somebody knocked on the door after a storm with a free inspection and a contract. Urgency after a hail event is a sales technique before it is a diagnosis.
You are within a year or two of selling. The improvement is real and the payback horizon is long.

What changes the answer

The covering type and its published damage threshold: a lightweight three-tab and a heavyweight laminate do not fail at the same hail size.
The age and condition of the existing covering — every one of these tests grades new material, and the standards say so.
How big the hail actually gets where the building is, which is a different question from how often hail is reported there.
Whether your policy carries a flat deductible or a percentage of the dwelling limit for wind and hail.
Whether your policy settles the roof at replacement cost or actual cash value, and whether an endorsement changes that at a certain roof age.
Whether any cosmetic-damage exclusion is attached, and what it defines as cosmetic.
What the adopted code where the building sits requires on a re-cover or replacement, which is a question for the authority having jurisdiction and not for this page.
Whether a competent installer is available. A Class 4 shingle installed with a nail line that misses the reinforcement is a Class 4 shingle in name only.

The mechanism

The test reads the back of the shingle. The roof only shows you the front.Section link

This one geometric fact explains most of what is confusing about hail: the laboratory grades a surface that, on a finished roof, nobody can see.

One asphalt shingle in section, greatly magnified, lying over underlayment and deck with its head lapped under the course above. The dashed line along the underside is where UL 2218 decides pass or fail. Numbers key to the list below. Schematic, not to scale, and not an inspection procedure.Original diagram, Understanding Roofing.

Granules are knocked loose. The granules are the mineral surfacing that protects the asphalt from ultraviolet light. An impact scatters them, and that is the part a person can see from above. NRCA’s published guidance is blunt about what it means on its own: granule loss is an ongoing process that begins at manufacture and is considered normal weathering, and it is not, by itself, treated as functional damage.
The asphalt and mat below it are dented. The result is the soft spot roofers call a bruise — NRCA’s description is that it may not be obvious on the top surface but can be felt with the fingertips as a localised soft spot, like a bruise on an apple.
The fracture starts underneath. This is the sentence to carry away. Per NRCA, impact-caused fractures start on a shingle’s underside — the region of maximum tensile strain — and propagate toward the granule surface. Some are visible on top. Some are not. All of them are detectable on the underside.
That underside is the pass/fail line. Under UL 2218, IBHS records any evidence of opening — tearing, fracturing, cracking, or rupturing — on the back of the shingle as a test failure, and the absence of an opening visible on the back as a pass. Crushed or dislodged granules, dents, and openings visible only on the top are not failures under those criteria.
On a roof, that view does not exist. The shingle is nailed to the deck and lapped under the course above. IBHS states the consequence directly: claims adjusters will rarely be able to view the back of shingles after a hail event without causing more damage. The laboratory turns the specimen over. The person standing on your roof cannot.
Underneath sits everything that keeps water out. A steep-slope covering sheds water; the underlayment and the deck are the layers behind it. This is why a fractured shingle is a service-life problem long before it is a leak, and why hail damage is usually argued about rather than observed dripping.

Why this makes hail claims contentious rather than obscure

Put those six points together and the dispute structure falls out of the physics. The condition that defines failure in the laboratory is on a surface no one can inspect in the field. The condition that is easiest to see in the field — missing granules — is explicitly not a failure in the laboratory and is explicitly not functional damage in NRCA’s protocol. So two competent people can look at the same roof, describe the same marks, and disagree about what they mean, without either of them being dishonest.

IBHS said the same thing about its own results in 2014: performance criteria used in that assessment do not necessarily reflect the performance criteria used by insurance companies in deciding whether to repair or replace a shingle roof after a hail event. Three different questions — did the product pass a standard, is the roof functionally damaged, and does the policy pay — have three different answers, and a sales conversation that collapses them into one is the thing to watch for.

What is actually hitting the roof

Hailstones are not steel balls and they are not laboratory ice spheres either. FM Approvals says so inside its own test standard: hailstones are variable in properties such as shape, density and frangibility, and those properties affect the duration and magnitude of the impulsive force acting on the roof and the area over which the impulse is distributed. It adds that ice balls are generally harder and denser than hailstones, so an ice ball simulates the worst-case hailstone, and that the major difference between the two is that hailstones are more variable.

Direction matters less than people assume, and differently than they assume. NRCA’s guidance is that hailstones fall in random patterns with a predominant fall direction influenced by wind, but that wind generally only minorly contributes to their impact forces. The practical consequence of wind is therefore mostly about which slopes and which walls get struck square, not about turning a modest stone into a heavy one. That is also why damage assessment is done elevation by elevation rather than as one verdict for a whole roof.

And size distribution matters more than any single storm story. NRCA states the obvious thing that gets lost in a bad week: small hailstones are much more common than large hailstones. Most hail events do nothing measurable to a sound roof. The tail of the distribution is what a covering decision is actually about.

The central comparison

Three impact tests, and what a pass in each one actually provesSection link

This is the table the page exists to publish. Read the last two columns first: what the pass proves is narrow, and what it does not prove is where every misunderstanding lives.

UL 2218, FM 4473, and the IBHS impact protocol compared by projectile, delivery, failure criterion, and the inference each one supports. Compiled from FM Approvals' published standard and from IBHS's published summaries of its own testing.
Test	Projectile	How it is delivered	What counts as failure	What a pass proves	What a pass does not prove
UL 2218 — steel ball	Steel balls: 1.25 in (Class 1), 1.50 in (Class 2), 1.75 in (Class 3), 2.00 in (Class 4).	Dropped from the height needed to give the ball the kinetic energy a hailstone of the same size would have in a thunderstorm. IBHS impacts each target location twice, at edges, corners, joints, and centres, on panels built to the manufacturer's instructions.	Any tearing, fracturing, cracking, or rupturing visible on the BACK of the specimen. Crushed or dislodged granules, dents, and openings visible only on the top are not failures.	That a new specimen of this product, laboratory-conditioned, took that energy from a uniform hard sphere without opening at the back.	Nothing about aged material, irregular or softer stones, oblique impacts, or the rest of the assembly — and IBHS says the criteria do not necessarily reflect what insurers use to decide repair or replacement.
FM 4473 — freezer ice ball	Ice balls cast from distilled water: 1¼ in / 15.3 g, 1½ in / 26.5 g, 1¾ in / 42.1 g, 2 in / 62.9 g for Classes 1 to 4.	Propelled to a target kinetic energy — 3.72, 7.77, 14.95 and 26.81 ft-lb (5.0, 10.4, 20.3 and 36.4 J) — perpendicular to the panel within 90° ± 5°, twice at each location, with impact points chosen at edges, corners, unsupported areas, overlaps and joints.	Visible cracking or breakage of the specimen.	That new rigid roofing — the standard's stated scope is tiles or planks of slate, concrete or clay — survived an energy-matched ice impact at that class.	FM says so itself: ice balls are harder and denser than hailstones and simulate a worst case, hailstones are far more variable, and weathering has the potential to significantly lower hail resistance. FM also states that meeting the standard does not by itself qualify a product for Approval.
IBHS Impact Resistance Test Protocol	Laboratory-manufactured hailstones, 1.5 in and 2.0 in, made to mimic natural hail in density, strength, hardness, mass, kinetic energy and terminal velocity.	Propelled under controlled conditions at shingles bought from retail shelves rather than supplied by manufacturers.	Not pass/fail. Products are scored across three damage modes — dents (deformation), tears (breaches), and granule loss — and rated Excellent, Good, Marginal, or Poor.	Relative performance against a more realistic projectile, judged on the damage modes people actually argue about after a storm.	Still a laboratory comparison on new product. IBHS's own caveat is that the tests do not exactly replicate natural hailstone impacts and provide a way to compare relative product performance.

