The main reason why this is illusionary is that you can't see it when data is used against you. With a sample size of one it would be perfectly indistinguishable from bad luck or just bad individual performance.