Law of large numbers

The law of large numbers explains why long-run averages are predictable even when single outcomes are not. For example, a few flips of a fair coin might give 70 percent heads, but after 10,000 flips the proportion of heads sits very near $0.5$. Formally, as the number of trials $n$ (the sample size) increases, the sample mean $\bar{x}$ (x-bar, the average of the results) approaches the true mean $\mu$ (mu). It describes long-run behavior, not a force that makes short-run streaks balance out.