Read this table one item at a time

UL 2218 — steel ball

Projectile: Steel balls: 1.25 in (Class 1), 1.50 in (Class 2), 1.75 in (Class 3), 2.00 in (Class 4).
How it is delivered: Dropped from the height needed to give the ball the kinetic energy a hailstone of the same size would have in a thunderstorm. IBHS impacts each target location twice, at edges, corners, joints, and centres, on panels built to the manufacturer's instructions.
What counts as failure: Any tearing, fracturing, cracking, or rupturing visible on the BACK of the specimen. Crushed or dislodged granules, dents, and openings visible only on the top are not failures.
What a pass proves: That a new specimen of this product, laboratory-conditioned, took that energy from a uniform hard sphere without opening at the back.
What a pass does not prove: Nothing about aged material, irregular or softer stones, oblique impacts, or the rest of the assembly — and IBHS says the criteria do not necessarily reflect what insurers use to decide repair or replacement.

FM 4473 — freezer ice ball

Projectile: Ice balls cast from distilled water: 1¼ in / 15.3 g, 1½ in / 26.5 g, 1¾ in / 42.1 g, 2 in / 62.9 g for Classes 1 to 4.
How it is delivered: Propelled to a target kinetic energy — 3.72, 7.77, 14.95 and 26.81 ft-lb (5.0, 10.4, 20.3 and 36.4 J) — perpendicular to the panel within 90° ± 5°, twice at each location, with impact points chosen at edges, corners, unsupported areas, overlaps and joints.
What counts as failure: Visible cracking or breakage of the specimen.
What a pass proves: That new rigid roofing — the standard's stated scope is tiles or planks of slate, concrete or clay — survived an energy-matched ice impact at that class.
What a pass does not prove: FM says so itself: ice balls are harder and denser than hailstones and simulate a worst case, hailstones are far more variable, and weathering has the potential to significantly lower hail resistance. FM also states that meeting the standard does not by itself qualify a product for Approval.

IBHS Impact Resistance Test Protocol

Projectile: Laboratory-manufactured hailstones, 1.5 in and 2.0 in, made to mimic natural hail in density, strength, hardness, mass, kinetic energy and terminal velocity.
How it is delivered: Propelled under controlled conditions at shingles bought from retail shelves rather than supplied by manufacturers.
What counts as failure: Not pass/fail. Products are scored across three damage modes — dents (deformation), tears (breaches), and granule loss — and rated Excellent, Good, Marginal, or Poor.
What a pass proves: Relative performance against a more realistic projectile, judged on the damage modes people actually argue about after a storm.
What a pass does not prove: Still a laboratory comparison on new product. IBHS's own caveat is that the tests do not exactly replicate natural hailstone impacts and provide a way to compare relative product performance.

UL 2218's own text is not published free of charge; the description here is taken from IBHS's published summaries of its UL 2218 testing and from IBHS's Roof 101 material, both of which state the projectile sizes and the back-of-specimen pass criterion. FM 4473's figures are read directly from FM Approvals' published standard. None of these three tests is a building code, and none is a warranty.

A worked comparison

Same energy, different blow: why an energy-matched steel ball is not a hailstoneSection link

The tests are designed to match the kinetic energy of hail. Energy is not the whole of an impact, and the standards say so in their own words before this page does any arithmetic.

FM Approvals states the principle inside FM 4473: hailstones vary in shape, density and frangibility, and those properties affect the duration and magnitude of the impulsive force acting on the roof and the area over which the impulse is distributed. Two impacts can carry identical energy and load a shingle differently.

Here is the size of that difference for the Class 4 case, as arithmetic. This is the site’s own calculation from published inputs, not a test result, and it is not a claim that either test is more severe on a real roof.

mass of a sphere = density × (4/3)πr³

A 2-inch ice ball has a mass of 62.9 g — that figure is FM’s, from the standard’s own table. A 2-inch sphere has a volume of about 68.7 cm³, so at a carbon-steel density of roughly 7.85 g/cm³ a 2-inch steel ball weighs about 540 g — roughly 8.5 times the ice ball.

KE = ½mv², so equal KE ⇒ v ∝ 1/√m

To carry the same kinetic energy with 8.5 times the mass, the steel ball has to arrive at about one-third the speed (1/√8.5 ≈ 0.34). Same energy; roughly three times the momentum; a slower, heavier, harder, non-deforming contact against a faster, lighter, softer, self-destroying one.

That is the concrete reason the three tests on this page are not interchangeable and why IBHS built a third protocol using hail-mimicking stones rather than adopting either of the first two. It is also why “this shingle passed a 2-inch impact” is a statement about a laboratory event, not a description of what a 2-inch hailstone will do to a roof.

Three things this arithmetic is not

It is not a criticism of either standard. A uniform, repeatable projectile is exactly what a comparative test needs, and FM says plainly that ice balls are the nearest hailstone approximation known with known, repeatable properties.
It is not a claim that steel-ball testing is easier or harder than ice-ball testing on a real product. That is an empirical question, and this page does not have a source that answers it.
It is not a substitute for either result. If you want to compare products, use the published classifications and the published ratings — not this paragraph.

The uncomfortable finding

What happened when a testing laboratory put Class 4 products through Class 4 impactsSection link

In August 2014 IBHS published the results of UL 2218 testing on 22 asphalt shingle products from five manufacturers, bought from local vendors rather than supplied by the manufacturers. The products included basic three-tab, basic architectural, premium architectural, and Class 4-rated impact-resistant lines of both construction types — scrim-backed and polymer-modified.

The headline sentence is worth reading twice. IBHS reported that impact-rated products “should withstand testing without cracking or tearing on the shingle back for all four projectile sizes” — and then that none of the products tested, basic or impact rated, passed more than Class 2 impacts without at least one double-impact location failing the UL 2218 performance criteria.

Two qualifications belong immediately next to that, or the sentence becomes propaganda rather than evidence.

The impacts were aimed at the weak points. IBHS struck edges, corners, joints, and centres, including over framing where the deck is stiffest and the shingle has least room to flex. That is a legitimate and deliberately demanding application of the standard; it is not the same as a random storm.
Impact-rated products were still clearly better. IBHS reported impact-rated three-tab products passing roughly 55 to 225 percent more impact locations than basic three-tabs across the classes, and impact-rated architectural products roughly 25 to 80 percent better than basic architectural. Polymer-modified impact-resistant shingles outperformed scrim-backed ones by roughly 20 to 50 percent, most noticeably at the larger ball sizes.

Read together, those three findings say something precise and useful: the classification is a real and measurable difference between products, and it is not a threshold below which damage does not occur. That is the whole argument of this page in one sentence, and it came out of a laboratory rather than out of an opinion.

IBHS has since moved to its own protocol with hail-mimicking projectiles and publishes product-level ratings from it. In the round announced on 19 November 2025, IBHS reported that no product earned its top rating of Excellent, and that it does not recommend products rated Marginal or Poor in areas that experience hail. This page does not reproduce the product list, because it changes; the point to carry is the shape of the distribution, not this year’s entries.

Reading the data

Hail climatology, and the difference between frequent hail and large hailSection link

Every hail map you have seen was built from reports made by people. That does not make them wrong — it makes them a record of reporting as well as a record of weather, and reading them well starts with knowing which is which.

Two thresholds do most of the work. The National Weather Service defines a severe thunderstorm as one producing hail at least 1 inch in diameter or larger. The Storm Prediction Center defines a significant severe thunderstorm as one producing hail at least 2 inches or larger. Those two thresholds produce two different national pictures, and SPC publishes them as separate climatologies for exactly that reason.

The distinction matters for spending. A place with frequent 1-inch hail has a granule-loss, soft-metal, and nuisance-claim problem. A place where 2-inch-plus hail turns up every few years has a covering-replacement problem. Hold that against NRCA’s published damage thresholds, which put a lightweight three-tab covering at 1 inch and a heavyweight laminate at 1¼ inches, under the assumptions NRCA states with them and reproduced with that table further down this page. The severe-hail threshold and the damage threshold of a basic three-tab roof are the same number, and moving up to a heavyweight laminate buys a quarter of an inch. Neither choice puts a covering anywhere near the 2-inch significant-severe threshold. Choosing a tougher covering is a decision about how much damage a storm does, not about whether it does any.

