Because the word "bias" pertains to whether a sample is representative or not. It has to do with probabilities. GGGP isn't saying that long-term successful out

Because the word "bias" pertains to whether a sample is representative or not. It has to do with probabilities. GGGP isn't saying that long-term successful outcomes are the probable or most likely result of being large at one point in history, so your criticism is inapposite. What he's saying is that, contrary to the OP's seeming position, a paradigm shift isn't always fatal to established players, i.e., "doom articles are a dime a dozen."