The concept known as the "hot hand" refers to the belief that an athlete, particularly in sports like basketball, has an increased chance of success after a series of recent successes. This idea suggests that a player who has made several shots in a row is more likely to make their next shot. For many years, researchers largely dismissed this belief as fallacious, meaning it was considered a mistaken idea. The initial and most influential work challenging
the existence of the hot hand was published in a 1985 paper, which laid the groundwork for understanding how people perceive streaks and randomness.
The Groundbreaking 1985 Study
The hot hand fallacy was first formally described in a pivotal 1985 paper titled "The Hot Hand in Basketball" by Thomas Gilovich, Amos Tversky, and Robert Vallone. This study directly questioned the common hypothesis that basketball players possess "hot hands." The researchers defined the hot hand as the claim that a player is more likely to make a successful shot if their previous shot was also successful. Their investigation delved into the human inability to accurately understand randomness and random events, suggesting that a lack of statistical intuition, or innumeracy, can lead individuals to form incorrect assumptions about such occurrences.
To illustrate this point, the 1985 study used the example of a coin toss. Respondents in their research often expected even short sequences of heads and tails to be approximately 50% heads and 50% tails. This expectation highlights a common misconception about randomness: people anticipate that random sequences will alternate more frequently than they actually do. This tendency to see patterns where none exist, or to misinterpret the nature of random streaks, was central to the researchers' argument against the hot hand.
Biases in Perceiving Randomness
The 1985 study proposed two main biases that arise from the thought patterns observed in the coin toss example. One bias is the gambler's fallacy, where an individual might believe that the probability of heads or tails increases after a long sequence of the opposite outcome has occurred. This is the expectation of a reversal following a run of one outcome. The other bias is the tendency for an individual to reject randomness altogether if a streak of either outcome appears too "lumpy" or unrepresentative of a truly random sample, a phenomenon known as the clustering illusion.
These biases contribute to the hot hand fallacy by leading people to misinterpret streaks of success as evidence of a player's increased skill or a temporary boost in performance, rather than as a natural occurrence within random or semi-random events. The researchers argued that people are inherently wired to seek and identify patterns in data, even when those patterns are merely products of chance. This pattern-seeking behavior, combined with an imperfect understanding of probability, fuels the belief in the hot hand.
Experimental Approach and Initial Findings
The 1985 paper employed a multi-faceted approach, conducting several studies to systematically eliminate external factors and isolate the core phenomenon. The first study involved a questionnaire administered to 100 basketball fans from Cornell and Stanford colleges. Subsequent studies analyzed the individual records of players from the 1980–81 Philadelphia 76ers, examined free-throw data, and conducted a controlled shooting experiment. The progression from general fan perception to controlled environments aimed to remove variables such as defensive strategies, shot selection, and game-day distractions.
Across these various studies, the primary finding consistently indicated that the outcomes of both field goal and free throw attempts were independent of each other. Even in the controlled shooting experiment, where external influences were minimized, the results remained the same. The researchers concluded that the subjective feeling of being "hot" did not predict future hits or misses. This initial research strongly suggested that the hot hand was indeed a fallacy, rooted in cognitive biases rather than actual statistical reality.