There is also a step in the record. Effective 5 January 2010, every NWS forecast office raised the minimum hail size for issuing a severe thunderstorm warning from 3/4 inch to 1 inch. A warning threshold is not a reporting threshold, but the two are entangled in practice, and any long-run comparison that crosses 2010 has to account for it before it can claim a trend.

The public hail datasets a property owner can actually reach, and what each one can and cannot support. All are NOAA products; none of them is a forecast for a specific building.
Source	What it is	What it can tell you	What it cannot tell you
SPC preliminary storm reports	Daily listings automatically collected from local storm reports sent by NWS offices.	Whether hail was reported near a location on a given date, and at what estimated size — quickly, within days.	SPC states the list is preliminary and likely does not contain all severe weather reports; a report arriving more than ten days late does not appear at all.
NCEI Storm Events Database	The official record of NOAA Storm Data, entered by the NWS, running from January 1950 onward.	The final, official record for a date and place, with damage estimates — the source SPC itself points to for official data.	It takes months to compile, periods of record differ by event type because collection procedures changed over time, and absence of a report is not evidence that no hail fell.
SPC severe weather climatology maps	Probability surfaces estimated from severe weather reports for 1982–2011, gridded at 80 km × 80 km and smoothed with Gaussian filters of 15 days in time and 120 km in space.	The regional shape of hail risk, and — by comparing the all-hail and significant-hail products — where large hail is common rather than merely where hail is common.	Anything about a specific address. The smoothing is explicitly designed to generalise across 80 km cells, and the report period ends in 2011.
SPC hail size table	The object-to-size conversion used to translate reports: 1.00 in quarter, 1.75 in golf ball, 2.00 in hen egg, 2.75 in baseball, 4.00 in softball.	How to describe a stone you have, and how to read a size in someone else's report.	Nothing about damage. It is a translation table, and SPC's own instruction on it is that measurement, not estimation, is encouraged.

Read this table one item at a time

SPC preliminary storm reports

What it is: Daily listings automatically collected from local storm reports sent by NWS offices.
What it can tell you: Whether hail was reported near a location on a given date, and at what estimated size — quickly, within days.
What it cannot tell you: SPC states the list is preliminary and likely does not contain all severe weather reports; a report arriving more than ten days late does not appear at all.

NCEI Storm Events Database

What it is: The official record of NOAA Storm Data, entered by the NWS, running from January 1950 onward.
What it can tell you: The final, official record for a date and place, with damage estimates — the source SPC itself points to for official data.
What it cannot tell you: It takes months to compile, periods of record differ by event type because collection procedures changed over time, and absence of a report is not evidence that no hail fell.

SPC severe weather climatology maps

What it is: Probability surfaces estimated from severe weather reports for 1982–2011, gridded at 80 km × 80 km and smoothed with Gaussian filters of 15 days in time and 120 km in space.
What it can tell you: The regional shape of hail risk, and — by comparing the all-hail and significant-hail products — where large hail is common rather than merely where hail is common.
What it cannot tell you: Anything about a specific address. The smoothing is explicitly designed to generalise across 80 km cells, and the report period ends in 2011.

SPC hail size table

What it is: The object-to-size conversion used to translate reports: 1.00 in quarter, 1.75 in golf ball, 2.00 in hen egg, 2.75 in baseball, 4.00 in softball.
What it can tell you: How to describe a stone you have, and how to read a size in someone else's report.
What it cannot tell you: Nothing about damage. It is a translation table, and SPC's own instruction on it is that measurement, not estimation, is encouraged.

A report requires an observer, so report density reflects where people are as well as where hail falls. That is a limitation of the method rather than a criticism of it, and it is the single most important thing to hold in mind when a hail map is used to justify a purchase.

The distinction that decides claims

Functional damage, cosmetic damage, and why competent people disagreeSection link

NRCA’s published position is a definition rather than a judgment call: a roof system is not functionally damaged by hail unless the roof’s water-shedding capabilities are diminished or the roof’s expected service life is reduced. For an asphalt shingle, that means the unit must be punctured or fractured. A mark that removes granules and nothing else does not meet it. Functional versus cosmetic is therefore not a euphemism invented by insurers; it is a technical line that the trade drew and that policy language then borrowed.

Where it gets genuinely hard is that the two categories are not independent. NRCA notes that for a shingle in relatively good condition, a hailstone able to dislodge granules far enough to expose the coating asphalt will generally also have fractured the reinforcement and produced a detectable bruise. So severe “cosmetic” granule loss on a sound roof is often evidence of a fracture underneath — while the same granule loss on a roof near the end of its life may mean the shingle was already flaking. Same observation, opposite conclusion, depending on the condition of the roof it is on.

Add the geometry from the diagram above — the fracture is on the underside, and nobody in the field can look at the underside — and the dispute structure is complete. This is why hail claims turn on documented protocol and photographs rather than on a verdict, and why “my contractor says it is damaged and the adjuster says it is not” is usually two people applying different definitions to partial evidence rather than one of them lying.

NRCA's published smallest hail sizes that typically damage common steep-slope roof coverings. Reproduced from NRCA/Professional Roofing, May 2009, with the source's own stated assumptions.
Roof covering	Threshold hail size	What that means in practice
Asphalt three-tab, lightweight	1 in	The same size the NWS uses to call a thunderstorm severe. A basic three-tab roof and the severe-hail definition sit at the same number.
Asphalt laminated, heavyweight	1¼ in	A quarter of an inch of headroom over a lightweight three-tab. Small in absolute terms, and it lands squarely in the size range that is most common.
Wood shingles	1¼ in	Damage mode is splitting along the grain with a coincident dent, rather than the mat fracture that governs asphalt.
Gray slate, ¼ in thick	1½ in	Higher threshold, brittle failure. It resists more and then breaks rather than bruises.
Green slate, ¼ in thick	1¾ in	Higher again — the same nominal thickness in a different stone behaves differently, which is a caution against reasoning from material category alone.
Clay tile	1½ in	Brittle, individually replaceable, and vulnerable to being broken by the people who come to inspect it.
Concrete tile	1¾ in	The highest threshold in the published set, and the heaviest — which makes it a structural conversation before it is a hail one.

Read this table one item at a time

Asphalt three-tab, lightweight

Threshold hail size: 1 in
What that means in practice: The same size the NWS uses to call a thunderstorm severe. A basic three-tab roof and the severe-hail definition sit at the same number.

Asphalt laminated, heavyweight

Threshold hail size: 1¼ in
What that means in practice: A quarter of an inch of headroom over a lightweight three-tab. Small in absolute terms, and it lands squarely in the size range that is most common.

Wood shingles

Threshold hail size: 1¼ in
What that means in practice: Damage mode is splitting along the grain with a coincident dent, rather than the mat fracture that governs asphalt.

Gray slate, ¼ in thick

Threshold hail size: 1½ in
What that means in practice: Higher threshold, brittle failure. It resists more and then breaks rather than bruises.

Green slate, ¼ in thick

Threshold hail size: 1¾ in
What that means in practice: Higher again — the same nominal thickness in a different stone behaves differently, which is a caution against reasoning from material category alone.

Clay tile

Threshold hail size: 1½ in
What that means in practice: Brittle, individually replaceable, and vulnerable to being broken by the people who come to inspect it.

Concrete tile

Threshold hail size: 1¾ in
What that means in practice: The highest threshold in the published set, and the heaviest — which makes it a structural conversation before it is a hail one.

