Skip to content

FeatureShingle wind ratings

Shingle wind ratingsIBHS research report, September 2026

After 10 years outdoors, shingles rated for 190 mph winds had failure odds of 0.8% to 99.8% at 90 mph in IBHS tests

IBHS tested nine shingle products rated for 190 mph after 10 years outdoors and put their failure odds at 90 mph between 0.8% and 99.8%. The shingle makers’ trade group disputes the study’s methods.

Fig. 1 | Nine products with the same wind rating ranged from 0.8% to 99.8% failure odds at a 90 mph gust.
ArchitecturalImpact-resistant3-tabProduct F0.8%Product I10.6%Product G14.7%Product A30.6%Product H59.0%Product E60.1%Product C78.8%Product D98.8%Product B99.8%020406080100Chance at least one of 60 shingles fails at a 90 mph gust (%) ArchitecturalImpact-res.3-tabProduct F0.8%Product I10.6%Product G14.7%Product A30.6%Product H59.0%Product E60.1%Product C78.8%Product D98.8%Product B99.8%0255075100Chance one of 60 shingles fails (%)

Chance that at least one of 60 shingles fails in a 90 mph 3-second gust, percent, for products aged 10 years at Richburg, South Carolina. Marker shape shows the shingle type IBHS lists; impact-resistant means a Class 4 rating under UL 2218 or FM 4473, and Product G is also polymer-modified. All carry ASTM D3161 Class F and ASTM D7158 Class H wind ratings. IBHS labels products by letter. Source: IBHS, Roof Aging Farm: 10-Year Wind Performance of Asphalt Shingles, Table 3.

Asphalt shingles that spent 10 years outdoors at a South Carolina test site differed widely in how well they held down in wind, the Insurance Institute for Business & Home Safety (IBHS) reported on September 22, 2026. IBHS is a research group whose members include insurance companies. All nine products it modeled carry the top class under both ASTM wind test standards, and one of those classes means a shingle is expected to perform at 190 mph.

After a decade, IBHS put the chance of a shingle failure at a 90 mph gust between 0.8% and 99.8%, depending on the product. The Asphalt Roofing Manufacturers Association (ARMA), the shingle makers’ trade group, published a review in October that disputes the study’s methods.

1IBHS modeled the odds that at least one of 60 shingles fails

IBHS began building roof sections on what it calls aging farms in 2013, at its research center in Richburg, South Carolina. Member companies later hosted smaller farms in Madison, Wisconsin, and Amelia, Ohio. Every roof faces north and south. Fourteen products from seven manufacturers were tested, and the report labels each one with a letter.

Sections came off the roofs in 2024 and 2025. In the lab, IBHS measured how much of each panel had come unsealed and the force needed to break the sealant bond, the strip of adhesive near a shingle’s lower edge that glues it to the shingle below. It set those results against wind loads measured on a full-size test house in its wind chamber. The result is the chance that at least one of 60 shingle positions fails at a given 3-second gust. The report notes that a typical roof can have about a thousand shingles, which would raise the odds.

The wind ratings on a shingle wrapper come from these two tests.

Fig. 2 | Every product carries Class F, 110 mph, and Class H, 190 mph, and the IBHS model runs from 90 to 190 mph.
ASTM D3161 fan testASTM D7158 uplift testGusts in the IBHS modelStrongest airport gust since 1996Class A, 60 mphClass D, 90 mphClass F, 110 mphClass H, 190 mph90 mph19074.9 mph, June 24, 2010050100150200Wind speed (mph) ASTM D3161 fan testASTM D7158 uplift testGusts in the IBHS modelStrongest airport gust since 1996A 60D 90F 110H 1909019074.9 mph, June 24, 2010050100150200Wind speed (mph)

Wind speeds for the classes of the two ASTM wind tests for asphalt shingles, as the IBHS report describes them: ASTM D3161 puts a shingle panel under a fan for 2 hours, and ASTM D7158 measures the force needed to lift sealed shingles against calculated wind uplift. Filled markers are the classes the nine products carry. The airport gust is the strongest 5-second gust at Philadelphia International Airport from 1996 through October 28, 2026. Sources: IBHS report; NOAA NCEI daily summaries, WSF5.

