AHJ Radar · Mason Mathews Facility Assurance, LLC
Sprinklers, smoke alarms and the fires that kill: what the data says
Three official datasets say the same thing from different directions. Sprinklers work when they are on: the most common reason one failed was that the system had been shut off. The fires that kill several people at once happen between midnight and 6 a.m., spread past the room of origin 93% of the time, and occur in buildings with no sprinklers 91% of the time. And deaths in nonresidential buildings rose 70% over ten years while the fires themselves rose 19%.
Sprinklers: reliable, and defeated by the valve
NFPA, structure fires large enough to activate the system, 2015 to 2019 annual averages. NFPA may have published a newer edition since; the 2021 report is the one we could read.
| Measure | All structure fires |
|---|---|
| Sprinklers operated | 92% |
| Effective when they operated | 96% |
| Operated and effective | 88% |
| Civilian death rate per 1,000 fires, with sprinklers vs none | 89% lower |
| Civilian injury rate | 27% lower |
| Fire confined to the object or room of origin | 95% with sprinklers; 71% with none |
| Why sprinklers failed to operate | Share of failures |
|---|---|
| System shut off | 57% |
| Manual intervention defeated the system | 18% |
| Lack of maintenance | 10% |
| System components damaged | 9% |
| Inappropriate system for the type of fire | 6% |
NFPA's sentence: "The most common reason that sprinklers failed to operate was the system being shut off at some point before the fire." Where sprinklers operated but did not control the fire, water did not reach the fire in 50% of cases and not enough water was discharged in 31%.
NFPA 25 in the Citation Decoder · Recalled sprinkler heads: CPSC's September 2026 notice
Multiple-fatality fires in residential buildings
U.S. Fire Administration, NFIRS data 2021 to 2023. "Multifamily" includes condominiums and apartment buildings.
| Measure | Figure |
|---|---|
| Fires per year, deaths, injuries, loss | About 200 fires; 720 deaths; 125 injuries; $42 million |
| Share of all residential fire deaths | 26%, from less than 0.1% of residential fires |
| Property type | One- and two-family 81%; multifamily 11% |
| Smoke alarms, occupied buildings | Present 27%; not present 20%; undetermined 52%. Where present, operated in 47%, failed in 17% |
| Automatic extinguishing systems | Not present in 91% |
| Fire spread beyond the room of origin | 93%, against 80% of single-fatality fires |
| Time of day | Midnight to 6 a.m., 53%; peak 2 to 3 a.m., 12% |
| Leading causes, where determined | Other unintentional or careless 24%; under investigation 19%; intentional 13%; electrical malfunction 12% |
Nonresidential buildings: fewer injuries, more deaths
| 2023 estimate | Ten-year change, 2014 to 2023 |
|---|---|
| 110,000 fires | Up 19% |
| 130 deaths | Up 70% (up 93% excluding the 35 deaths of the 2016 Oakland fire) |
| 1,200 injuries | Down 14% |
| $3,164,400,000 loss | Up 10%, inflation-adjusted |
Residential buildings: the 2024 estimates
U.S. Fire Administration national estimates from NFIRS data; the pages carry a review date of September 28, 2026. "Residential" includes one- and two-family homes, apartments and condominiums together; USFA does not split multifamily on these pages.
| 2024 estimate | Ten-year change, 2015 to 2024 |
|---|---|
| 343,000 fires | Down 8% |
| 2,975 deaths | Up 9% |
| 9,225 injuries | Down 13% |
| $11,561,100,000 loss | Up 25%, inflation-adjusted |
| Leading causes, 2024 | Fires | Ten-year change |
|---|---|---|
| Cooking | 160,300 | Down 15% |
| Other unintentional or careless | 34,400 | Up 40% |
| Heating | 27,200 | Down 26% |
| Electrical malfunction | 26,400 | Up 5% |
Fewer fires, more deaths: the pattern of the nonresidential table above repeats in homes. Cooking accounts for 160,300 of the 343,000 fires. USFA's companion nonresidential pages for 2015 to 2024 were refreshed the same day; our monitor could not open the summary page, so the nonresidential table on this page is still the 2014 to 2023 edition (110,000 fires, 130 deaths in 2023) until it is read.
What this means for a hotel or a condominium
| The number | The habit it argues for |
|---|---|
| 57% of sprinkler failures: system shut off | A written impairment procedure (tag, notify the alarm company and the fire department, fire watch, restore and log) and weekly or monthly control-valve inspection under NFPA 25, with the record kept |
| 18%: manual intervention | Nobody closes a valve to stop water damage without the impairment procedure; the training record shows who was told |
| Smoke alarm present in only 27% of multiple-fatality fires; failed in 17% where present | In-unit alarm checks at turnover and on a schedule, and the dated record of them, in every condominium and extended-stay property |
| 53% between midnight and 6 a.m. | Night staff know the alarm panel, the elevator recall and the evacuation plan; the drill record says when they last practiced |
| 160,300 cooking fires in residential buildings in 2024, out of 343,000 | In a condominium or an extended-stay property, the range hood, the in-unit extinguisher or suppression device and the smoke alarm nearest the kitchen are where the count is held down; the fire marshal sets what is required in each unit |
| Nonresidential deaths up 70% in ten years | Occupancy and egress are not a one-time certificate; the annual fire door, emergency lighting and exit inspections are where that number is held down |
Intervals and who performs each inspection are set by NFPA 25, NFPA 72 and the fire code your authority has adopted; confirm them with the fire marshal and a licensed contractor.
What is not on this page, and why
NFPA's newer reports (High-Rise Building Fires 2019 to 2023, November 2025; Fire Loss in the United States 2024, February 2026; Structure Fires in Hotels and Motels) could not be read by our monitor: nfpa.org serves them to a browser only and asks automated readers to stay out of its document store. Their figures will be added when read directly. NFPA code text is not quoted; the standards are paraphrased.
Get the free digest
AHJ Radar™ is a trademark of Mason Mathews Facility Assurance, LLC. Official sources read daily since September 2026; every entry names its source. Content and compilations © 2026 Mason Mathews Facility Assurance, LLC, for personal and internal use under the Terms of Use; no redistribution or resale.
Reference only. This page restates a published official document; it is not legal advice and not an assessment of any property. Confirm requirements with the official text linked above, your authority having jurisdiction and your own counsel. Mason Mathews Facility Assurance, LLC · all reference pages · how our monitor reads sources.