NRCA states these thresholds are the culmination of more than 45 years of laboratory testing with simulated hail and tens of thousands of field inspections, and are given for a specific set of assumptions: hard hail, perpendicular impacts, reasonable support, and roof systems in relatively good or midlife condition. Change any assumption and the threshold moves. The source's wood-shake row is omitted here because its printed value could not be read with confidence from the published PDF, and this site does not publish a number it cannot verify. Nothing in this table is a determination about a specific roof.

After a storm

What a property owner can usefully do, all of it from the groundSection link

The evidence that is actually worth something

Date and time, written down. Dating the storm is frequently the contested element of a claim, and memory degrades faster than evidence.
A measured hailstone, if one survives. Photograph it beside a ruler or a coin. SPC’s own instruction on its size table is that measurement, not estimation, is encouraged, and a measured stone beats every adjective.
The soft metals, photographed from the ground. Gutters, downspouts, drip edge, vent hoods, chimney caps, garage doors, and the aluminium fins on the outdoor air-conditioning unit. NRCA treats metals as a permanent record of every storm they have endured, and puts dent width in air-conditioner fins at about 80 percent of hailstone diameter.
Which elevations took it. Note the wind direction during the storm and which sides of the building faced it. Hail has a predominant fall direction, and damage is assessed slope by slope.
Spatter marks, before they fade. Clean spots where grime, oxide, or algae have been knocked off surfaces. NRCA says they typically linger about a year depending on exposure and climate, which makes them a rough date stamp.
Your declarations page and endorsements. Before anyone quotes anything, find out whether the wind and hail deductible is a dollar figure or a percentage, whether the roof settles at replacement cost or actual cash value, and whether any endorsement changes that at a stated roof age.

What not to do

Do not climb up to look. See above. It is the one step that converts a property question into a hospital question.
Do not sign anything at the door. Post-storm urgency is manufactured. Verify the contractor first — licence or registration, insurance certificate, physical address, and complaint history — using the contractor verification steps.
Do not let anyone create damage to demonstrate it. NRCA’s guidance illustrates intentionally inflicted mechanical damage precisely because it is a recognised phenomenon in this field. If a demonstration involves a tool touching your roof, stop it.
Do not assume the roof needs replacing because there was hail. NRCA opens its guidance by naming this as a common mistake among professionals and homeowners alike. Sometimes the answer is that the gutters and the air-conditioner need attention and the roof is fine — and that is a good outcome, not a failed claim.

When the question is whether to repair a hailed roof or replace it, the economics belong on the repair-or-replace triage, and the general symptom routing on roof care. Neither of them, and not this page, can make the determination for a specific building.

What a claim pays, and why

The settlement arithmetic, using regulators' own worked examplesSection link

These are teaching examples published by state insurance regulators to show the structure. They are not market prices, and they are emphatically not a prediction of what any carrier will pay on any claim.

Replacement cost coverage: $6,000 paidTexas Department of Insurance illustration: a $10,000 roof replacement, a house insured for $200,000 with a 2% ($4,000) deductible. TDI's point is that this figure is the same no matter how old the roof is.
Actual cash value, 10-year roof: $3,000 paidSame TDI example. TDI puts the actual cash value of the 10-year-old roof at $7,000; the $4,000 deductible comes off that.
Actual cash value, 20-year roof: $0 paidSame TDI example. TDI puts the actual cash value at $4,000, and the $4,000 deductible consumes it entirely.
A percentage hail deductible: $16,250Montana Commissioner of Securities and Insurance illustration: a 5% split deductible for hail on a home insured for $325,000. The homeowner pays that first.

Units: U.S. dollars, exactly as printed in the cited regulator publications
Scope included: The settlement arithmetic for one roof-covering claim: deductible, depreciation, and the payment that results
Not included: Real roof prices in any market, contractor pricing, supplements, code-upgrade coverage, matching, additional living expense, and any question of whether a particular loss is covered at all
Geography: Texas and Montana regulator publications, used as structural illustrations — not as prices or practices anywhere
Data as of: Texas Department of Insurance page last updated 24 January 2024; Montana Commissioner of Securities and Insurance article published 25 April 2023
Confidence: High as arithmetic, none as prediction. Both are deliberately simplified teaching examples from regulators, chosen here because they are public and checkable.
Method: How this figure is built

The three levers, separated

Almost every surprise in a hail settlement comes from one of three places, and they compound rather than substitute.

The deductible may be a percentage rather than a dollar figure. Montana’s regulator describes split deductibles — losses from certain causes such as hail and wind carrying a much higher deductible, which may change from a fixed price to a percentage of the total amount of insured coverage. In its example, 5% of a $325,000 limit is $16,250. That is not a share of the roof; it is a share of the dwelling limit.
The roof may settle at actual cash value rather than replacement cost. NAIC’s consumer explanation is that replacement cost pays to repair or replace without deducting for depreciation, while actual cash value pays the depreciated cost — and that insurers usually calculate depreciation from the property’s condition at the time of loss, what a new item would cost, and how long it would normally last.
An endorsement may change the rules by roof age. Iowa’s Insurance Division notes that many homeowners policies now include an endorsement specifically about how a damaged roof will be settled — one that may replace replacement cost with actual cash value, or set a specific depreciation amount based on the age and type of roofing material. Montana’s regulator gives the same structure a threshold: once a roof reaches 10 to 15 years old, hail damage may be paid at actual cash value rather than replacement cost value.

Stack them and Montana’s worst-case illustration follows: a 12-year-old roof, $20,000 to replace, a $250,000 limit with a 2% split deductible, a policy that pays actual cash value at ten years and older. Fifty percent depreciation takes the settlement basis to $10,000, the $5,000 deductible comes off that, and the regulator’s figure for what the homeowner is left carrying is $15,000. The roof was covered. The homeowner still paid three-quarters of it.

The recoverable-depreciation wrinkle

Many replacement-cost policies pay in two steps: an actual cash value payment first, and the remainder — sometimes called recoverable depreciation — only after the work is actually done and documented. Whether a given policy works that way, what proof it requires, and what deadline applies are in the policy, not here. The reason to know the structure exists is that the first cheque is a common cause of a homeowner concluding the claim was underpaid when it was not yet finished — and an equally common lever for a contractor to argue that a cheaper scope is affordable.

Nothing in this section is coverage advice, a coverage opinion, or a prediction. Coverage, deductibles, depreciation schedules, endorsements, deadlines, and dispute rights are set by the specific policy and by the law of the state it was written in, and they change. Read the declarations page and the endorsements, and take questions to your carrier, to a licensed agent, or to your state insurance department.

Considerations

What changes this on a real buildingSection link

Hail and impact

An impact classification is a property of a tested specimen under a defined laboratory procedure. It is not a rating of your roof, and it is not a claim that the product will be undamaged. The two mechanisms that break the inference are stated inside the standards themselves: the projectile is not a hailstone, and the specimen is new.

There is also an installation dimension that no rating covers. Impact-rated asphalt shingles achieve their resistance either through a scrim or mesh backing or through polymer-modified asphalt, per IBHS. Both depend on the shingle being fastened in its designed nailing zone, sealed, and supported. A product tested to Class 4 and installed badly is not a Class 4 roof in any meaningful sense.

“Class 4 impact resistant” does not mean hail proof. UL 2218 and FM 4473 grade new material against a standardised projectile in a laboratory; a Class 4 roof can be and routinely is damaged by hail.

Climate

Two different questions get answered with the same map, and they should not be. How often is hail reported here and how often does large hail fall here have different answers in different places, which is why NOAA’s Storm Prediction Center publishes separate climatologies for all severe hail and for significant severe hail. The Storm Prediction Center defines severe hail as at least 1 inch in diameter and significant severe hail as at least 2 inches.

A place with frequent 1-inch hail and a place with occasional 2.5-inch hail present different problems and justify different spending. Frequent small hail is mostly a granule-loss and soft-metal problem; rare large hail is a covering-replacement problem. Buying for the wrong one is the most common way this decision goes wrong.

