The Anecdotal Fallacy occurs when people ignore statistics and quote a story of events that happened to them. Often, it will turn out not to have even happened to them, but to “someone they know.” While this latter step is an additional move away from constituting useful data, it is not the most malign effect of this bias. The main problem is that assessing probabilities on the basis of personal experiences is almost completely useless even when those personal experiences actually occurred.
There is only one way to assess probabilities, and that is to use statistics of similar event frequencies. This is extremely hard. In fact, even understanding it when it has been competently done by scientists or statisticians is extremely hard. It needs a lot of training and it seems as though our psychology is almost designed to trip us up.
The Anecdotal Fallacy is extraordinary widespread. It’s use seems in many circumstances to be almost automatic. If you give anyone any data on anything at all, people will generally respond with what they think is a counterargument from their own experience. Apparently intelligent and successful people fall into this error, so those qualities are not a prophylactic here. For example, Rupert Murdoch recently tweeted a photo accompanied by the text: “Just flying over N Atlantic 300 miles of ice. Global warming!”
This is a fairly extreme example which may have been deliberately provocative, but in my view it is just quite stupid. The ideas that global warming has to have happened already in all locations and that it would eliminate all ice on the planet betray a non-existent understanding of the problem. The only way to assess the probability that global warming is a genuine threat is to look at graphs showing correlations between greenhouse gas concentrations and temperature rises over several decades.* A personal experience is simply irrelevant to that task.
We also tend to over-estimate the probability of vivid events. I see this as an aspect of the Availability Heuristic, which I think is related to the Anecdotal Fallacy. We use the Availability Heuristic when we assess the probability of events by considering how difficult it is to think of an example of that type of event. Obviously we will make systematic errors in probability judgment if some events are easier to recall than others, and more vivid events are more easy to recall. I discuss this aspect of our psychology in the context of financial markets in my new book:
Why is the Anecdotal Fallacy relevant to the Bitcoin bubble?
Because everyone who is buying Bitcoin is doing so based on one of two events. Either they themselves have recently made a large amount of money from buying it or someone they know has. Twitter is full of stories of people doing so. This is extremely vivid and alluring. It draws more people in, which of course is what helps to sustain the bubble and indeed any Ponzi scheme.
Note that the problem is not that these stories are false. A lot of people have indeed made a lot of money out of Bitcoin. However, it is still a terrible investment — in fact, I don’t think we can even call it an investment — because it has no fundamental value and can crash to zero at any moment. It will definitely do so; we just don’t know when. So the problem is rather that people are using the Anecdotal Fallacy to assess the probability that Bitcoin will rise forever. They do not consider the statistics on bubble which have occurred widely throughout financial history. Any “asset” which rises this quickly has been a bubble which has eventually crashed to zero value. It will do so as quickly as it ascended.
So the statistics are diametrically opposed to our psychology here. Stay away from Bitcoin at all costs.
*The reason I say “several decades” is because we have only been taking detailed measurements for about 150 years. However, we have enough data from ice cores etc going back much, much further, just with bigger error bars.