2Product B, nearly half unsealed, had the highest odds at 99.8%

IBHS calls unsealing the weakest link. Product B, an architectural shingle, had nearly half of its north- and south-facing shingles unsealed after 10 years and the highest failure odds, 99.8%. The two 3-tab products, the flat single-layer shingles with three tabs along the edge, came in at 78.8% and 98.8%, and IBHS says 3-tab shingles are more prone to unsealing than architectural ones. Product F, with very little unsealing, a moderate average bond strength and a narrow spread, stayed at 0.8% at every wind speed.

Product A had relatively high bond strength after 10 years, but its results were spread widely and some of its shingles had unsealed. IBHS put its failure odds at 30.6% and says products like it are open to failures that start at their weakest shingles, because one failed shingle can set off failures across the roof.

The odds barely moved with wind speed. Product H went from 59.0% at 90 mph to 60.1% at 190 mph.

Fig. 3 | From 90 to 190 mph the odds barely moved: Product H went from 59.0% to 60.1%.
ArchitecturalImpact-resistant3-tab0255075100901101301501701903-second gust (mph)Failure chance (%)Product B 99.8%Product D 99.1%Product C 80.4%Product E 61.0%Product H 60.1%Product A 30.9%Product G 14.9%Product I 10.9%Product F 0.8%Product H, 59.0% at 90 mph ArchitecturalImpact-res.3-tab0255075100901101301501701903-second gust (mph)Failure chance (%)B 99.8%D 99.1%C 80.4%E 61.0%H 60.1%A 30.9%G 14.9%I 10.9%F 0.8%H, 59.0% at 90

Chance that at least one of 60 shingles fails, percent, by 3-second gust speed, for each 10-year product from the Richburg aging farm. Labels give the value at 190 mph. Source: IBHS report.

IBHS measured the spread within each product with the coefficient of variation, the standard deviation as a share of the average. It averaged 0.24 for new shingles, 0.33 at 5 years and 0.29 at 10. Most products gained bond strength in their first 5 years and began to decline after that, the report says. Of the nine 10-year products from Richburg, only A, G and I did not lose strength from 5 to 10 years.

Fig. 4 | The spread in uplift resistance averaged 0.24 new, 0.33 at 5 years and 0.29 at 10.
New5 years10 yearsProduct FProduct AProduct HProduct EProduct D (3-tab)Product IProduct BProduct C (3-tab)Product GAverage of 120.240.290.330.00.10.20.30.40.50.60.7Coefficient of variation of uplift resistance New5 years10 yearsProduct FProduct AProduct HProduct EProduct D, 3-tabProduct IProduct BProduct C, 3-tabProduct GAverage of 12New 0.24, 5 yr 0.33, 10 yr 0.290.00.20.40.6Coefficient of variation of uplift resistance

Coefficient of variation of uplift resistance at each age, Richburg aging farm. Higher means a wider spread among shingles of the same product. The baseline for Product G is blank in the IBHS table. The average covers the 12 products in the IBHS table, not only the nine shown. Source: IBHS report.

3Conshohocken sits about 5.4 degrees north of the Richburg farm

The same products aged differently on different farms. For the products installed at both Richburg and Madison, the report says bond strength at Richburg stayed about the same or fell from 5 to 10 years, while at Madison it held steady or rose, and Product B showed the same pattern at Amelia, where it had about a tenth as much unsealing as at Richburg.

IBHS links the difference to sun and heat. Sealant activates at about 140°F, and the report says south-facing roofs at Richburg spend as many as 1,000 hours a year above that, against about half as many on north faces.

Conshohocken sits between the farms. On Census Bureau coordinates it is about 1 degree of latitude north of Amelia and about 5.4 degrees north of Richburg.

Fig. 5 | Conshohocken is about 1 degree north of the Amelia farm and 5.4 north of Richburg.
34° N36° N38° N40° N42° N44° NNorthRichburg34.72° N Aging farmAmelia39.03° N Aging farmConshohocken40.08° N For comparisonMadison43.09° N Aging farmabout 1°about 5.4°

Latitude of the three aging farms with 10-year results and of Conshohocken, degrees north. Source: Census Bureau, 2024 place gazetteer.

4The airport’s strongest gust since 1996 was 74.9 mph, in 2010

The strongest 5-second gust at Philadelphia International Airport from 1996 through October 28, 2026 was 74.9 mph, on June 24, 2010. The airport recorded 112 days with a gust of 50 mph or more from 1996 through 2025, 12 of them in 2025.