Code and jurisdiction

There is no nationwide building code for site-built houses in the United States. Whether an impact-classified covering is required, encouraged, or irrelevant where a building sits is decided by the code that jurisdiction has adopted, as amended, and by the authority having jurisdiction — not by a product label and not by this page. This page makes no code claim at all, because it is a national page and a code claim without a jurisdiction is not a claim.

The one adopted-law example on this page is deliberately narrow: a 1998 Texas insurance rating order, which is insurance regulation rather than building code, cited for the structure it created rather than as a statement of current Texas practice.

Any code requirement affecting a roof covering carries a jurisdiction, an adopted edition, local amendments, and an effective date. Confirm all four with the authority having jurisdiction before treating anything as a requirement.

Insurance — a structure to read, never an outcome to expect

Texas is the clearest publicly documented example of how impact ratings and insurance became entangled. Under Commissioner’s Order No. 98-0390, effective 2 May 1998, the Texas Department of Insurance adopted mandatory homeowners premium credits for roof coverings meeting UL Standard 2218, graduated by class and by rating territory — from 4% for a Class 4 covering in the lowest-credit territory to 35% in the highest.

In the same order, TDI adopted Endorsement HO-145, Exclusion of Cosmetic Damage to Roof Coverings Caused by Hail, which may be attached only to a policy that is receiving that credit. Its text excludes cosmetic loss or damage, defined as damage that alters the physical appearance of the roof covering but does not allow water penetration or cause the covering to fail to perform its intended function. TDI required the endorsement to be signed by the insured, and required insurers to inform consumers of the possibility that cosmetic damage may be excluded.

Read that structure carefully, because it is the honest version of the trade: in that regulatory scheme, the same classification that bought a premium credit was also what made a cosmetic exclusion attachable. That is not an argument against impact-rated roofs. It is an argument for reading the declarations page and the endorsements before deciding what the upgrade bought.

TDI also stated plainly in the same rules that the credits programme is solely for enabling property owners to obtain a premium reduction, and is not intended to require or create any express or implied warranty by the manufacturer, supplier, or installer. A classification is not a promise from anyone.

That order is dated 1998, it is Texas, and it is insurance rating regulation rather than building code. This page does not assert it is current practice in Texas or anywhere else. Confirm current rules with your state insurance department and current terms with your carrier. This section needs legal review before the page is registered.

Wind

Hail rarely arrives alone. The same storms produce straight-line wind, and wind performance is a completely separate, site-specific question — basic wind speed, exposure, building height and geometry, pressure zone, enclosure, risk category, attachment, and the tested assembly all matter, and a marketing “mph rating” on a shingle wrapper is not a code determination. A covering chosen for impact still has to be specified and fastened for the wind at that address. That conversation belongs on the tornado and straight-line wind page.

Wind performance is site- and building-specific and is determined for a building, not for a product. Nothing on this page is a wind-design determination.

Fire

Impact classification and fire classification are independent, and neither implies the other. A roof’s Class A, B, or C fire rating applies to a tested assembly — deck, underlayment, and covering together — rather than to the visible covering by itself, and swapping in an impact-rated covering does not carry a previous fire classification with it. If the building is anywhere fire classification matters, the two requirements are specified together or the specification is incomplete.

Maintenance

The most useful hail habit costs nothing: after any storm that sounded serious, photograph the soft metals from the ground — gutters, downspouts, drip edge, vent hoods, chimney caps, and the aluminium fins on the outdoor air-conditioning unit. NRCA notes that metals are a permanent record of every hailstorm they have endured, and that a dent’s width in air-conditioner fins is typically about 80 percent of the hailstone diameter. It also notes that spatter marks — where grime, oxide, or algae have been cleaned off by impacts — typically linger for about a year. That is a date stamp, and dating a storm is frequently the contested part of a claim.

Structural weight

Impact resistance is not the same as weight or strength, and heavier is not automatically better. Concrete and clay tile carry higher published hail thresholds than asphalt and are also brittle: they resist a larger stone and then break rather than bruise, and broken tiles are a fall-and-breakage hazard for anyone working on the roof afterwards. Any change between covering families — asphalt to tile, tile to metal — is a structural and detailing question for the specific building before it is a hail question.

Access and site conditions

Every diagnostic step a homeowner should take here is ground-based: photographs of soft metals, photographs of the ground and the deck furniture, notes on which elevations took the storm, the date and time, and a measurement of a hailstone if one is still intact — the Storm Prediction Center’s own instruction on its size table is that measurement, not estimation, is encouraged. The roof itself is inspected by somebody with fall protection and a reason to be up there.

Do not climb onto a roof after hail to look for damage. Loose granules on a slope are the specific footing hazard, and the marks that matter are not visible from a standing position anyway.

Warranty and repair

Which document is in play when hail damages a roofSection link

Four documents get pulled off the shelf, and they answer four different questions. None of them is a substitute for reading it.

The impact classification itself: Not a warranty and not a promise. Texas’s own rating rules said so explicitly when the state created its credit programme: the programme is for obtaining a premium reduction and is not intended to require or create any express or implied warranty by the manufacturer, supplier, or installer. A UL or FM classification is a statement about a tested specimen, made by a testing body, to a defined procedure.
The manufacturer's limited warranty: A separate document with its own terms. Whether hail is covered, excluded, or treated as an outside cause; whether an impact-rated line carries different terms; what proration applies; and what voids it are set by that specific document for that specific product. Find the exclusions section and read it before the storm rather than after — and see how to read a roofing warranty properly for what the categories mean.
The installer's workmanship warranty: Hail damage often surfaces installation defects that predated it — fasteners outside the nailing zone, unsealed tabs, poor flashing. That is a workmanship question, not a hail question, and it is the document that answers it. Read it for length, for what triggers a callback, for whether the diagnostic visit is chargeable, for transferability, and for what happens if the company stops trading.
The insurance policy and its endorsements: The declarations page tells you the dwelling limit and the deductible structure. The endorsements tell you whether the roof settles at replacement cost or actual cash value, whether that changes at a stated roof age, and whether any cosmetic-damage exclusion is attached. This is the document that decides money, and it is the one most homeowners have never opened. This page cannot tell you what yours says.

Repairability

Hail repairability is unusual, because the thing that decides it is usually availability rather than technique.

Matching. Asphalt shingle lines are discontinued and colours shift between production runs. A slope-only repair on a ten-year-old roof frequently cannot be made to match, and whether that matters contractually is a policy and contract question rather than a roofing one.
Sealing. Replacement units on a cold or shaded slope may not seal down for weeks, which is a wind exposure in the interim. Timing a hail repair into the wrong season creates a second problem.
Soft metals. Gutters, hoods, caps, and fascia are usually replaced rather than repaired, and they are frequently the only visibly damaged components after a moderate storm.
Tile and slate. Individually replaceable, which is a genuine advantage — but the replacement has to match in profile and colour, and walking the roof to do it breaks more units. Specialist labour, not general roofing labour.
The fracture you cannot see. A bruised shingle that has not opened is not repairable in any practical sense; it is a shortened service life sitting in place. That is precisely what makes the functional-versus-cosmetic argument financially real rather than semantic.

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.

Ask before you sign

Questions to ask an installerSection link

Hail brings two different people to the door — a roofing contractor and, separately, an insurance adjuster. These questions are for the contractor, and the useful ones are the ones that reveal whether they are diagnosing or selling.

