> if we recalculate the odds in terms of a particular in-group, and tell you that, of the 25 people you know and care about, there's a 33 percent chance that on

> if we recalculate the odds in terms of a particular in-group, and tell you that, of the 25 people you know and care about, there's a 33 percent chance that one of them will die, that seems like a large risk, and one you would work hard to avoid. Isn't that 25% chance that someone will die of the personal cohort?