Fig. 6 | No gust at the airport since 1996 reached the 90 mph where the IBHS odds begin; the strongest was 74.9 mph.
aStrongest 5-second gust each year (mph)025507510090 mph, lowest gust in the IBHS model74.9 mph, June 24, 2010bDays with a gust of 50 mph or more04812122000200520102015202020252026 through October 28, hollow aStrongest 5-second gust each year (mph)025507510090 mph, lowest gust in the IBHS model74.9 mph, June 24, 2010bDays with a gust of 50 mph or more04812122000200520102015202020252026 through October 28, hollow

a, strongest 5-second gust each year, mph. b, days each year with a 5-second gust of 50 mph or more. Philadelphia International Airport, 1996 to 2026; 2026 runs through October 28, and a few years have one or two days without data. Source: NOAA NCEI daily summaries, WSF5.

IBHS’s model starts at 90 mph, above any gust in that record, so its table gives no odds for the gusts this area sees most. The report’s summary says shingle failures have been observed in storms at wind speeds well below design levels. The weather guide covers wind and hail here.

5ARMA says the study covered 14 products in a market of about 50 product lines

ARMA’s technical review, October 2026

ARMA’s review says the study covered 14 products in a market of about 50 product lines, relied on samples from single production batches and switched its uplift method partway through, from clamping a shingle’s edge to gluing a wooden beam on top.

It says the IBHS lab operates without independent auditing, that IBHS withheld raw data from industry groups, and that one-third of the 10-year specimens kept or gained uplift resistance. It puts 3-tab shingles at less than 4% of today’s residential market.

ARMA urges insurers not to use the study for age limits, including what it describes as a proposed 15-year age cap on IBHS’s FORTIFIED designations.

What the IBHS report says about the method change

The IBHS report addresses the method change. It tested six new products both ways, found no systematic bias between the two methods and nearly identical medians, and says the change, if anything, could make 10-year results look better.

6Both groups tell owners to know which shingle is on the roof

  1. Keep the record

    Both groups tell owners to know what is on the roof. IBHS says knowing the installed product matters for judging risk, and ARMA tells homeowners to keep records of the specific shingle and its product class. Ask that the contract name the manufacturer, the product line and the color, and keep the invoice and a wrapper. The contract guide lists what Pennsylvania requires a home improvement contract to include.

  2. Ask about fastening

    The wind class on the wrapper does not separate the products IBHS modeled. ARMA tells contractors to follow the manufacturer’s installation instructions, including the enhanced fastening some manufacturers specify at eaves and rakes. Eaves are the roof’s lower edges and rakes its sloped edges at the gable ends. Ask whether the quote follows those instructions.

  3. Ask the insurer which rule applies

    If an insurer raises the roof’s age at renewal or in a claim, ask the insurer which rule in your policy it is applying. The Pennsylvania Insurance Department’s Consumer Help Center takes questions, and the insurance claims guide covers the claim process.

7All nine products, with the odds at 90 and 190 mph and the spread at each age

Failure odds for each 10-year Richburg product and the coefficient of variation of its uplift resistance. Source: IBHS report, Table 3 and the variability table.
ProductShingle typeFailure chanceCoefficient of variation
90 mphBar, 0 to 100%190 mphNew5 years10 years
Product FArchitectural0.8%0.8%0.300.300.20
Product IArchitectural, impact-resistant10.6%10.9%0.310.180.30
Product GArchitectural, impact-resistant, polymer-modified14.7%14.9%—0.280.24
Product AArchitectural, impact-resistant30.6%30.9%0.230.340.36
Product HArchitectural, impact-resistant59.0%60.1%0.190.260.37
Product EArchitectural60.1%61.0%0.260.380.34
Product C3-tab78.8%80.4%0.220.370.24
Product D3-tab98.8%99.1%0.280.630.22
Product BArchitectural99.8%99.8%0.130.300.35
Average of the 12 products in IBHS’s table0.240.330.29

A dash marks a blank cell in the IBHS table. Failure odds at 90 mph (CSV) · Odds by gust speed (CSV) · Variability (CSV) · Airport gusts (CSV) · Latitudes (CSV) · Wind ratings (CSV)