The Value of Probabilistic Thinking

Clock Icon

4 mins read

  • Tags:
  • #STEM

  • #Quantitative method

  • #Article

Why do I need to know about probabilistic thinking?

How long can I stay in bed and still make it to school on time? How can I one-up Sainsbury’s with tactical meal deal choices?

Our daily lives are studded with probabilistic decisions. In fact, our relationship with probability goes deeper than that. Our brains are hardwired to think in terms of crude probabilities and heuristics. Without these powerful mental shortcuts we would be easily overwhelmed.

Probabilistic thinking is essentially trying to estimate, using some elements of maths and logic, the likelihood of a specific outcome actually happening. If you can learn the tools of high-level probabilistic thinking, your value in the workplace is likely to be huge. All complex commercial, political and technological decisions involve some element of probabilistic thinking. The world is unpredictable, perhaps unknowable. Our best chance in making good decisions is to weigh up different eventualities. Even then we can get things wrong.

As Joe Biden once said: “if we do everything right, if we do it with absolute certainty, there’s still a 30 per cent chance we’re going to get it wrong.”

How can I use probabilistic thinking?

Prediction is a skill which can be improved with practice. Robert Rubin, who served as Bill Clinton’s Treasury Chief, believed that smart decisions started with posing questions you don’t want to ask: what else could happen, what might happen next, what if you’re wrong? The true enemy of probability is certainty. Proven strategies to overcome the desire to reach after fixed conclusions include: thinking in teams, being aware of your own biases and trying to stay open-minded.

If you want to deep-dive into probabilistic thinking, you might have a look at: Bayesian statistics, metaprobability, and the Dempster-Shafer theory.

Which real-world problems can be tackled with probabilistic thinking?

Probabilistic thinking is key to building statistical models. If we want to tackle knife crime, estimate obesity rates, or predict the effects of using different materials in a supply chain, then probabilistic thinking provides a powerful point of departure.

Consider the headline ‘Stabbings on The Rise’. Without high-level probabilistic thinking, you might be genuinely worried because your chance of being stabbed is higher than it was a few months ago. But high-level probabilistic thinking (in this case, a Bayesian approach to probability) means that you take into account what you already know about knife crime – you place this headline in the context of your wider knowledge and circumstances.

You know that violent crime has been decreasing steadily for years and that your city is, statistically speaking, the safest it’s ever been. Last year, 1 in every 10,000 people (0.01%) of people in your city were stabbed. This year, it’s doubled and now 2 in every 10,000 people (o.o2%) are stabbed. Is this a serious cause for worry? When we factor in our prior knowledge, we realise that our safety isn’t massively at risk.

When faced with a decision, a Bayesian asks “What do I already know that might better help me understand the reality of the situation?”

High-level probabilistic thinking means roughly identifying what matters, doing a check on our assumptions, coming up with the odds, and making a decision. In a world as complex and unpredictable as ours, we can never know the future with absolute certainty. But probabilistic thinking is an extremely useful tool to evaluate how the world will most likely look, so that we can strategise effectively.

Notify of
Inline Feedbacks
View all comments