A bicentenarian is about to die but I doubt many will notice. Farmers’ Almanac is calling it quits after being in business for 208 years. Its 2026 issue will be its last.
The very word “almanac” emerges in my mind with a thick coat of dust. It was already on its way out when I was growing up: I doubt many youngsters have even heard of it today. Almanacs, which predate newspapers, were handbooks filled with general information, astronomical data, and a calendar for the upcoming year. The origin of the word itself is . It sounds Arabic but isn’t, and may have been invented by European astronomers who wanted a word with a tinge of exoticism. One of the more historically famous almanacs was Benjamin Franklin’s, whose Poor Richard’s Almanack, first issued in 1732, inspired future publications of its kind.
Farmers’ Almanac was first published in 1818—the year Mary Shelley’s Frankenstein was released and the year was sung (it was “Silent Night, Holy Night” in German). The man behind it was David Young, a poet and astronomer, and his goal was to give farmers long-range weather predictions at a time when satellites, atmospheric modelling, and The Weather Channel did not exist.
But 26 years earlier, a similar publication had already been born, one that would get renamed The Old Farmer’s Almanac. In front of me is a print copy of their 2026 almanac, Canadian edition. The 256-page softcover has much more than weather predictions. Recipes, essays, contests, and tables of upcoming astronomical events like eclipses and meteor showers all vie for room. The magazine also has questionable ads, although a disclaimer states that the presence of an ad does not represent an endorsement. Supplements pushed by a Dr. Al Sears raised my eyebrow; an ad for a booklet called “God’s Remedy for a World Gone Mad” is a plea for simply trusting the Bible that a deity will soon come in to fix all of this; and the almanac’s classified section is plagued by ads for psychics, astrologers, and mediums, including one “Doctor Jombo,” an African voodoo doctor who can help you win the Powerball lottery. Allegedly.
The raison d’être of these two almanacs is the nearly week-to-week weather predictions for an entire year. How do they do it, and how did they do it back in 1792 and 1818 respectively? Unfortunately for the scientist in me, it’s all very hush-hush. Farmers’ Almanac employs a secret mathematical formula, while Old Farmer’s Almanac uses its own formula, yes, but supplements it with data from prevailing climate patterns and from our atmosphere… as well as the study of solar activity. More on that strange beast later.
I cracked open my 2026 Old Farmer’s Almanac, which predicts the weather up until October of this year. The thing about predictions is that they can be checked after the fact.
Weather the rubber meets the road or not
I looked at the Old Farmer’s Almanac’s forecast from November 2025 to April 2026, focusing exclusively on Southern Quebec, where I am. Each month comes with a predicted average temperature. I plotted that (in orange), as well as the actual average temperature measured for that month, picking weather stations from the three main cities included in the almanac’s definition of Southern Quebec (they include Ottawa even though it is in Ontario because of the similarity in climate):
As you can see, it’s a pretty good fit… except for February. We all remember the “forever winter” we just survived. Average temperatures in February for Southern Quebec were -9.3ºC, but the almanac predicted a much warmer -2ºC.
Beyond these predicted numbers, the Old Farmer’s Almanac uses vague wording to hint at what the weather might be like. “Sunny, warm,” reads their entry for March 1-6, 2026. Does every day need to be sunny for it to count as a hit? The average temperature in Montreal for this period was -6ºC, with temperatures getting warmer over the course of these six days. Is that “warm”?
Many of the predictions failed quite clearly, though. We were supposed to get a snowstorm between November 26 and 28; Montreal got no snow and Ottawa had heavy rain on the 26th, followed by a mild snow fall. The 19thof December was supposed to be “snowy, bitter cold,” but Montreal reached a high of 12.7ºC, springtime weather!
Farmers’ Almanac, whose for Southeastern Quebec I also looked at, fares no better at a glance. Yes, there are hits, but also many misses. For April 12 to 15, the almanac promised “dry skies in southern Quebec;” the reality is that we received a lot of rain. “Blizzard conditions” were heralded for the end of January, but neither Montreal nor Ottawa had any snowfall for that period. Are “blizzard conditions” the same as an actual blizzard?
This mish-mash should not surprise anyone who knows a thing or two about weather forecasts: you simply cannot predict the weather this far in advance. That’s because what we call weather is the result of chaotic systems that cannot be fully modelled in a computer.
Science-based approximations
Our planet is a vast system of heat distribution that tries to stay in equilibrium. It does so, I am told, by “redistributing heat from the equator toward the poles, and weather is the result of this process.” This explanation comes from Julien Pellerin, who has a Bachelor’s degree in Earth and Atmospheric Sciences with a specialty in meteorology. He works at Environment and Climate Change Canada—the federal department in charge of monitoring the weather and the overall climate—as a Warning Preparedness Meteorologist, and he spent a lot of time dutifully answering my questions on the topic of science-based weather forecasting, as opposed to whatever the fine folks at farmers’ almanacs do.