Which elevations did you assess, and what did each one show? Not the roof — the elevations.
Hail arrives with a predominant fall direction influenced by wind, so exposure differs by slope. A single verdict for a whole roof, delivered quickly, has skipped the step that produces the verdict.
What did you find in the soft metals, and does it match what you found on the covering?
NRCA treats metals as a permanent record of the hailstorms they have endured, with dent width in air-conditioner fins running about 80 percent of hailstone diameter. Marks in the covering with nothing in the metals — or the reverse — is a discrepancy worth an explanation.
Are you calling this functional damage or cosmetic damage, and on what basis?
NRCA’s protocol defines functional damage as damage that diminishes water-shedding capability or reduces expected service life. Somebody who cannot say which one they mean is not going to be able to defend it to an adjuster either.
Are you counting granule loss as damage? If so, say why in this specific case.
NRCA is explicit that granule loss on its own is normal weathering rather than functional damage, and that shingles losing enough granules to expose the coating asphalt without fracturing the reinforcement are typically already near the end of their service life. There are defensible arguments here — but they have to be made, not assumed.
How old is this roof, and what does my policy do to an older roof?
A regulator’s own example puts a common threshold at 10 to 15 years, after which some policies settle hail damage at actual cash value. A contractor who has not asked what the policy says is quoting into an unknown.
If you are proposing an impact-rated covering, which classification, to which standard, and where is the label?
Class, standard, and product identity are three separate facts. Texas’s credit programme required individual units to bear a testing-laboratory label showing the classification, the manufacturer, the year, and the brand. A proposal that says “impact resistant” and nothing else is not specifying anything.
Will you put in writing that the covering can still be damaged by hail?
It is a fair question and an easy yes for an honest contractor. IBHS’s own testing found Class 4-rated products failing below Class 4 impacts, and FM’s standard says weathering can significantly lower hail resistance. Anyone who will not say it in writing has told you something.
Does anything you are proposing affect my premium, and have you seen the endorsement it is attached to?
A credit and an exclusion can travel together — that is exactly the structure Texas created in 1998. The contractor is not your agent and should not be answering coverage questions, but they should know enough to send you to somebody who can.
What is your relationship to the insurance side of this claim, and who is paying you?
Scope negotiation, supplement handling, and any assignment of benefits are contractual arrangements with real consequences and state-specific rules. You want the answer before you sign, not in the paperwork afterwards.
If the answer turns out to be that this roof is fine, will you say so?
NRCA’s guidance opens by saying that many roofing professionals, insurance professionals, and homeowners mistakenly assume a roof needs replacement after a hailstorm, and that this is not always the case. “No damage” has to be an available finding, or the inspection was never a question.

Require these in writing

The date of the storm the claim relates to, and what evidence dates it.
Findings recorded by elevation, with photographs, not a single roof-wide verdict.
An explicit statement of which marks are being called functional damage and why.
Product identity: manufacturer, product line, colour, classification, and the standard the classification is to.
Fastening specification, including the nailing zone and fasteners per unit, since impact resistance depends on it.
What is being replaced in soft metals — gutters, drip edge, hoods, caps, flashings — as separate line items.
Underlayment and any ice barrier as separate line items rather than folded into a square price.
Whether the work triggers a permit in your jurisdiction, and who pulls it.
A written statement that an impact-rated covering can still be damaged by hail.
Payment schedule and any deposit, with no term making payment contingent on a claim outcome unless you and your own adviser have read it.

What goes wrong

Misconceptions and failure modesSection link

Common misconceptions

Common belief
A Class 4 roof will not be damaged by hail.
What is actually true
The rating is a laboratory result on new material against a standardised projectile. IBHS’s published 2014 testing of 22 asphalt shingle products, including Class 4-rated impact-resistant lines, reported that none of the products tested — basic or impact rated — passed more than Class 2 impacts without at least one double-impact location failing the UL 2218 performance criteria. Impact-rated products did clearly better than basic ones, which is the real case for buying them. “Better” is not “immune.”
Common belief
Missing granules after a storm mean the roof is damaged.
What is actually true
Granule loss begins at manufacture and continues for the life of the roof, and NRCA’s protocol treats it as normal weathering rather than functional damage. NRCA goes further: shingles that lose enough granules to expose the coating asphalt without fracturing the reinforcement are typically already near the end of their service lives. That may still be a reason to plan a replacement. It is not, on its own, hail damage.
Common belief
The ice-ball test is realistic, so a pass means real-world protection.
What is actually true
FM Approvals says otherwise in its own standard. Ice balls are generally harder and denser than hailstones, so an ice ball simulates the worst-case hailstone — and the major difference between the two is that hailstones are far more variable in shape, density, and frangibility, which changes the force and how it is distributed. A worst-case uniform projectile is a good engineering test and a poor model of a storm.
Common belief
A roof rated ten years ago is still rated today.
What is actually true
The classification of the product is unchanged; the roof is not the product any more. FM 4473 states in its scope that exposure of roof coverings to the elements over an extended period has the potential to significantly lower their hail resistance, and that the standard is intended for testing new roof covers. There is no published derating curve, which is why this page does not offer one — but there is no basis for treating a fifteen-year-old covering as equivalent to the tested specimen either.
Common belief
Wind-driven hail hits much harder, so a windy storm is always worse.
What is actually true
Wind matters, but mostly for aim. NRCA’s guidance is that wind imparts a directional component to falling hailstones while generally contributing only minorly to their impact forces. The practical effect is on which slopes and walls are struck square versus at a glancing angle — which is a real effect worth assessing elevation by elevation, and a different claim from “wind turns small hail into big hail.”
Common belief
The hail map shows my area is high risk, so I need the best roof available.
What is actually true
Check which map. The Storm Prediction Center publishes separate climatologies for severe hail (1 inch and larger) and for significant severe hail (2 inches and larger), and the two are not the same picture. Frequent small hail and rare large hail justify different spending. Both climatologies, moreover, are built from reports — and a report requires somebody present to make it.
Common belief
The storm-report record proves hail is getting worse here.
What is actually true
It cannot prove that on its own. The minimum hail size for a severe thunderstorm warning changed from 3/4 inch to 1 inch effective 5 January 2010, which puts a step in the record. SPC’s own storm-report listings are described as preliminary and likely not to contain all severe weather reports, with NOAA’s Storm Events Database as the official final source. Report databases are records of reporting as much as records of weather.
Common belief
My insurer gave me a discount for the impact-rated roof, so hail damage is covered.
What is actually true
Those are separate provisions and can even move in opposite directions. In Texas’s 1998 rating order, the endorsement that excludes cosmetic hail damage to roof coverings could be attached only to policies receiving the impact-resistant credit, and had to be signed by the insured. Whether anything similar applies to your policy today, in your state, is a question for the declarations page and the endorsements.
Common belief
A free post-storm inspection is free.
What is actually true
It is a sales call with a ladder. That does not make it dishonest — a competent contractor genuinely does need to look — but the incentive structure is worth naming out loud. Ask whether “no damage found” is a result they will put in writing, and read how to verify a contractor before anyone gets on the roof.

