How often does the power actually go out in Connecticut?

From December 18 to December 20, 2023, one storm delivered 77.9 percent of Windham County's outage time for the entire year. Not just the worst storm of the year. Most of the year, in one storm.

That single number explains almost everything about how power outages work in Connecticut, and why the answer to "how often does the power go out" is stranger than you would guess.

The short answer

Connecticut's grid is fine most of the time. On the headline reliability numbers that utilities file with regulators, PURA's annual reliability reporting shows this state comparing well with the rest of the country, and we will keep saying so because it is true. The catch is what those headline numbers set aside: the major storm days. A quick definition first: a customer in these statistics is a metered account, roughly a household. In 2023, the average Eversource customer's filed outage time was 189.8 minutes counting all days, and 75.6 minutes with major event days set aside. In 2024, a milder year, the pair was 158.8 and 76.9. In both recent filed years, the set-aside storm days held half or more of the total. The storms are not a footnote to the record.

So the honest answer has two halves. Day to day, you will barely notice the grid. And then, some year, a storm parks on your county for three days and the freezer, the well pump, and the furnace all stop at once.

What the federal records show, county by county

The Department of Energy has tracked Connecticut outages for roughly a decade. Here is what eight complete years of that record look like (the federal feed itself has gaps, and 2025 switched to a new county geography, so we compute rates only across the complete years and say exactly which ones on the methodology page):

A major event here means at least 1,000 customers out for at least 2 hours. The duration columns cover recorded events of every size, including small ones, which is why a typical event can run shorter than the major-event definition. Fairfield County sees about one major event a week on average; Tolland County about one a month. And the causes are not all storms: in Fairfield County, 40.5 percent of major-outage customer impact came from non-storm causes like equipment failures. Full figures for all eight counties are on the open data page, and your town's page shows its county record with every caveat attached.

The storm tail is the whole story

Windham County recorded 683,489 customer-hours of outage in 2023 and 249,796 in 2024. A customer-hour is one metered account without power for one hour. Same county, same grid, 2.7 times apart, and the reason fits in one sentence: December 18 through 20, 2023 happened, and nothing like it happened in 2024.

That is what risk looks like on a grid that works. It does not fail a little bit all the time. It fails a lot, occasionally, and the occasions are storms. Which is why we built CheckYourGrid around storm risk rather than around scaring anyone about a grid whose day-to-day record is genuinely good.

What to do with this

If your home would have a genuinely bad week without power, a well pump and no water, a sump pump in a wet basement, medical equipment, a freezer full of food, the useful question is not "is the grid reliable." It is "what is my county's storm tail, and am I set up for it." Your town's page has the record. Storm-driven risk alerts, scored publicly for accuracy, are coming; the waitlist opens on this site first.

Where these numbers come from

Every statistic above is declared in this post's source file and machine-checked against the datasets on our open data page on every site build; if the data changes, the build fails until this post is corrected. Outage records: DOE EAGLE-I, window 2016 to 2024 excluding 2019 (the federal feed has documented gaps in May 2015 and May through August 2019). Filed reliability: EIA-861 utility filings, both with and without major event days. Method, definitions, and limitations: how we compute things.