The Signal and the Noise: Why So Many Predictions Fail-but Some Don’t
Nate Silver is a statistician, writer, and founder of the FiveThirtyEight blog. He specializes in analyzing baseball and election data and is a firm believer in Bayesian statistics.
In his book, Silver describes the methods of building a mathematical model using probability and statistics. He gives a broader picture of probability and statistics using case studies from elections, baseball, weather, poker, and financial markets. Ultimately the underlying principle to all of these issues and statistical analysis itself is trying to separate out the real signals from the underlying noise.
For those who want some perspective on what we do and don’t know about statistics as it’s applied to real-world problems, Silver provides actionable ways to implement prediction techniques to makes better decisions in investing and in life.
Key Takeaways
- The signal is the truth. The noise distracts us from the truth. More data means more noise in relation to the signal.
- Some situations are particularly difficult to forecast especially when:
- The event is an anomaly.
- You are forecasting a dynamic system.
- Meaning current behavior influences future outcomes.
- There is a lack of theory in the domain.
- Correlation does not equal causation
- Make sure there is logical causation for the mathematical relationship before using it in your predictions.
- We can become better predictors by:
- Thinking more probabilistically and articulating a range of possible outcomes.
- Updating forecast when presented with new evidence.
- Looking for consensus.
- Use Bayes’ theorem or think in a Bayesian way when new information presents itself.
- Bayes’ theorem: the probability of event A given event B has occurred.
If you liked this post, you can subscribe down below. 👇👇