It has been widely commented in the yellow and not so yellow press that The Simpsons (the TV series) have repeatedly predicted the future. Comprehensive online articles about it are here and here. If you google "simpsons predict the future" you get millions of hits and videos.

Perhaps the most remarkable "prediction" (at least to me) is Trump as US president (made in 2000!). The latest seems to be the legalisation of Cannabis in Canada.

The question is, why this apparent success?

My guess is that, (i) The Simpsons make a lot of "predictions" (rather, scenario building), (ii) statistically speaking, the proportion of hits is actually very low (not that I have compute it). The apparent "success" is just a cognitive bias.


closed as off-topic by Tim Nov 25 '18 at 21:27

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "This question appears to be off-topic because EITHER it is not about statistics, machine learning, data analysis, data mining, or data visualization, OR it focuses on programming, debugging, or performing routine operations within a statistical computing platform. If the latter, you could try the support links we maintain." – Tim
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    $\begingroup$ +1 because (a) I want to reward questions that are unique and (b) there is likely an apt "model" you could formulate to explain this $\endgroup$ – Mark White Oct 19 '18 at 13:32
  • 32
    $\begingroup$ Did anyone count the times that prediction is wrong? $\endgroup$ – user158565 Oct 19 '18 at 13:35
  • 5
    $\begingroup$ There's also this: en.wikipedia.org/wiki/Donald_Trump_presidential_campaign,_2000 $\endgroup$ – swit Oct 19 '18 at 15:10
  • 6
    $\begingroup$ What if Simpsons actually inspire people to do the things they show? Maybe Trump watched it and then decided he could actually become a president. $\endgroup$ – Gherman Oct 19 '18 at 16:20
  • 4
    $\begingroup$ One of the highest profile executive billionaires of the last 100 years winning the presidency at any time after 2000 doesn't seem incredible when you consider that another business billionaire, Ross Perot, far less well known at the time he began his campaign, took almost 19% of the 1992 vote--after dropping out of the race that July and re-entering. And who thinks Canada ultimately legalizing marijuana is in any way surprising? $\endgroup$ – Chelonian Oct 19 '18 at 18:32

Quick thoughts:

Let's pretend—instead of scenario-building or making jokes—each of their plot lines are actual predictions.

  1. They make a lot of predictions, so their Type I error is very high, but their Type II error is also very low. If every creative choice is making a prediction AND they have been around for decades, then their show is similar to a medical test that almost always says you have whatever disease you are testing for: You will almost never miss a positive case, but you will be telling a lot of people that they have a disease which they do not have.

  2. People probably only consider a subset of scenarios ("predictions") that are feasible. If the Simpsons were visited by aliens, nobody would consider this a prediction—because we know the odds of it happening are very low. So the universe of predictions we are considering is highly correlated with the prior probability that they will come true—this is stacking the deck in favor of the Simpsons.

  3. The writers of the Simpsons are smart people that also live in the same society in which they are making their "predictions." They are trying to be funny, so what they do, in a Bayesian sense, is construct funny situations that are not assuredly going to happen (these are boring predictions) and are not never going to happen (these are absurd predictions). So again we see the prior probability of these things happening stacking the deck toward the Simpsons being correct: If they write about things with a solid-ish probability of happening (like Canada legalizing weed), then we shouldn't be too surprised when their predictions are correct.

  4. There is no time limit on their predictions. This gives us an unlimited amount of "trials" (let's say the unit of analysis is days or elections or news cycles or celebrity careers, etc.), and all we ever need to do is hit truth once and the Simpsons are "correct."

When you consider all of these together, we can see that making a ton of predictions over an unlimited number of trials where you only have to hit once to be "correct," and people define "success" by ignoring Type I errors and shaping the universe of possible as things with only some probability of occurring, and the creators themselves generally make predictions in areas that have some prior probability of occurring—we can get to the conclusion that the Simpsons can "predict the future."

