That is true for any distribution. The mean (expectation) of the part after X+epsilon has to be larger than after X. There is no possible age distribution where

That is true for any distribution. The mean (expectation) of the part after X+epsilon has to be larger than after X. There is no possible age distribution where your claim isn't true.