Allstate Should Pull this Ad and Apologize for Misleading the Public

December 17th, 2018 by Roy W. Spencer, Ph. D.

I’ve been meaning to comment about this TV ad for Allstate insurance, which enraged me the first time I saw it. Allstate knows better (the insurance business deals with probability and statistics) and they knew this was a lie when they put the ad together:

In the ad, actor Dennis Haysbert says:

“A once-in-500-year storm should happen once every 500 years, right? The fact is, there have been 26 in the last decade.”

Setting aside the fact that we don’t have enough statistics to say anything meaningful about what happens over 500 years (hydrologists prefer to stick to 100 years as justified when talking about rare events), Allstate knows very well that such statistics refer to the repeat period for the same location… not for (say) the whole United States. It is not unusual for once-in-100 year weather events to occur more than once, maybe several times, each year somewhere in the U.S.

I consider this false advertising because Allstate knows better, and is purposely misleading the public to make more money.

76 Responses to “Allstate Should Pull this Ad and Apologize for Misleading the Public”

Toggle Trackbacks

  1. Tim Rhyne says:

    I was mad when I saw it, too. But I bet that the obfuscation didn’t register with 95% of the public.

    • Curious George says:

      I deeply distrust all x-hundred-years predictions. When oil companies were building platforms in the North Sea, they positioned the deck safely above a 500-years-wave which was determined to be (I am guessing) 16 metres tall. Two waves damaged the deck in the first year.

      I wonder how the satellite altimetry copes with rogue waves?

    • David Appell says:

      Cut your cable. Then you won’t see any ads at all, and your house will be much more peaceful.

      • Bill F says:

        David! We finally agree on something. Its a Christmas miracle.

      • just a thought says:

        I don’t have cable, so I didn’t see that ad till I came here. And yes it does make my house more peaceful. But that doesn’t alter the fact that they are engaging in an unethical business practice that will harm others and possibly myself. Why shouldn’t I be upset by that?

        The more we tolerate crooks like them, and activists like yourself, the worse they and you will become. By defending them, which is what you are doing however indirectly, you show yourself to be no better.

  2. Patrick says:

    That was the first thing that crossed my mind: “How do they know what happened in 1518?”.

  3. ßri says:

    In the 500 year time scale Communism never existed so that’s good.

  4. Bri says:

    got any predictions for the winter , so far we have had a pretty cold one. I put my beehives in a temporary greenhouse because the late cold snaps killed them last year. I saw an all black woolly worm so that’s not good. LOL

  5. garyh845 says:

    “It is not unusual for once-in-100 year weather events to occur more than once, maybe several times, each year somewhere in the U.S.”

    There are certainly more than “100 locations” in the US – try many thousands.

    Odds seem to be in favor of 100-yr events, someplace . . all the time

  6. Gary says:

    Seriously? When has any advertisement told the truth? I learned at the age of six to be impervious to their inducements. Pity fewer people are.

  7. Entropic man says:

    Let me see….

    If there is a 1 in 500 probability that a particular strength of storm will occur in a location and there are 2.6 storms each year, then there are 500 / 2.6 = 192 locations in the US.

    Put another way each storm occupies 1/192 or 0.5% of the area of the US, or 3.8 million * 0.05 = 190,000 square miles.

    Roughly 490 miles in diameter.

    • Entropic man says:

      From the Texas hurricane handbook

      “What is the Size of a Hurricane?
      The size of a hurricane is usually determined by the diameter of its hurricane- and gale-force winds. The average diameter of hurricane-force winds is 100 miles, while the average diameter of tropical storm (gale) force winds is 300 to 400 miles. However, this varies from storm to storm. For example, Hurricane Carla had a diameter of hurricane-force winds of 300 miles, and diameter of tropical storm (gale) force winds of 500 miles.”

    • Bart says:

      That is 5%.

  8. Bob Tisdale says:

    The ALLSTATE slogan was “You’re in Good Hands with Allstate.”

    Now they’re doing something else with those hands.


  9. Jim Mossy says:

    Unfortunately we are witnessing a total erosion of honesty in US society. Lies and distortions of truth are no longer seen as problematic by a large portion of the country. Our president can’t seem to go a single day without saying or tweeting something demonstrably false, and 42% of the country still support him.

    • Martha says:

      You just lied. But, hold on! There is a catch.

      When Trump invented the lie to become POTUS, he should have patented it and then collected back royalties from previous administrations, media, and people like you, in short, from those that traveled back in time with his malevolent invention and have been using it at every opportunity since eons past.

    • David Appell says:

      You’d think all of Trump’s lying would bother Republicans. I don’t know how they don’t see that their moral standing is approaching zero.

