In most statistical analysis, it's useful to throw out the outliers before calculating the results. For example, many latency metrics on websites use the 99th p

In most statistical analysis, it's useful to throw out the outliers before calculating the results. For example, many latency metrics on websites use the 99th percentile or "P99" as a measure of success. If we throw away the top 1% of incomes (and the bottom 1% as well to be fair) how well has America actually recovered from 2008? I can't imagine it looks nearly as good.