16 Randomness versus causality

 

This chapter covers

  • Laplace’s rule of succession
  • The hot hand
  • Recognizing randomness versus order and causality
  • Simulating coin flips
  • Inserting ggplot2 objects into other ggplot2 objects

Way back in the 18th century, a French scholar and polymath by the name of Pierre-Simon Laplace developed a formula to compute probabilities starting from low observation counts. Assume there have been n independent trials that can only result in success or failure; we might then compute the probability (p) of success (s) on the next trial (n) by applying the following formula: p = s / n. However, let’s say there have been just five independent trials and five successes. The probability of success on the sixth trial would therefore equal 100%; alternatively, if there had instead been five failures, the probability of success would equal 0%. The traditional or customary way of computing probabilities doesn’t make an allowance for a different outcome where and when the observation count is low and the opportunity for variance is relatively minimal, and therefore not terribly meaningful.

16.1 Loading packages

16.2 Importing and wrangling data

16.3 Rule of succession and the hot hand

16.4 Player-level analysis

16.4.1 Player 1 of 3: Giannis Antetokounmpo

16.4.2 Player 2 of 3: Julius Randle

16.4.3 Player 3 of 3: James Harden

16.5 League-wide analysis

Summary