  • 2
    $\begingroup$ Point 3 is particularly interesting. The Bayesian stuff. But naturally, predictions are not always random. Even is "the model" is non bayesian but frequentist, is still an "informed" prediction, afaik. $\endgroup$ – luchonacho Oct 19 '18 at 13:58
  • 4
    $\begingroup$ To expand a little on your point 3 (which is a good one!) - the Simpsons' writers aren't all typically "only" writers. Many of them have technical backgrounds in maths/physics/art/etc. that allows them to pull from a wide field of knowledge, which helps their predictions be a little more accurate and reality based, than "regular" comedy writers might be able to. $\endgroup$ – BruceWayne Oct 19 '18 at 14:41
  • 2
    $\begingroup$ @BruceWayne Actually, given that Trump thought of running for President in 1988 and also 2000 pretty much explains it. See comment at bottom of question. $\endgroup$ – luchonacho Oct 19 '18 at 15:18
  • 6
    $\begingroup$ A side effect of point 2 is that even if we do consider how many of their predictions failed, we are likely to discount the truly implausible ones as not being serious predictions. But on the rare occasion when one of those does come true, we count it as a successful prediction of an event with a very low prior probability. $\endgroup$ – Ray Oct 19 '18 at 15:34
  • 2
    $\begingroup$ Hm. I don't think we can necessarily say there are unlimited trials. The specific example of Donald Trump becoming president could not happen after his death, for instance. $\endgroup$ – jpmc26 Oct 19 '18 at 21:24

Perhaps the most remarkable "prediction" ... is Trump as US president (made in 2000!).

That is indeed a pretty impressive prediction, but it is less outlandish than you are probably aware. In regard to this prediction, it is worth bearing in mind that even by 2000, Trump had established himself as a popular business figure in the USA, and had made well-known forays in entertainment and politics, including an aborted presidential run. During the 1980-90s Trump had made regular forays into political issues and published various newspaper advertisements setting out his views on foreign policy and crime control. In 1999 he sought candidacy as the Presidential nominee for the Reform Party of the USA, but he withdrew from his attempted candidacy in February 2000, citing problems with the party. He indicated that he might run for President in a future election.

Trump was one of the most admired business figures of the 1980-90s. He had made a successful career in New-York real-estate and also commonly featured in popular culture (e.g., in a bit part as himself in the 1992 movie Home Alone II). He was regularly interviewed in the media in relation to political and social issues in New York. In September 1987 he published a major advertisement in multiple newspapers advertising his foreign policy views (see e.g., this NYT article). His spokesman told the media, "There is absolutely no plan to run for mayor, governor or United States senator. He will not comment about the Presidency." In 1989, during a period of high crime in New York, he published another advertisement calling for reinstatement of the death penalty and increases in police. In a 1989 Gallup poll he was listed as the tenth most admired man in America.

Aside from The Simpsons there have been various other items of popular entertainment that featured Trump in his aspirations for the presidency at around the same time. The video clip for the 1999 Rage Against the Machine song "Sleep Now in the Fire" shows the band holding a concert on Wall Street, with a mixture of head-bangers and bankers, and they show one of the bankers holding up a sign promoting Trump's 2000 presidential run (at 1:04 min).

The episode of The Simpsons that you are referring to would probably have been written during the run-up to the 2000 presidential election, and so the writers would have been aware that Trump was a candidate for a minor party. It is likely that the episode was making fun of the fact that he was (at that stage) an outside candidate for a minor party, who had some populist support, but was unlikely to win. Trump's aborted candidacy for President in 2000 foreshadowed his later runs, and even when he withdrew from that attempt, he indicated that he might run again. In January 2000, prior to withdrawing from his presidential run, he released his political book, "The America We Deserve".

I suspect that this episode of The Simpsons was just poking a bit of fun at a possible future with a populist candidate who was running for a small party. However, even back then, it was not a stretch to imagine that Trump would run again for the presidency. It is certainly an impressive prediction, but less so if you know the history of Trump in politics.


Not the answer you're looking for? Browse other questions tagged or ask your own question.