All generations have their beliefs that are considered superstitious. Avoiding sailing on some days was a way of sailors in the past. “Lucky” socks are worn by athletes today. Today millions of people are refreshing applications at certain moments, thinking that the algorithm is ‘testing’ them, or that clicking on one video means that the algorithm will show a different one next.
These are not necessarily irrational beliefs, the nature of the human brain in dealing with uncertainty. The algorithms behind today's online world today seem like magic, making us feel as though we're being addressed on a first-name basis. If we're feeling a certain way and a playlist plays exactly what we're thinking of, it can be easy to take the recommendations for granted, thinking that they're operating on some kind of secret formula.
Randomness, probability and personalization are particularly interesting occurrences in digital environments. When communities like HellSpin Italia are in discussion, they often talk about streaks, timing and patterns. These talks usually are about probability and not certainty, but they also hint at something bigger: a human phenomena we all experience when complicated algorithms collide with human psychology – the ability to form superstitions.
What are Digital Superstitions?
Online superstitions are when people believe that certain Internet activities are going to bring them good or bad luck, without any evidence to back it up.
Digital superstitions are unlike traditional superstitions, that are typically passed down from generation to generation, as it is through repeated usage of technology that they are formed. An individual could think that an update to a page before buying will generate more recommendations, or that the chances of getting to see the wanted content will be higher if opening an app at a certain time of the day.
These beliefs are born out of people's tendency to form patterns. The brain is highly proficient at very rapidly discovering patterns. Patterns were a key way to survival in prehistoric environments. Movement of tall grass could be a predator (sometime the wind).
Today, the same mental system is at work in a constructed world as opposed to a natural world.
Because the brain is easily fooled into believing patterns that aren't there.
A survival mechanism that is pattern recognition.
Why the Brain Easily Believes in Invisible Patterns
Pattern Recognition as a Survival Mechanism
The brain doesn't just process each event individually but continually models what it anticipates in the future. Neuroscientists refer to this as predictive processing – the brain is continually predicting what it will see and then adjusting these predictions in light of what it does see.
This system is extremely efficient most of the time!
From time to time, however, it sees patterns that it won't.
Picture it: You like a video and then get a list of related videos that you are interested in.Suppose that you like a video and then you see a list of videos that are related and that you are interested in. This could be the case even if the recommendation engine wasn't only relying on this one action, but rather your brain might think that you found a secret rule.
That's the start of a digital superstition.
Dopamine and Reward Prediction
The dopamine is sometimes referred to as the "pleasure chemical. In practice it has a much greater part in learning and anticipation.
Surprise and novelty causes particularly high dopamine levels.
This is what makes for a great and engrossing experience on digital devices.
Examples include:
- unexpected notifications
- surprising recommendations
- rare achievements
- limited-time offers
- viral posts
Rewards that are variable are more likely to lead to repeat visits because it may be something valuable that is gained each time.
Being unsure is its own reward!
It's a fact of behavioral economics that inconsistent pay is more effective at developing habits than consistent pay. This works great for the digital platforms, they don't actually make them up on purpose like casino bonus offers, but it's because they want people to interact again.
How Algorithms Reinforce Superstitious Thinking
One of the tasks of an algorithm is to find the patterns in user behavior. So humans start looking for patterns within the algorithms themselves – ironically.
Personalized Recommendations
Recommendation systems are learning systems that work on huge sets of behavioral signals:
- watch time
- clicks
- scrolling speed
- searches
- purchases
- pauses
- shares
Personalization can change over time, so that users see all the changes as significant.
Someone might believe:
I have seen one documentary and all my feed is changed now.
In fact, there could have been hundreds of factors that played a role at the same time.
The system is complex, giving the impression of mystery.
Ranking Systems and Hidden Logic
It's hard for few users to comprehend how one gets millions of reach while the other doesn't get a single view.
This is a fertile ground for speculation for this uncertainty.
Individuals come up with rules like:
- Only post at specific times
- When posts are not successful, they are removed right away
- avoiding specific words
- refreshing before publishing
Sometimes all these strategies coincide with increased visibility and you may end up believing them even when they're not actually magic, but simply statistical.
Everyday Examples of Digital Superstitions
Digital rituals can be found nearly everywhere on the Web.
Social Media Rituals
There are many creators who have pre-habitations before they post. Some do not publish unless they periodically check on the number of engagements. Others, however, think that certain hashtags are always more effective at achieving visibility. The habits serve as psychological reassurances even though it may disregard them, according to the algorithm.
Streaming Platforms
The spectre of "training" the recommender system is a common one that people think they could do, for instance by carefully structuring their viewing sessions.
Some make a conscious effort to end shows that they don't like because they believe that they are sending a negative message.
Other people repeatedly look for the same pages wondering that they can get better recommendations.
Recommendation systems can, of course, learn from users' behavior, but do not always follow the rules that the users think of.
Online Gaming and Probability
When people talk about a 'hot streak' or 'lucky number' on platforms, they share anecdotes and insights, particularly in the context of games they've played or observed. The talks tend to be more about understanding, rather than proving, hidden algorithms, usually based on experiences of random events.
It's not so much a matter of whether or not a particular belief is right. Instead, when they experience repeated outcomes that are not predictable, they are compelled to look for explanations and thus, they are going to turn into telling stories of probability.Instead, repeated exposure to uncertainties will motivate people to look for explanations and thus they will tell stories of probabilities.