How it actually fails

Mat fracture with no opening on the surface: The impact starts a fracture on the underside — the region of maximum tensile strain — which runs upward and may never break through the granule surface. NRCA’s account is that such fractures are often accompanied by fractures in the shingle reinforcement, which is what diminishes water-shedding; service life is shortened either way, with nothing conclusive to see from above.What you can see: A localised soft spot that can be felt rather than seen. Marks in soft metals from the same storm. Leaks that appear a season or two later at no obvious point of entry.
Exposed asphalt after granule displacement: Enough surfacing is knocked away to expose the coating asphalt, which then ages under ultraviolet light much faster than the surrounding field. NRCA’s framing: for a shingle in good condition, a stone able to dislodge granules that far generally also fractures the reinforcement.What you can see: Dark spots against the field colour, concentrated on one or two elevations. Granule accumulation at downspout outlets after the storm rather than continuously.
Soft-metal damage without covering damage: Aluminium and thin steel deform at energies that a sound covering shrugs off, so gutters, hoods, caps, and air-conditioner fins mark first.What you can see: Dents in fins and gutters with no fractures in the field. This is a real and common outcome, and it usually means a genuine storm and a roof that is fine.
Aged covering failing at a size it once survived: Asphalt stiffens and the mat embrittles with age and heat, so the same stone that bruised a five-year-old roof can fracture a fifteen-year-old one. FM 4473 names weathering as a potential significant reduction in hail resistance.What you can see: Widespread fracturing after a storm that produced little elsewhere in the neighbourhood. Brittle, cracking behaviour when a unit is lifted by a professional.
Damage concentrated on one elevation and dismissed on the roof average: Hail has a predominant fall direction, so one or two slopes take square hits while the rest take glancing blows. An assessment that averages the roof loses the finding.What you can see: Marks that stop at a ridge or a hip line. Soft metals damaged on one side of the house only.
Impact damage misattributed to hail, and vice versa: Blisters, foot traffic, tree abrasion, manufacturing variation, and normal aging can all produce marks that resemble hail. NRCA names this explicitly as the reason the assessment is a protocol rather than a glance.What you can see: Marks in traffic paths rather than random distribution. Damage under overhanging limbs. Circular blisters with no coincident mark in any metal on the property.
The claim dated to the wrong storm: Two hail events a year apart, and only one is inside the policy period or the reporting deadline. Dating is often the contested element.What you can see: Spatter marks — which NRCA says typically linger about a year, depending on exposure and climate — present or absent. Weathered versus fresh fracture surfaces. Your own dated photographs, which is why taking them costs nothing and is worth a great deal.
Impact-rated shingle installed outside its nailing zone: The reinforcement that gives an impact-rated shingle its resistance depends on the unit being fastened where the manufacturer says and sealing down. High nails, overdriven nails, and unsealed tabs undo it.What you can see: Units lifting or sliding in wind before any hail arrives. Shiners visible from the attic side. Damage patterns that track the fastener line.

Sources and further readingSection link

Understanding Roofing / Published August 26, 2026

Scope and limitations

It cannot tell you whether your roof is damaged.
The condition that defines failure in the laboratory is on the underside of the covering, and the marks that matter in the field are assessed on the building by somebody qualified to be on it.
It cannot tell you what hail will do at your address.
Hail climatology is built from human reports, the severe-hail reporting threshold changed in 2010, and no report record is a forecast for one building.
It does not publish a derating curve for aged coverings.
FM 4473 states that weathering has the potential to significantly lower hail resistance; no public standard converts that into a number, and inventing one would be decoration.
It does not name or rank products.
IBHS publishes product-level impact ratings; this page describes the method and points there rather than reproducing a product list that changes annually.
It cannot tell you whether your insurance responds to hail damage, what your deductible will be, whether depreciation is recoverable, or how a dispute would be resolved.
Those are set by the policy and by the law of the state it was written in.
It makes no code claim.
Whether any impact classification is required where a building sits is a jurisdiction-specific question for the authority having jurisdiction.
It publishes no roof prices.
The dollar figures here are regulators' own illustrations of settlement arithmetic, not market data.

