Randomness, and its omnipresence in everyday life

What is randomness?

Everyone has experienced randomness in many different situations in life. This could be when playing a board game, looking at radioactive isotopes, or tossing a coin. Random events occur all of the time whether we notice them or not and influence the outcomes of nearly everything we do. Some "random" events can be predicted if you know enough details. For example, you could theoretically predict if a coin will land on heads or tails if you knew enough facts like the speed the coin was flipped and the rate the coin was spinning (well until it hits the floor) (Matthews, 2020). However, some things are truly random and there is no way to calculate their outcomes regardless of what you know. Finding random events in board games is relatively easy but finding them in real life is a much harder task. Differentiating between what is random and what is predictable is important and could save you time, money, and even help you better understand the world.

Predicting the unpredictable

Professionals of all fields make predictions on a daily basis from stock traders to football pundits. Many make a career out of predicting results and many listen, but what if the modern world is not predictable. At all.

Stock predictions from companies like Bloomberg and the World Financial Review are used by experienced traders and amateurs alike, but how accurate are the experts and should we be putting money on the line based on these "expert opinions". Based on an article from Financial Mentor by Todd Tresidder, the answer is no. "It is very important to separate what is knowable from what is not and to be able to tell if something is useful financial information or a waste of time," he says. The information is not only harmful because it is incorrect but because "it appears credible, causing you to factor it into decisions, but it has no basis in fact" (Tresidder, 2020).

Forecasting tools such as graphs have been used for a long time to try and predict where prices are headed however using this software leaves you just as susceptible to making wrong decisions. Finding trends in the market has been the go-to method for many professionals and large companies. These findings are used as the basis of many critical decisions and the allocation of large sums of money can be decided using this information. But what if the reasons that we give to these trends such as a decrease in supply, a change in consumer habits, or the increasing dominance of the tech sector are just fabricated to help us try and understand a world that is fundamentally impossible to understand.

Comprehending the incomprehensible

We are constantly creating stories to try and explain everything we see and experience but sometimes the correlations and patterns we see are not caused by the reasons we automatically assume. The better the story that we create sounds, the more convinced we are even if we have very little infomation. We can also become less convinced of a hypothesis after we have been given more information. This is because new information might contradict the original story that has been fabricated and instead of making you re-evaluate , just makes you less sure (Kahneman, n.d.). Most people automatically attempt to explain everything and cannot accept that somethings cannot be predicted or controlled.

Cause and effect

The problem of correlation vs causation is clearly described in the article "Why eating ice cream is linked to shark attacks" by Eric Siegel. Data collected from days at the beach show that when the number of ice creams sold increased, the number of shark attacks that took place also increased displaying a correlation between the number of ice creams sold and the number of shark attacks. These events are obviously not linked (the sale of ice cream causes sharks attacks or shark attacks cause the sale of ice cream) so what is going on? This correlation, although backed up by the data, does not prove causation (that one event increasing causes the other to increase). The real reason behind both the ice cream sales and the shark attacks, is an increase in people visiting the beach (Why eating ice cream is linked to shark attacks, 2020). Explaining this example is quite easy as the linking factor is obvious but most real-world examples the reasons are not so trivial.

Drawing the conclusion that all correlations mean causation can have disastrous consequences. Clinical trials are important and are used to make life or death decisions about medications and treatments. Millions of pounds could be invested in a pill or treatment that is ineffective based on an act of randomness or hidden bias.

As I mentioned earlier, differentiating between true randomness and events that can be assigned an explanation is very important and the key to this is looking for causation. The importance of causation is covered in an article from the New Scientist by Ciarán Gilligan-Lee. The article talks about an algorithm from the 1990s that was trained on hospital admissions data and made a surprising prediction, "It said that people who presented with pneumonia were more likely to survive if they also had asthma " (Gilligan-Lee, 2020). This prediction disagrees with all modern science and was universally opposed. So was the prediction an error in the algorithm, a problem with the data, or a medical revelation?

It turns out that "doctors treating pneumonia patients with asthma were passing them straight to the intensive care unit, where the aggressive treatment significantly reduced their risk of dying from pneumonia"(Gilligan-Lee, 2020). Without knowing this fact about how the hospital was run, it is perfectly reasonable to believe the prediction or at the very least carry out further research

How to deal with randomness?

Finding correlations and patterns in everyday life is something that humans are naturally good at but we disregard the most important part. Finding causation. Without causation, there is no way to tell if a correlation is a completely random occurrence or a useful insight. Resisting the urge to skip finding causation is difficult but important as biases and inaccuracies can come from anywhere, from the way you collect data to the people you are collecting data on. One famous example of this is the "Hawthorne Effect". This effect is talked about in the article "10 Correlations That Are Not Causations" by Nicholas Gerbis.

In a series of experiments from 1924-1932, researchers studied the worker productivity effects associated with altering the Illinois factory's environment, including changing light levels, tidying up the place and moving workstations around. Just when they thought they were on to something, they noticed a problem: The observed increases in productivity flagged almost as soon as the researchers left the works, indicating that the workers' knowledge of the experiment, not the researchers' changes, had fueled the boost. Researchers still call this phenomenon the Hawthorne Effect. (Gerbis, 2020)

Citations:

Tresidder, T., 2020. The Great Financial Forecasting Hoax: Why Stock Market Predictions Are Dangerous To Your Wealth. [online] Financial Mentor. Available at: <https://financialmentor.com/investment-advice/investment-strategy-alternative/financial-forecasting-hoax-stock-market-predictions/18251>.

Big Think. 2020. Why Eating Ice Cream Is Linked To Shark Attacks. [online] Available at: <https://bigthink.com/correlation-causation?rebelltitem=1#rebelltitem1>.

Gilligan-Lee, C., 2020. Correlation Or Causation? Mathematics Can Finally Give Us An Answer. [online] New Scientist. Available at: <https://www.newscientist.com/article/mg24632790-700-correlation-or-causation-mathematics-can-finally-give-us-an-answer/>.

Matthews, R., 2020. Are Coin Tosses Really Random?. [online] BBC Science Focus Magazine. Available at: <https://www.sciencefocus.com/science/are-coin-tosses-really-random/#:~:text=While%20a%20coin%20toss%20is,coin%20tumbling%20in%20the%20air.&text=So%20the%20outcome%20of%20tossing,air%2C%20or%20allowed%20to%20bounce.>.

Gerbis, N., 2020. 10 Correlations That Are Not Causations. [online] HowStuffWorks. Available at: <https://science.howstuffworks.com/innovation/science-questions/10-correlations-that-are-not-causations1.htm>.

Kahneman, D., n.d. Thinking, Fast And Slow.