Three key ideas I learned from reading fooled by randomness by Nassim Nicholas Taleb.
Key idea 1: Law of large numbers
The first key idea that I gathered from reading his book is that things that happen to us occur at random and any events that deviate from the mean will soon converge back to the mean over time . He stresses on one of the important rules in statistics, where law of large numbers can affect your calculation on finding the mean. Example, calculating a mean of smaller sample size may not be reflective of the population mean. Only through larger samples (Law of large Numbers), will you be able to calculate the mean that is most reflective of the population. Further, selecting sample at random may not be a task for the human to complete due to ‘human biases' of selecting sample at random. This can make forecasting of future and managing risk based on the forecast a rubbish attempt.
In his book, he also refer to ‘rare occuring' (large deviation or outliers) events as the black swan 🦢 events, where it maybe hard to showcase large deviation as non-outliers in a short sample space. But in bigger sample size, the large deviation may prove as another normal data point in the graph that is mistakenly as outliers. An example to this is the pandemic. Over bigger sample sizes, the pandemic is not necessarily a rare occuring events. There was another pandemic that occured just over 100 years ago that devastated the economy on a global scale.
He gave other examples in his book that relates to trading but I think through this, he stresses the point of large number and how we shouldn’t get too worked up on small sample sizes and start to zoom out our vision to the population size.
His writing on population size make me curious to see how Monte Carlo simulations work. Here is a good lecture by a professor at MIT:
Key idea 2: Decision making using expectation
I like how he evaluate decision making using the calculation of expectation value. Example of this is say you have an option of winning $50 with a probability of 70% and loose $60 with a probability of 30%. Will you take that bet? A rational person, would calculate the overall expectation where [$50 × 70% - $60 × 30% =$35 -$18 = $17 ] and take the bet because they are $17 better off. But we may act the opposite because we have in-built heuristic biases where we may be afraid to lose due to the emotional attachment of $60.
Key idea 3: Correlation does not imply causation
Human tend to remember better through stories than viewing events as independent of each other. It is inherently human to establish a causal link in observable events. It is important to be aware of this heuristic biases and not to oversimplify events to be interrelated of each other. The best example accentuating this key point is the study listed in another book titled thinking fast and slow by Daniel Kahneman. In Daniel’s book, he show a study of how human oversimplify the causation link of employees taking more breaks from work and people being more healthy. This is untrue if the breaks involve employees going for smokes every five minutes.
There are many more key ideas one can extrapolate from reading the book Fooled by Randomness. But the three above tend to stick out to me the most (perhaps my heuristic biases is kicking in seeing the ideas get repeated from thinking fast and slow by Daniel Kahneman).
Anyway, I highly recommend reading the book. Get yours today from https://www.amazon.com.au/Fooled-Randomness-Nassim-Taleb/dp/0812975219