Class Number 4473: Specification Test Standard for Impact Resistance Testing of Rigid Roofing Materials by Impacting with Freezer Ice Balls
FM Approvals / July 2005; effective date stated in the document as 1 September 2005
The Class 1–4 ice ball diameters (1¼, 1½, 1¾ and 2 in / 31.8, 38.1, 44.5 and 50.8 mm) and masses (15.3, 26.5, 42.1 and 62.9 g); the target kinetic energies of 3.72, 7.77, 14.95 and 26.81 ft-lb (5.0, 10.4, 20.3 and 36.4 J); the perpendicular 90° ± 5° impact and two impacts per location; the pass criterion of no visible cracking or breakage; the statement that ice balls are generally harder and denser than hailstones and simulate a worst-case hailstone while hailstones are more variable in shape, density and frangibility; that those properties affect the duration and magnitude of the impulsive force and the area over which it is distributed; and that exposure of roof coverings to the elements over an extended period has the potential to significantly lower their hail resistance, the standard being intended for new rigid roof covers.
A test standard, not adopted law and not a building code anywhere. Its stated scope is new rigid roofing materials — tiles or planks of slate, concrete or clay — so it does not by itself describe how asphalt shingles are graded, and FM states that meeting it does not by itself qualify a product for Approval.
Relative Impact Resistance of Asphalt Shingles: Summary of UL 2218 Impact Tests
Insurance Institute for Business & Home Safety (IBHS) / August 2014
That UL 2218 Class 1–4 projectiles are steel balls of 1.25, 1.50, 1.75 and 2.00 in dropped from the height needed to achieve the kinetic energy a hailstone of the same size would have in a thunderstorm; that any evidence of opening — tearing, fracturing, cracking or rupturing — on the back of the shingle is recorded as a failure and its absence as a pass; that crushed or dislodged granules, dents, and openings visible only on the top are not failures under UL 2218 criteria; that 22 products from five manufacturers were tested and none — basic or impact rated — passed more than Class 2 impacts without at least one double-impact location failing; that impact-rated products nonetheless outperformed basic products at every class; that crushed granules were seen on every laboratory panel but that damage mode is not seen in the field because hailstones are not hard enough to crush granules; that claims adjusters will rarely be able to view the back of shingles after a hail event without causing more damage; and that the performance criteria used do not necessarily reflect those insurers use to decide repair or replacement.
A 2014 research summary on products purchased in 2013, testing a small number of impact locations per product. IBHS itself notes the sample sizes are small, that rating impacts is partly subjective, and that a repeat series was underway. It is not a product-selection guide and does not describe today's market or IBHS's current rating protocol.
Roof 101
Insurance Institute for Business & Home Safety (IBHS)
That impact resistance in asphalt shingles is achieved either with a scrim backing or with polymer-modified asphalt; the UL 2218 Class 1–4 steel ball diameters and the pass criterion of no crack visible on the back of the shingle; that FM 4473 propels frozen water ice balls to the kinetic energy of hailstones with matching Class 1–4 classifications; that the IBHS protocol uses laboratory-manufactured hailstones made to mimic natural hail density, strength, hardness, mass, kinetic energy and terminal velocity and evaluates deformation, granule loss and breaches; and that shingles rated Good or Excellent in the IBHS Hail Impact Ratings qualify for the FORTIFIED High Wind & Hail Supplement.
An educational overview published by a research organisation funded by property insurers. It is not adopted law, not a code requirement, and not a warranty.
Relative Hail Impact Resistance of Asphalt Shingles
Insurance Institute for Business & Home Safety (IBHS)
That the IBHS Impact Resistance Test Protocol uses 1.5- and 2.0-inch diameter laboratory-manufactured hailstones informed by documentation of natural hailstone size, density, strength and kinetic energy; and IBHS's own caveat that while these tests do not exactly replicate natural hailstone impacts, they provide a way to compare relative product performance in controlled, repeatable laboratory tests.
Summary page. The detailed test parameters and the product-level results sit in the linked protocol document and in member reports, which this page has not reproduced.
IBHS Releases Its Most Expansive Impact-Resistant Shingle Ratings to Date
Insurance Institute for Business & Home Safety (IBHS) / 19 November 2025
That the 2025 ratings round evaluated impact-resistant shingle products purchased from retail outlets using laboratory-made hailstones, scoring dents, tears and granule loss; that no product earned the top rating of Excellent; and that IBHS does not recommend using a product rated Marginal or Poor in areas that experience hail.
A news release about a ratings round that is updated periodically. This page deliberately does not reproduce the product count or the product list, both of which change; readers should consult the current ratings directly.
Learn how to identify hail damage on roof systems
National Roofing Contractors Association / Professional Roofing (Scott J. Morrison, P.E.) / May 2009
That a roof system is not functionally damaged by hail unless its water-shedding capability is diminished or its expected service life is reduced; that impact-caused fractures start on a shingle's underside, the region of maximum tensile strain, and propagate toward the granule surface, so an impact fracture is always detectable on the underside; that such fractures may or may not be visible on top and are often felt as a localised soft spot called a bruise; that granule loss is normal weathering and is not considered functional damage, and that shingles losing enough granules to expose the coating asphalt without reinforcement fractures are typically already near the end of their service lives; the published threshold table of the smallest hail sizes that typically damage common steep-slope coverings, with its stated assumptions of hard hail, perpendicular impacts, reasonable support, and roof systems in relatively good or midlife condition; that small hailstones are much more common than large ones; that wind imparts a directional component to hailstones but generally contributes only minorly to their impact forces; that spatter marks typically linger about a year depending on exposure and climate; that metals are a permanent record of hailstorms and dent width in air-conditioner fins is typically about 80 percent of the hailstone diameter; and that many professionals and homeowners mistakenly assume a roof needs replacement after a hailstorm.
Trade technical guidance from 2009, written for roofing professionals, drawing on one engineering firm's laboratory and field work. It is not adopted law, it is not an insurance coverage standard, and its damage thresholds are explicitly conditional on the assumptions it states. The second part of the series, covering low-slope roof systems, is a separate article this page has not used.
Severe Weather Climatology (All Hail)
NOAA National Weather Service Storm Prediction Center
That SPC's hail climatology probabilities were estimated from a 30-year period of severe weather reports from 1982–2011; that reports for each day are placed on an 80 km × 80 km grid; and that the annual cycle is smoothed in time with a Gaussian filter of 15 days standard deviation and in space with a two-dimensional Gaussian filter of 120 km standard deviation in each direction. SPC publishes a parallel significant-hail climatology at the same address with the parameter changed.
A report-based climatology, not a measured hail record and not a forecast. Its smoothing is explicitly designed to generalise across 80 km cells, so it cannot describe a specific address, and the report period ends in 2011.
Storm Prediction Center Frequently Asked Questions
NOAA National Weather Service Storm Prediction Center
That the NWS defines a severe thunderstorm as one producing hail at least 1 inch in diameter or larger, and/or wind gusts of 58 mph or greater, and/or a tornado; that SPC defines a significant severe thunderstorm as one producing hail at least 2 inches in diameter or larger; and that SPC's storm reports list is preliminary and likely does not contain all severe weather reports, with the NCEI Storm Events Database providing the official final data.
Definitions and operational practice, not a climate dataset. It does not quantify reporting bias.
Service Change Notice 09-52: National change in minimum hail size criterion for issuing Severe Thunderstorm Warning and Severe Weather Statement products
NOAA National Weather Service Headquarters / 4 December 2008
That effective 5 January 2010 all NWS Weather Forecast Offices changed the minimum hail size criterion for Severe Thunderstorm Warning and Severe Weather Statement products from 3/4 inch (penny) diameter or larger to 1 inch (quarter) diameter or larger, with no change to the 58 mph wind criterion.
A warning-criterion change. It governs when warnings are issued; this page's inference that it puts a step in the long-term report record is an inference from that change, not a statement made by the notice.
Hail Size as Related to Objects
NOAA National Weather Service Storm Prediction Center
The object-to-size conversions used when assessing hail reports — 1.00 in quarter, 1.75 in golf ball, 2.00 in hen egg, 2.75 in baseball, 4.00 in softball — and SPC's own instruction that measurement, not estimation, of hail size is encouraged.
A translation table for reports. Object analogies are approximations and are not a damage scale.
Storm Events Database
NOAA National Centers for Environmental Information
That the Storm Events Database is the official record of NOAA Storm Data, entered by the National Weather Service, covering events including hail from January 1950 onward, with periods of record that differ by event type because collection procedures changed over time; and that NCEI reformats and standardises event types without changing reported values.
A record of reported events. Absence of a report is not evidence that no hail fell, and damage values in the database are estimates entered at the time.
Commissioner's Bulletin B-0030-98: Adoption of amendments of endorsements to the Texas Personal Lines Manual (Commissioner's Order No. 98-0390)
Texas Department of Insurance / Effective 2 May 1998
That Texas adopted mandatory homeowners and dwelling premium credits for roof coverings meeting UL Standard 2218, graduated by class and rating territory, with Class 4 homeowners credits ranging from 4 percent to 35 percent depending on territory (the separate dwelling-form credits run higher, to 46 percent, and are not quoted on this page); that individual roofing units installed on or after 1 January 1999 had to bear a testing-laboratory label showing the classification, manufacturer, year and brand; that Endorsement HO-145, Exclusion of Cosmetic Damage to Roof Coverings Caused by Hail, may be attached only to a policy receiving that credit, must be signed by the insured, and defines cosmetic loss or damage as damage that alters the physical appearance of the roof covering but does not allow water penetration or cause the covering to fail to perform its intended function; that hail damage which does allow water penetration or causes such failure remains covered; that an insurer must inform a consumer of the possibility that cosmetic damage may be excluded; and that the credits programme is solely for obtaining a premium reduction and is not intended to require or create any express or implied warranty by the manufacturer, supplier or installer.
Texas only, insurance rating regulation rather than building code, and dated 1998. Texas rating rules and personal-lines forms have changed since. This page cites it as documentary evidence that this structure exists and how it was written, not as a statement of current Texas practice or of practice in any other state.
Home policies: Replacement cost or actual cash value?
Texas Department of Insurance / Page states last updated 24 January 2024
The worked settlement example reproduced on this page: a house insured for $200,000 with a 2 percent ($4,000) deductible and a $10,000 roof replacement; replacement cost coverage paying $6,000 regardless of roof age; and actual cash value coverage valuing the roof at $8,500, $7,000 and $4,000 at 5, 10 and 20 years, paying $4,500, $3,000 and $0 respectively after the deductible.
A consumer teaching example from one state regulator, with figures chosen to illustrate the arithmetic. It is not a price, not a depreciation schedule any carrier uses, and not a statement about any particular policy.
Insurance may not cover a hail damaged roof
Montana Commissioner of Securities and Insurance (Office of the Montana State Auditor) / 25 April 2023
That split deductibles, under which losses from causes such as hail and wind carry a much higher deductible, are increasingly common and may change from a fixed dollar amount to a percentage of the total amount of insured coverage — with the regulator's example of a 5 percent hail deductible on a $325,000 limit equalling $16,250; that some policies change how claims are paid on older roofs, with hail damage paid at actual cash value rather than replacement cost value once a roof reaches 10 to 15 years old; and the regulator's worked illustration of a 12-year-old roof, $20,000 to replace, a $250,000 limit with a 2 percent split deductible and 50 percent depreciation, leaving the homeowner $15,000 out of pocket.
An opinion column by a state insurance commissioner, using deliberately illustrative figures the article itself calls a likely worst-case scenario. It is Montana-specific commentary, not a rule, and the depreciation assumption in the example is the author's for the purpose of the illustration.
Consumer Connection: Roof Coverage Options
Iowa Insurance Division / 8 July 2024
That the roof is covered under the dwelling (Coverage A) portion of a homeowners policy; that actual cash value settlement pays the roof's depreciated value based on age and condition and will not replace the roof; and that many homeowners policies now include an endorsement addressing how a damaged roof is settled, which may replace replacement cost coverage with actual cash value or set a specific depreciation amount based on the age and type of roofing material.
General consumer guidance from one state regulator. It does not describe any specific policy, and the roof-cost figure it quotes is not used on this page.
Rebuilding After a Storm: Know the Difference Between Replacement Cost and Actual Cash Value When It Comes to Your Roof
National Association of Insurance Commissioners
That replacement cost value coverage pays the cost to repair or replace damaged property without deducting for depreciation while actual cash value coverage pays the depreciated cost; that insurers usually calculate depreciation based on the property's condition when damaged, what a new item would cost, and how long the item would normally last; and the recommendation that a policyholder ask their insurer whether separate deductibles apply to wind or hail damage.
Consumer education from the association of state insurance regulators. It is not a regulation and does not bind any insurer.

Suggest a correction