Random: Meaning in Everyday Life and in Science

“That’s so random!” Common phrase used to express something is unexpected or without a pattern. Does random mean something different in science? The short answer is no, the meaning of the word “random” is the same in any context. However, there is a difference in human perception of randomness and the true meaning of random.

Merriam-Webster defines “random” as made, done or chosen at random. Dictionaries don’t differentiate between random in science and in everyday life but there are discrepancies in how this word is used in practice which stems from the fact that people are very bad at recognizing randomness. 

Imagine you want to test your luck and visit a casino. You sit down and start to play a game. Unfortunately, you repeatedly lose and you think to yourself “okay, one more time, I must be close to winning”. This is called a “Gambler’s fallacy”. Researchers in Psychology and Cognitive Science cannot always agree on which psychological concepts exist, but most of them agree the Gambler’s fallacy is a cognitive bias clouding human judgement not only at the casinos. 

Gambler’s fallacy is an irrational belief that probability of a future event is determined by the previous events in a sequence. This can sometimes be a useful heuristic, but it’s not helpful when we consider independent and identically distributed (IID) events such as games at a casino. 

Another issue humans have concerning randomness is the phenomenon called Pareidolia. It occurs when you recognize a known object or a person in a random pattern. Have you heard about those stories of Virgin Mary appearing in a toast or seen farm animals in the shape of clouds? Well, then you get it.  

A field that lives and breathes ”random” is cryptography. When we can detect a pattern in a process, we can also predict how the process might change in the future. That’s very useful in data science and machine learning, but not really what cybersecurity is about. In Serious Cryptography, Jean-Philippe Aumasson describes two mistakes people make when it comes to recognizing randomness 

Which one of these sequences of 0s and 1s is random?

a) 00000000

b) 11010110

Surprise, surprise. They are both random because they have the same chance of being generated. Neither a), nor b) are more likely to occur. But b) can look like it is more random because it has less obvious pattern.

 So what does random mean then? 

Scientific methodologies were developed to minimise human erroneous perception of (non)randomness. Meaning, methodology enforces the statistical definition of random as something unpredictable, i.e. coming from chaos and uncertainty. When researchers are designing experiments, they want to control the environment as much as they can, i.e. minimize the uncertainty. For example, when a new drug is developed, it needs to go through a series of clinical trials. After just one experiment with two groups, one placebo and one taking the new drug, they might find the drug works. But what if the methodology was not followed correctly, the number of study subjects was low etc.? Well, in that case, the results could be observed by pure chance in which case, the new drug would not be helpful.

Going back to cryptography, randomness is not something to avoid. Instead, random means secure so cryptographers need to generate randomness at will. How? First, they need a source of uncertainty in the form of a random number generator and second, they need an algorithm that helps them produce new sequences of random bits from the source of uncertainty, i.e. pseudorandom number generator. 

It might sound as if random numbers should be preferred over pseudorandom numbers, which is true in theory but generating random numbers from the source of uncertainty is very expensive. The uncertainty can come from electrical static or acoustic noise or it can be harvested from a running operating system and the attached sensors or devices. Meaning, the “random” comes from the analog world. It’s not easy to generate random bits fast enough this way. Therefore, the pseudorandom number generators come to the rescue. They can produce many artificial bits from a few truly random bits and so they are very important in practice. 

So no, randomness means the same in different contexts, except we as humans are not good at recognizing what is (not) random. I hope this story has helped you understand what random really means. 

You might also enjoy reading physicist’s perspective on randomness.