How Accurate Is The Old Farmer's Almanac's Weather Forecast?

Looking Back on Our Winter 2020–2021 Forecast

August 20, 2021
Weather Seasons
Vaclav Volrab/Shutterstock

How accurate has The Old Farmers’ Almanac been at predicting long-range weather across the U.S. and Canada? Let’s take a look at the weather predictions from last year (2020–2021).

Every year around Labor Day, The Old Farmer’s Almanac comes out with its long-range weather predictions for the year! Our annual edition includes 12 months of extended forecasts to help gardeners, farmers, weather watchers, and folks from all walks of life plan for the seasons. We cover 18 regions of the United States (and our Canadian edition covers 5 regions of southern Canada).

As well as publishing the year’s extended forecast, we also publish an an analysis of the forecast from the previous year

How Accurate Is The Old Farmer’s Almanac Weather Forecast?

We believe that nothing in the universe happens haphazardly, that there is a cause-and-effect pattern to all phenomena. However, although neither we nor any other forecasters have as yet gained sufficient insight into the mysteries of the universe to predict the weather with total accuracy, our results are almost always close to our traditional claim of 80% accuracy.

The 2020–2021 Weather Forecast

As a reminder, The 2021 Old Farmer’s Almanac predicted that the 2020–2021 winter season would bring warmer-than-average temperatures across the eastern and norther parts of the United States, with below-normal average temperatures limited to the western portion of the nation. 

  • For winter, we predicted that temperatures would be colder than normal in Maine; the Intermountain; the Desert Southwest, and Pacific Southwest regions; and eastern Hawaii and near or normal elsewhere.
  • Precipitation was forecasted to be below normal from Delmarva into North Carolina; in the southern Appalachians, George, and Florida; from the Ohio Valley westward to the Pacific and southward to the Gulf and Mexico; and in western Hawaii and above or near normal elsewhere.
  • Snowfall was predicted ti be greater than normal in the Northeast, Wisconsin, Upper Michigan, the High Plains, and northern Alaska and below normal in most other areas that receive snow.

See this map to understand the regions referenced.

How Accurate Were We Last Year (2020–2021)?

Our overall accuracy rate was lower than most years. Specifically, the accuracy rate in forecasting the change in temperature versus the prevoius winter across all 18 regions of the U.S. was 72.2%, while our accuracy rate in forecasting the change in precipitation was 77.8% This is below are traditional average rate of 80%.

  • Our temperature forecasts were incorrect for the Intermountain, Desert Southwest, and Pacific Southwest regions. We also count the Northeast and Hawaii as “misses,” even though in these regions we came infinitesimally close to predicting the proper temperature trends.
  • In precipitation, we were correct in all regions except for the Southeast, Lower Lakes, Heartland, and Intermountain. 
  • When it comes to snowfall, our prediction that snow would be greater than normal in the Northeast was close, but the actual area was from Connecticut to Philadelphia. We were also close in forecasting above-normal snowfall in Wisconsin, with the actual area from southern Wisconsin to Chicago. Our forecasts of above-normal snowfall in the High Plains and northern Alaska were correct.  
  • Upper Michigan had mostly below-normal snowfall, despite our forecast for above, while the northern Ohio Valley, portions of the Deep South, and portions of Washington and Idaho had more snow than we forecasted. In most other areas, our forecast of below-normal snowfall was correct.

See our table that details predictions and actuals for major cities in The 2022 Old Farmer’s Almanac.

How We Make Our Forecast

We derive our weather forecasts from a secret formula that was devised by the founder of this Almanac, Robert B. Thomas, in 1792. Thomas believed that weather on Earth was influenced by sunspots, which are magnetic storms on the surface of the Sun.

Some people don’t believe that the minute amount of energy the sun bears could have any influence on the atmosphere and therefore the climate and the weather, but we are among those that do. In addition to solar science (the study of sunspots), we rely on climatology (the study of weather patterns) and meteorology (the study of the atmosphere).

Of course, our exact formula is a secret. But rest assured, the meteorological technology and methods are continually updated. While principals upon which the forecasts are made are essentially the same, we use the latest state-of-the-art satellite data, all the latest technology and equipment. Read more about how we predict the weather.