      • Bart says:

        If you like your insurance, you can keep your insurance.

      • John F. Hultquist says:


      • Martin says:

        LOL… approaching zero….. like the Dems, depending on if one wants their private opinion or public opinion published. But we digress, we should have free storms for everyone. We need a fair and decent working climate model that allows all alarmists an opportunity for belief…

    • Gordon Robertson says:

      Jim…” Our president cant seem to go a single day without saying or tweeting something demonstrably false, and 42% of the country still support him”.

      Propaganda is not new to these times. I think the media is as much to blame as anyone else. Trump certainly likes to bs, but so do the media and politicians throughout the world.

      Even our scientific institutions like NOAA and NASA are into it.

      Then there is political correctness which has spread through the Western world like a plague. People into that mind set not only want to agree with each other, they want to suppress the thoughts of those who don’t agree.

  10. Entropic man says:

    By my calculation and Allstates’ figures, your house has a 0.5% chance of being caught in a severe storm each year.

    It also has a 0.3% chance of burning down.

    If you think fire insurance is worthwhile, you should probably have storm insurance as well.

  11. Ossqss says:

    When I saw that commercial, I immediately reached over and hit this!

  12. Dan Pangburn says:

    This whole idea starts off with a misleading concept. What is called a 1 in 100 year event actually means the probability of it occurring in any one of the 100 years is 1%. Similarly, 1 in 500 years means the risk in any one of the 500 years is 0.2%. If the intensity of events is normally distributed, 1% (one sided) is beyond 2.57 standard deviations and 0.2% is beyond 3.1 s.d.

    • David Appell says:

      Yes, they use standard deviations to come up with these numbers. Why is that a “misleading concept,” as you claimed?

    • John F. Hultquist says:

      If the intensity of events is normally distributed,

      These such events are usually described by the Pareto distribution.
      If insurance companies are using the Normal Distribution . . .

      I’ll let you finish that.

    • Craig T. says:

      I think “misunderstood concept” better describes it. A lot of people take “500 year flood” too literal.

    • Gordon Robertson says:

      Dan…the chances of winning the 6/49 Lotto in Canada are 1 in 13 million, yet someone wins it every couple of weeks. It stands to reason that someone will beat those odds if several million people play using different combinations of the numbers.

      I don’t place much faith in odds or luck. The chances of getting hit by lighting are fairly small in Vancouver, Canada for the simple reason we seldom get thunderstorms. If you live on the Canadian prairies, you can almost bet on getting one every day or so.

  13. Correct Roy. And from Allstate it was definitely deception.

    • Entropic man says:

      I disagree.

      My 1.48pm post unpacks Allstate’s figures and gets a probability of 0.3% that a customer’s house wiill be hit by a serious storm in any one year.

      That is comparable to the risk of it burning down.

      The advert does not specify probabilities, but does specifically give false numbers. A rational customer is expected to do their own due diligence from there.

      • Entropic man says:

        I hate this tablet!

        That should be

        “The advert does not specify probabilities, but does NOT specifically give false numbers.”

  14. bk says:

    Put your money where your mouth is and sue them Roy!

    No Money?

    Then get the Heritage Foundation to find it.

    If a bunch of kids can sue the USA Government, whats stopping you and your buddies?

    • Roy W. Spencer says:

      Because I have not suffered damages from their deception. I use State Farm 🙂

      • bk says:

        You state above the damage to yourself (enraged, you will note, about what amounts to an error of semantics, not technical fact) but if you really want to “serve the public good” on this issue, I suggest stepping out beyond the confines of this rather parochial blog to reveal this “outrage” of fake news more broadly.

        Fox News perhaps.

        Tucker Carleson I suggest, quickly, before he is stage hooked off into journalistic oblivion like so many before him!

  15. gbaikie says:

    A once-in-500-year storm should happen once every 500 years, right? The fact is, there have been 26 in the last decade.”

    If you Don’t know he is lying, he saying,
    1 in 500, has recently become 26 * 50 =
    1300 in 500 years. Or in last decade the risk of having a 1 in 500 year storm has become 1300 times more likely.

    And chance of having a 1 in 500 year storm has not increased- or there is zero evidence of any such increase.

  16. Stevek says:

    Technically they are not lying. But deceiving yes.

  17. Stevek says:

    By misleading do you mean people take 1 in one hundred year event to mean that the event only occurs once every 100 years ? Or that if event occurs it is guaranteed not to occur for at least another 100 years ? The general public is pretty clueless and l could see them believing that.

    • David Appell says:

      Consider the set of all daily high temperatures for the day D of the year.

      This set has an average, and a standard deviation. The temperatures will typically be normally distributed (but perhaps not in a changing climate; or perhaps so but skewed with average 0).

