Law of Large Numbers
Imagine flipping a coin ten times and getting heads eight times in a row. You might feel that the next flip must be tails to balance out the results. This assumption feels natural, but it ignores how random events actually behave over long periods of time. The Law of Large Numbers explains why individual short-term results often look erratic while long-term patterns become predictable and stable.
The Mechanics of Long-Term Stability
When you perform a random experiment a small number of times, the results can vary wildly from your expectations. If you roll a fair six-sided die only six times, you might not see every number appear exactly once. You could roll three ones and no sixes at all, which makes the outcome seem biased or unfair. However, the math changes as you increase the number of trials you perform. As the count of trials grows, the actual observed frequency of an event moves closer to its theoretical probability. This convergence happens because the random fluctuations of early trials get smoothed out by the sheer volume of subsequent data points.
Think of this process like a professional basketball player attempting free throws during a long practice session. If the player takes only three shots, they might miss all of them or make all of them by pure chance. These early shots do not accurately represent their true skill level or their long-term shooting percentage. If they continue to shoot for an entire hour, their success rate will eventually settle near their actual average accuracy. The initial streak of misses or makes becomes a tiny, insignificant part of the total data set. The total average becomes a reliable reflection of the player's performance because the large sample size minimizes the impact of random noise.
Applying Probability to Consistent Outcomes
Key term: Expected value — the long-term average outcome of a random process calculated by multiplying each possible result by its probability.
Understanding this principle allows us to predict the behavior of games of chance with high precision. Casinos rely on this mathematical reality to ensure they remain profitable over the course of thousands of games. They know that a single player might win a large jackpot due to a rare lucky streak during a short visit. When thousands of players gamble over an entire year, the results align almost perfectly with the calculated house edge. The house does not need to predict the outcome of one specific spin of a wheel. They only need to ensure that the total number of spins is large enough for the Law of Large Numbers to guarantee their statistical advantage.
| Trial Count | Observed Heads | Expected Heads | Deviation |
|---|---|---|---|
| 10 | 8 | 5 | 3 |
| 100 | 54 | 50 | 4 |
| 1,000 | 507 | 500 | 7 |
| 10,000 | 5,012 | 5,000 | 12 |
This table demonstrates that while the absolute deviation increases as the number of trials grows, the relative error shrinks significantly. At ten flips, an error of three represents a thirty percent variance from the expected result. At ten thousand flips, an error of twelve represents a tiny fraction of one percent. This reduction in relative error is the core reason why large systems become predictable. It proves that randomness is not chaotic when viewed through the lens of a sufficiently large sample. By focusing on the aggregate rather than the individual, we can remove the mystery from games of chance and see the underlying structure of probability.
The Law of Large Numbers dictates that the average of results from a large number of trials will converge toward the expected value as the sample size increases.
The next Station introduces Gambler's Fallacy Errors, which determines how humans incorrectly perceive these long-term trends.