A Statistician’s Approach to Coincidences: Part 4

Illustration by Tom Brown
Illustration by Tom Brown

Coincidences emerge in the minds of the beholders. Without a human mind to detect them, most coincidences would not exist.

Cognitive processing errors serve for statisticians like Persi Diaconis and David Hand as a bulwark against the potential meaning of coincidences. Their perspective shows us how our minds help to create meaningful coincidences.

We can pop coincidences into existence by perceiving patterns where there are none. When taken to an extreme, this tendency has a name: “apophenia.”

The Oxford dictionary defines coincidence as “a remarkable concurrence of events or circumstances without apparent causal connection.” We can perceive a concurrence by overemphasizing or stretching the similarities of the events and by selectively remembering events.

Let’s examine these two very common tendencies.

Just how similar is “similar”? Computer software developers are actively seeking an objective answer to this question. But, for now, degree of similarity remains subjective. Human beings are still better than computers at finding patterns and judging similarity.

Sometimes we may stretch similarity beyond what is reasonable to create coincidences out of two or more unrelated events. We see similarities that may not be there because we want the connection to be there.

But what are the limits of “reasonable” similarities? It’s hard for me to clearly say.

Similarity between two specific patterns can be judged on a gradient by human raters, and eventually by computer programming. For now we can be satisfied with knowing that we probably aren’t too bad at discerning similarities and that there will always be someone who will claim that my similarity is not similar enough.

Degree of similarity plays an important role in judging the probability of a coincidence. The more similar the two (or more) events of the coincidence are, the lower the probability of the coincidence.

Let’s say you and a friend meet up and you’re both wearing the exact same shirt and pants bought from the same store. The probability of that happening is lower than both of you wearing pants and shirts that are the same color but different designs. The closer the similarity, the lower the probability.

We select what we see and remember. What else is new? We have to select information from the huge onslaught of stimuli coming at us. To not select is to overload our brains.

We can, and do, selectively remember certain details and then match those details to a current event. If we did not do that, there would be many fewer coincidences. We also would be living in an ever-present now without links to past experience.

Some people overdo this remembering and matching—selecting just the right memory to create the coincidence. Others may be smacked in the face with a coincidence and not notice it.

What other factors could be influencing the probability of the coincidence? This question challenges students of Coincidence Studies to examine the variables contributing to the coincidence beyond the base rates of each intersecting event. For example, actor Mike Myers was visiting famed author, physician, and alternative-medicine advocate Deepak Chopra. As Mike walks into Deepak’s office, he sees a card on the wall. Mike pulls out his own deck of cards, the first one of which is the same card as the one on the wall. Mike is amazed at the coincidence.

This coincidence was perhaps more probable than it seemed to Mike. The deck contained images of Hindu gods. Deepak relies heavily on Hindu ideas for his teaching. Mike knew that. In preparation for their meeting, Mike seemed to want to show Deepak what he knew that might be relevant to their discussion.

While the coincidence seemed amazing to Mike the context of their relationship increased the likelihood of a matching card. However, if you watch the video, you can see that there were many cards in the deck so Mike’s came placing this one, the Goddess of Wealth, on the top lowers the probability.

In summary, the main cognitive errors people make in estimating the probability of a coincidence include: stretching the similarities to make the two elements fit, selectively remembering past events to find a match with a current event, and neglecting the contextual influences that could increase the probability.


Co-authored by Tara MacIsaac a reporter and editor for the Beyond Science section of Epoch Times. She explores the new frontiers of science, delving into ideas that could help uncover the mysteries of our world.



Read more blog posts on The Coincidence Connector »