      Any particular date’s high temperature is X degrees above (or below) the mean. Divide this difference by the standard deviation to get a unitless number x.

      Calculate the probability p of x using the probability distribution of the normal distribution.

      A 100-yr temperature would have a probability distribution with p>=0.01.

      • Bart says:

        I think you missed the entire point of Dr. Spencer’s discussion.

        Let’s say you had 100 locations with a 1% probability. Then, the probability of at least one event is 1-(1-0.01)^100 = 63%.

  18. martinitony says:

    I saw a show on PBS, perhaps NOVA, several decades ago. It was about the planning of Hoover Dam.
    They had about 40 years of rainfall data at the time, in the 1920s, and believed average annual rainfall produced about 21 million acre feet of water in the Colorado River watershed. However, they were wrong. The actual average, determined later, was more like 14 million acre feet.
    My amounts may be recollected incorrectly, but I believe the 33% error is correct.
    Later, by developing techniques unavailable in the 1920s, they were able to get estimates over more like 500 years by examining tree rings, I think bristolcone pines, perhaps, a technique unavailable earlier.
    My recollections might be incorrect, but the point is that 40 years is a pretty short period in examining climates. Oh yah, those satellites have been up there for about 4 decades also. Perhaps given a little more time we will develop techniques that will tell us more about ice extent and temperatures than we now have an accept as accurate.
    As to 100 year floods and statistics, one doesn’t have to know very much about these statistics to realize some 100 year flood plains won’t flood over several hundred years and some will flood 3 or 4 times in your lifetime. Statistics will confirm these probabilities as routine events in a world this big.

  19. Denis Ables says:


    There are many thousands of events and each could be classified as a 500 year event. Early snow in Amarillo is one. Another is a high summer temp in Amarillo. Now repeat that, using other events such as rain, droughts, tornados, etc., for Amarillo. Then do the same for every city in the US and add every other city in the world. Then throw in various rural areas.

    If 26 such events occur in one year, that would be minuscule compared to the total number of potential events.

    Is that simple enough for you ?

  20. John F. Hultquist says:

    A bit early, but
    Merry Christmas, Roy & all who read here.

    If you are not the Christmas type — be of good cheer anyway.

  21. Milton Hathaway says:

    I think a number of commenters here are mangling the concept of “location”, then trying to apply numbers to that mangled concept. Instead of assuming a location has some areal value, assume it has a value of zero, a single GPS point on a map. This is how I read Dr Spencer’s phrase “repeat period for the same location”.

    But as long as we are mangling, how about this. Weather is like peanut butter getting smeared over the surface of the earth with a knife. The goal is to get a nice even layer, but the weather is spread more or less unevenly, piled up thick in some locations and spread thin in others. The total amount of peanut butter weather is always the same, it’s just constantly being smeared around the planet.

    So how many 500-year weather events can we expect to see every year? I have no idea, but I suspect the answer lies in the peanut butter analogy.

  22. Entropic man says:

    Are storms changing?

    In a warmer, more energetic world,it is probable.

    We’ll need a longer baseline to be sure.

  23. Aaron S says:

    First point. In a big company marketing department can be entirely detached from the core business. So it is likely that the math guys in the company are just as upset as anyone else. Second, some locations have high resolution records extending back thousands of years. Hurricanes for example leave a very distinct layer in an inland marsh with accomodation, but fluvial deposits on a flood plain are not as confident. So frequency estimates vary by region and event.

  24. barry says:

    The outrage against deceptive advertising is so cute!

  25. Craig T. says:

    Only 36% of Houstonians who were flooded by Harvey had flood insurance. I don’t see any harm in Allstate correctly pointing out that living in the 500 year flood plain doesn’t mean your house will only flood once every 500 years.

    I think some people are too sensitive here. Allstate didn’t blame anything on climate change.

  26. DM says:

    First, Merry Christmas a & Happy New Year to Dr. Spencer & all readers:-)

    Second, perhaps 26 500 year storms in a decade should prompt redefining a 500 year storm. Another reason to do so is the high profile hurricanes, tornadoes and N. American blizzards during the past decade were weaker than many of their counterparts during the 1800s and 1900s.

    Third, one might reasonably wonder if heavily regulated State Farm ran this ad to placate federal and state regulators. The green mafia has infiltrated all levels of government and abuses its positions of authority.

  27. aaron says:

    That’s why the phrased it as a question.

