We determine life-cycle trading strategies for portfolios subject to the US tax system. Our method employs Monte Carlo optimization. It accommodates long horizons (between forty and sixty years) and large numbers of trading periods (eg, 480), while accounting for the full cost basis history of the portfolio’s stock holdings, thus sidestepping the curse of dimensionality. We present many new results that provide insights into questions about taxable portfolio investing which were previously unexplorable. Some of our results challenge current conventional wisdom. For instance, we establish circumstances where raising the allocation of stock is optimal though counterintuitive and demonstrate the suboptimality of the 5/25 rebalancing rule, even as a rule of thumb.