Every aging roof eventually asks the same question after a leak or a windstorm, and roof repair vs replacement is decided well or badly depending on whether the answer weighs one variable or three: the roof’s age against its material’s real lifespan, whether the problem is localized or systemic, and whether the surrounding shingles can even survive being worked on. This comparison lays out the framework, the special cases, and the timing logic.
The framework: three questions
One: where is the roof in its life? Materials fail on knowable schedules, 15 to 20 years for three-tab asphalt, 20 to 30 for architectural, per the lifespan reference, and a roof’s age locates it on that curve. Damage at year 8 argues repair; the same damage at year 22 is the curve talking, and money spent there rehearses for the replacement rather than deferring it.
Two: is the problem localized or systemic? Localized failures have addresses and causes: a cracked vent boot, pulled counterflashing, a wind-stripped patch on one exposure, a valley detail, a skylight’s flashing kit, and each repairs durably at modest cost, the leak hunt supplying the address. Systemic evidence reads across whole slopes: granule loss showing bald patches broadly, curling silhouettes along the ridges, hail bruising at claim density, and leaks recurring at new locations each season, the pattern that says the field itself, not a detail, has retired.
Three: can the roof survive the repair? Asphalt’s late-life brittleness is the practical veto: repairs require lifting and re-nailing surrounding courses, and shingles that crack when raised turn every fix into three new defects, the property that makes some old roofs technically unrepairable regardless of budget. A roofer’s hand on the shingles answers this in a minute, and “they’re breaking as I work” is the sentence that ends the repair conversation.
The special cases
The layering question: an overlay, new shingles over old, saves tear-off cost and is generally code-legal once, but it conceals the deck’s condition (soft sheathing found during tear-off is exactly what overlays never find), runs the new layer hotter and shorter-lived, complicates every future flashing detail, and doubles the eventual disposal. It is a legitimate budget tool with those costs stated, and a corner to refuse where the budget allows honesty. The bundling logic: replacement is the once-per-decades moment when everything else on the roof renews at marginal cost, all flashing new rather than reused, ice membrane at eaves and valleys, ventilation corrected, skylights swapped on their 20-year clock, and gutters and drip edge coordinated, which is why a replacement quote missing those line items is a question list, not a bargain. The insurance case: storm damage runs on the claim’s evidence rules rather than this framework, with the deductible and the roof’s depreciated value setting the arithmetic.
Timing the one you choose
The strategic point of the whole framework is choosing the date: a replacement scheduled at year 21 on a dry week, with a vetted contractor’s full crew, beats the same replacement at year 23 in a storm season’s rush, tarps on the deck and the interior drying out. Repairs buy time honestly when the framework says repair, chronic small spending on a brittle roof buys rehearsals, and the interim ownership of a late-life roof is vigilance: the gutter granule ledger, the post-storm binocular pass, and the attic check that catches the next leak while it is still a detail.
The bottom line
Young roof, nameable defect, flexible shingles: repair it and expect the fix to last. Old roof, slope-wide symptoms, shingles that crack in the hand: the field has retired, and the honest spending is the tear-off, flashing and membrane and ventilation renewed with it, on a date the calendar chose rather than the weather.