Get Weather Predictions for 2022

How cold and snowy will this winter be? For specifics on snowfall and temperatures in your region, pick up The 2022 Old Farmer’s Almanac—now available to order!



The 2022 Old Farmer's Almanac


Reader Comments

Leave a Comment


Why is 2021 still active? Shouldn't 2022 be active?

Weather forecasts

The Editors's picture

Hi Sandra, In The 2022 Old Farmer’s Almanac, we report on how accurate we predicted the weather the prior winter (so, winter of 2020–2021). If you are looking for this coming season’s predictions (2021–2022), they are available in the new edition on sale now.  You can also find the general winter summary here:


this sucks last winter we in the north east philadelphia did not have a winter should be 20 below here around Sept/Oct/Dec what we get 70 to 80 degree temps this is not cally we got no snow no cold ...that is not the way it should be...hope this almanac is wrong this year and we here in philly get record cold maybe 14 to 20 inch snow storms freezing artic tempertures not summer give us real winter

Lower Lakes Region

Most of the projections for the Lower Lakes were close to the mark, although August turned out to be a near oven with quite not enough rain. Almost had to go buy jelly beans, unshelled peanuts and a spray bottle for water and go outside and yell, "Rain, willya?" at the sky gods, but then that derecho went through here and did so much damage. 15 tornadoes - almost a swarm occurred.
National Weather Service is predicting cooler weather and some rain in a couple of days, right in keeping with your September forecast.
Note: last year, the first snow was on Hallowe'en and no one in my neighborhood went trick-or-treating. I have photos, measured 4 inches on my front steps, where I feed the birds. The last snow of winter was April 30, same as 2018. I don't know if it's significant, but the cold weather did not really turn summery until after July 25. Butterfly and dragonfly count seems to be down in some areas, if that's significant.

Mid Atlantic Region

I thought your predictions were pretty close to what we had. It was warmer than normal (what's normal?) and we had very little snow. In late Dec. I turned the heat down and opened windows.
We had a quick, wet spring which preceded a very hot/humid June/July and beginning of August. It started to cool down a bit after the first tropical storm, which dumped around 8 inches of rain in a few hours.

Almanac weather accuracy

The Editors's picture

Hi, Diane,

Thank you for taking the time to let us know how the forecast worked out in your area.

As for what’s normal—good question; many people wonder. Here’s the answer: 

When you read the weather predictions in this Almanac (or listen to any TV or radio forecast), you may notice the terms “normal” or “average” used to compare current traditions.

Most of us may think of the weather last year or in recent history.  However, averages are based on 30-year periods prepared by government meteorological agencies and updated every 10 years. The most recent tabulations span the period 1981 through 2010. Before that, the base period used by the National Climatic Data Center (NCDC)—based in Asheville, North Carolina—was 1971 through 2000. This involves sweeping up the period’s daily temperature, precipitation, and other weather information from more than 7,000 weather stations around the country and computing new statistics to be rolled out.

The Old Farmer’s Almanac forecasts “deviations from averages,” using the official 30-year normals for temperature and precipitation, averaging each of our designated regions. Technically, we use the normals rather than the averages. (Although the spline-curve modifications affect the daily normals, they have no discernable effect on the monthly normals that we use—so in this case, the normals and averages are actually the same.) 

You can read more here: and watch an entertaining video about it.

Let’s hope that the coming seasons’ forecast is as accurate!

Winter 2019-20 forecast accuracy

For Region 2 (Atl Corr) and Region 5 (FL) you forecast above normal temps and they were but in this article you say you got them wrong.
For Region 15 (Pac NW) you forecast wetter than normal but it was drier.
I don't have time to go though every region to look for more mistakes.


it helpful information fpr weather 2018 to 2019 coming year will be winter snowfal it totaly change a weather

Long-Range View

The view from the summit of Mount Washington, or Katahdin, can be breathtaking; but if you're forecasting at ground level, I'm guessing you meant "a PEEK at 2018–2019 predictions".

Appreciate that you evaluate and report your own performance, by the way!


The Editors's picture

Thanks for giving us the peek and catching our oops, Barry. We appreciate that you evaluate our performance, too! ;-)