Feature of an ideally insurable riskWhy it matters
Many similar exposuresLets the law of large numbers work
Loss is accidental, not intendedRemoves the incentive to cause the loss
Loss is definite and measurableLets the claim be settled
Losses are not all likely at onceKeeps one event from ruining the pool
Chance of loss is neither negligible nor near certainA near-certain loss cannot be pooled affordably

Pure premium = probability × average loss. Gross premium = pure ÷ (1 − loading), when the loading is a share of the gross. More similar, independent risks make losses more predictable.

Key takeaways: compute the pure premium straight from probability × loss; divide, do not multiply, when the loading is a share of the gross premium; a deductible lowers expected claims and so the premium; the law of large numbers makes a big pool of independent risks predictable, while losses that strike together weaken it.