Scales with sqrt(time) for independent returns: a daily Sharpe of 0.1 annualises to 0.1 times sqrt(252) ≈ 1.6.
Estimation error is the underrated fact: the standard error is roughly 1/sqrt(years). Distinguishing a Sharpe of 1 from zero takes about four years of data, so a six-month track record carries almost no information.
The blind spot. Sharpe cannot see tail shape. Selling far out-of-the-money options produces a superb Sharpe for years followed by a catastrophic loss, and the ratio gives no warning. It also penalises upside volatility, which the Sortino ratio addresses.