17. Choose greedily when local choices are enough
You choose the best-looking next step when that choice can be proven safe. Activity selection, interval scheduling, Huffman coding, and coin-style problems show both the power and danger of greedy thinking.