Data and Decision Analytics

Week 3 Journal

This paper analyzes chapters 5 and 6 of the book superforecasting by Tetlock & Gardner

2015. The most important takeaway I have learned is that it does not require one to be Harvard
      smart to be a superforecaster. Superforecasters are only required to be knowledgeable. What sets
      forecasters apart from superforecasters is that they are knowledgeable people. They use past
      information, leverage it against current situations, and predict what an outcome will be.
      Secondly, superforecasters need to be good with numbers and judgement and, probabilistic
      thinking is a prerequisite to making precise and fine-grained predictions
      Superforecasters do not need to be 100% correct. It is impractical to assume that
      superforecasting means being right all the time. Tetlock & Gardner (2015) note that forecasting
      normally begins with an anchor, which really is just a rough estimate. Even if such an estimate is
      incorrect, it is better than beginning with a big number that has little or no meaning.
      Superforecasters realize that any forecast could be wrong and that it could be changed by the
      information they do not have yet. As such, Superforecaster look for evidence that could prove
      them right and wrong. Naïve forecasters look for information that will prove them right; in
      statistical language, this is known as confirmation bias. Superforecasters look for information
      that cuts both ways. Superforecasters are thorough in their arguments and discussions. They
      approach every discussion as an opportunity to gain new information. Superforecasters question
      the basic beliefs. To become a superforecaster, I must make it a habit to question even
      emotionally charged beliefs; this is the only way to increase the capacity for self-critical
      thinking.
      Superforecasters need to be comfortable with numbers (Tetlock & Gardner, 2015).
      Tetlock & Gardner (2015), in their experience, report that most superforecasters are good in

3
math and science. A lot of superforecasting is interpreting numbers and putting them into
practical use. Even more important than proficiency in numbers is judgment. Judgment is what
tells you how to find relevant information. Proficiency in math informs what to do with such
information. Information and judgment help forecasters to make accurate and reasonable
judgments. Through experience and good judgment, superforecasters learn to interpret
probabilities in ways that other people may not. Tetlock & Garner (2015) find that knowledge of
math and proper judgment are important factors preventing superforecasters from making basic
and elementary mistakes when dealing with percentages and probabilities. This mistake that
most people make is assuming that a high probability means something will happen.
Superforecasters understand the basics of probabilities. They understand that conclusions are
made with certain degrees of confidence and that new information may cause the level of
certainty to be changed. When dealing with probabilities, superforecasters need to adopt
probabilistic thinking, where nothing is certain, nothing is uncertain, but there is an awareness of
irreducible uncertainty. From this, I learned that becoming a superforecaster is having an
awareness of precise and fine-grained probabilities. The awareness of precision means that a 4%
probability has a significant difference from a 3% probability. Probabilistic thinking is, therefore
one way that superforecasters can use to improve the accuracy of their forecasts.
Superforecasters become well versed in their craft when they acknowledge the different
outcomes that some probabilities would take. Probabilistic thinkers move away from pro-fate
statements (certainties) and embrace pro-probability statements.

4

References

Tetlock, P.E & Gardner, D. (2015). Superforecasting: The art and science of prediction.
Retrieved from https://www.redshelf.com