  28. CO2isLife says:

    Bingo!!! Great article Dr. Spencer. Just look at property insurance rates and they prove that AllState is misleading the public.
    Property Insurance Rates Prove Climate Change and Extreme Weather is a Hoax

  29. Willful misunderstanding of longer time datasets is even worse. Glacial ice moves incessantly, leaks carbon gases, and supports a lot of mcrscopic life that uses it. Then it de[parts. Who in hell said glaial ice is a leak-proof container for carbon? So the claims and graphs of long past levels are nonsense. Going back, over hundreds of thousands of years, from that leakage they both reduce peak height and also, smooth out.

    Real driver is geomag shifts at the core mantle boundary and above, driven by gravity-induced shifts in the liquid metal ocean currents, mainly at spring tides. That sends location-variable heat up to the surface. Barycentre shifts top to bottom of mantle, lunar monthly. So now we can predict El Ninos and La Ninas way ahead, as the back record is perfect. Solar eclipse across equatorial Ppacific near year end, El Nino. Lunar eclipse, ditto for La Nina. Solar elipse across or near India in June, SW monsoon fails. NASA and WMO stone deaf. BBC also, of course.

    PSR, geologist and social anthropologist (so, amused at my fellow shamans) . 45 and 42 years. Details, anyone? 2,000 page independent report, always gratis, on usb or dvd. [email protected] ph 617 3289 4470 Australia.

  30. Colin Wright says:

    Maybe eight or nine years ago I was asked by the Canadian insurance company I worked for to pull together a response to a request from its British parent company to determine if we were “doing enough regarding climate change.” I met with the senior executives of the company, mostly a bunch of actuaries, to discuss the question and how to respond and then went off and roughed out a draft.

    In my reply I pointed out that as an insurance company we set rates based on historical data and identified trends. We also built in contingency reserves for catastrophic events that we couldn’t predict but knew would occur periodically. I highlighted that the Canadian subsidiary had invested time and money in the incorporating weather data, geographical data, e.g., proximity to flood plains, and other information with traditional historical loss data to build better predictive models. As well, I pointed out that we had undertaken to reduce unnecessary travel and other measures to minimize our overall environmental footprint. I concluded by saying that, given the nature of our business, we were doing enough to address the challenge of climate change.

    After reviewing my draft, an EVP and actuary asked for one change, the addition of a final sentence saying, “but we can do more.” That comment totally undermined the argument I had built and which he otherwise agreed with. It betrayed, I think, a singular lack of understanding of how the insurance industry, of all industries, is probably best able to adapt, and without making grand, symbolic gestures, to a changing climate over time.

    Sadly, the industry now treats catastrophic losses, which have always occurred, as evidence of climate change and as a means of conditioning policyholders and regulators to accept premium increases or limitations on exposure. They cynically use climate change fears, as Allstate does in its ad, as a promotional strategy. Instead of making a potentially useful contribution to the conversation, they attempt to exploit the issue for business advantage. Unsurprising perhaps, but disappointing nonetheless.

    • Gordon Robertson says:

      Colin….”given the nature of our business, we were doing enough to address the challenge of climate change”.

      I am not trying to be facetious, or express denial, but what climate change are you talking about?

      Can you give me an example of climate change anywhere in the world that we have not seen before? What we are seeing in the Arctic was witnessed in the early 1920s and around 1912 or so.

      Warming in the Arctic is not general, it moves around monthly, and the winters are about as cold as they have always been.

      And can you offer proof of what is causing it?

      There is no such thing as a generic or global climate change as inferred. Does not exist. Besides, there has been essentially no average warming since 1998 and no warming produces no climate change.

  31. caliact says:

    Not surprised. As someone who has worked in P&C insurance analytics for over a decide, Ive met modelers and actuaries who believe in catastrophic MMGW, and who support a global warming charge loaded into related insurance premiums. A few years back I had a youg, idealistic boss who tried to prove global warming using decades of company data, then sell the idea to the chief risk officer. The CRO, luckily, was both intelligent AND wise (something too many smart people lack), and didnt bite.

  32. caliact says:

    Gordon, hes not arguing about whether GW is true or not. His point, as I understand it, is that many in the insurance industry are already convinced, including, ironically, the actuaries, who should be motivated principally by solid data/analytics.

    • Gordon Robertson says:

      caliact…”Gordon, hes not arguing about whether GW is true or not. His point, as I understand it, is that many in the insurance industry are already convinced, including, ironically, the actuaries, who should be motivated principally by solid data/analytics”.

      I got that Cal. I was just adding my two bits worth. ☺

  33. Mopar440 says:

    So what, why all the outrage? so called “climate scientists” lie all of the time with their falsified, adjusted, homogenized, cheery-picked data so they can claim CO2 is causing man-made” climate change so they can continue to get huge sums of money to publish their pseudo-scientific garbage. The amount of lying going on about sea level rise, the melting arctic sea ice, extreme weather events, etc. has become sickening.

Leave a Reply