Which mathematical principle allows insurers to predict aggregate losses more accurately as the number of similar exposure units in the pool increases?

Why this is the answer

The Law of Large Numbers is the statistical foundation that lets insurers price coverage. As the number of homogeneous exposure units in a pool grows, the variance between actual aggregate losses and predicted aggregate losses shrinks. This is why insurers seek large, similar risk pools rather than insuring a handful of unique risks. Utmost good faith, indemnity, and adverse selection are insurance concepts, but they are not the statistical principle producing predictability.