Bayesian statistics

Can you predict the future?

Imagine you're trying to guess the number of jellybeans in a jar without counting them. You know the jar was filled in 10 batches, and each batch had around 50 jellybeans.

You start with a guess based on past experience (like a prior belief). Each time you guess and learn whether you were right or wrong, you adjust your guess for the next time (updating your belief).

Example

You guess 500 jellybeans (prior), then after a few guesses, you adjust to 520 jellybeans (updated guess).

Remember this

You're using past knowledge and new information to make better guesses over time. This is the essence of Bayesian updating.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

Markov chain Monte Carlo

MCMC samples from complex posterior distributions

Real options valuation

Why can't we always predict the future in business?

List of Call the Midwife episodes

Older financial institutions likely to survive longer

Recency bias

Recency bias overvalues recent events in decision-making

Glossary of logic

Ever wonder how scientists decide if their findings are just by chance or real?

Treynor ratio

Treynor ratio measures excess return per unit of systematic risk

Educational content, not financial advice.

Swipe through 100 ML concepts daily

Open Pocket Polymath