We may roll our eyes at a weather report when they erroneously predict rain, but these reports’ accuracy has improved over the years. It’s not perfect, though, and the problem is one of computing power. As Pellerin reminds me, we can’t measure and consider every cubic millimetre of air in order to predict with absolute certainty how our atmosphere will behave a month down the road. Approximations have to be used. It’s very hard to forecast daily weather beyond 10 days at the moment, but meteorologists can infer “broad seasonal trends and monthly tendencies” with “some level of confidence,” he tells me.
As an aside, I learned that will be playing a role in these predictions moving forward. No, the folks at Environment and Climate Change Canada are not simply asking Claude or ChatGPT to hallucinate whether or not a tornado will manifest on the 12th of June; they are using two AI models, one based on equations and the other on pattern recognition, with the first influencing the second, and this output is combined with conventional forecasts. The goal is to improve accuracy and to better predict severe weather events (and the paths they will take).
The kind of modelling that meteorology does dates back to the 1960s and 70s, with significant progress having been made in the 1990s, I learned. Proprietary equations developed 234 years ago and  in someone’s office, as with the Old Farmer’s Almanac? They don’t stand a chance.
And it’s not just me saying it. Predictions made by farmers’ almanacs have been tested, and again and again they fail at being accurate.
Will it rain a year from now? Flip a coin
In 1981, the scrutinizing the Old Farmer’s Almanac was published, courtesy of two meteorologists from the University of Illinois in Urbana. They checked predictions made for the entirety of the United States for a 60-month period encompassing the second half of the 1970s, specifically deviations from normal in temperatures and precipitations. Only about half of these predictions were in the right direction, meaning for example that the almanac had correctly predicted warmer-than-average temperatures in a particular month for a particular region, as opposed to cooler-than-average. Half.
You would think the almanac would do well at foretelling extreme weather events, but there too the results were middling. For example, it warned of a summer drought in 1980 for the south-central United States; in reality, it rained much more than usual.
If you know where to look, you can find a handful of meteorologists who, over the years, have published their own comparisons of farmers’ almanac predictions and weather measurements, always to mediocre results. Joel Gratz examined both almanacs for the 2013-2014 period, finding a smorgasbord of —a big one being the continued drought in California, on which the books were mum. Jan Null has done a similar exercise for years, and the results are all over the shop. Peter Goble analyzed Farmers’ Almanac’s predictions for 2019-2024: they were only more accurate than .
There is no peer-reviewed literature on the formulae used by these almanacs because of their “classified” nature. They are, I’m forced to conclude, pseudoscientific. Their custodians invoke mathematics, but the secrecy they are shrouded in and their clear lack of effectiveness put them outside the realm of science.
The Old Farmer’s Almanac’s obsession with sunspots is particularly odd. Studying sunspots is one of three prongs of their proprietary prediction algorithm, but sunspots don’t do much to influence the weather. Sunspots are patches on the Sun that momentarily look darker to us, because of they are cooler due to an increase in magnetic field lines. Julien Pellerin tells me that sunspots and other solar activity do not influence day-to-day weather; the Sun does have an influence on climate in the long run, and eruptions from it do have an impact here on Earth but mostly on our telecommunications systems, power grids, and satellites. Although we have known about sunspots for millennia, astronomer William Herschel (incorrectly) they were connected to temperatures here on Earth, and he did so in the early 1800s… around the time the two farmers’ almanacs came into being. One of the hallmarks of pseudoscience is an absence of progress. Sticking with sunspots when predicting the weather in 2026 sounds like an absence of progress to me.
As for the rest of the almanacs’ calculations, they rely heavily on past trends. When our climate was more stable, that’s not a bad approximation of the coming year, but with catastrophic climate change unfolding right now, “the past isn’t going to always be a predictor of the future for us now,” to quote Sarah Kapnick, the former chief scientist at the National Oceanic and Atmospheric Administration, who was interviewed on this topic for a few years ago.
Farmers’ almanacs fill a void that people abhor: uncertainty about the future. As with psychics, we remember the hits and forget the misses. When an almanac gets it right, we tell our friends. When it fails spectacularly, we shrug and tell ourselves that it’s still a marvel they can predict some things so far in advance.
It’s astrology. My copy of the Old Farmer’s Almanac even has astrological charts that are completely nonsensical. There are specific dates in each month that are best for cutting your hair to discourage growth, to lay shingles, to get married, and to castrate animals. The same goes for planting seeds according to the moon’s phase: in Quebec, planting lettuce is recommended for the entire month of June, but the moon apparently prefers the latter half of that month.
Farmers’ Almanac has bowed out, but the Old Farmer’s Almanac is still going strong. If you want long-term weather predictions, you might as well flip a coin, and you know what? The coin is cheaper than the price of the almanac.
Take-home message:
- Farmers’ Almanac and the Old Farmer’s Almanac use secret mathematical equations initially created over 200 years ago to predict the weather a year in advance.
- When compared to actual weather measurements, these predictions fare no better than flipping a coin, and their use of vague words like “warm” means it’s easy to make the predictions fit reality.
