In short: in a large enough population with a random distribution of outcomes, you'll eventually have one member of the population who sees "you" (who is picking results at random, but mutually exclusively to individuals) pick the correct result of a sportsgame everytime.
This is the VAERS dataset: because it records any event proximate to vaccination, eventually every event is proximate to vaccination on a long enough timeline with a large enough population. It exists because, yes, they should be looked into...but only if they're occurring at a statistically higher rate then the average rate for the population over the time after vaccination. This is, for example, how AstraZeneca's clotting risk was identified and how the Pfizer myocarditis risk was identified.
In short: in a large enough population with a random distribution of outcomes, you'll eventually have one member of the population who sees "you" (who is picking results at random, but mutually exclusively to individuals) pick the correct result of a sportsgame everytime.
This is the VAERS dataset: because it records any event proximate to vaccination, eventually every event is proximate to vaccination on a long enough timeline with a large enough population. It exists because, yes, they should be looked into...but only if they're occurring at a statistically higher rate then the average rate for the population over the time after vaccination. This is, for example, how AstraZeneca's clotting risk was identified and how the Pfizer myocarditis risk was identified.