if you can find the spontaneous rates of yachts sinking with eventually dead or missing people and spontaneous rates of people getting hit by a car and dying, you can calculate the likelihood for 2 such unordered events appearing within a surprisingly short amount of time.

obviously the p-value for the null hypothesis that this is just random coincidence is miniscule in this case.

in the long run formal verification and statistics will publically catch up with all these shenanigans.