ABSTRACT. As part of NASA's mission to explore habitable planets orbiting nearby stars, this article explores the detection and characterization capabilities of a 4 m space telescope plus 50 m starshade located at the Earth-Sun L2 point, known as the New Worlds Observer (NWO). Our calculations include the true spectral types and distribution of stars on the sky, an iterative target selection protocol designed to maximize efficiency based on prior detections, and realistic mission constraints. We conduct simulated observing runs for a wide range in exozodiacal background levels (ε ¼ 1-100 times the local zodi brightness) and overall prevalence of Earth-like terrestrial planets (η ⊕ ¼ 0:1-1). We find that even without any return visits, the NWO baseline architecture (IWA ¼ 65 mas, limiting FPB ¼ 4 × 10 À11 ) can achieve a 95% probability of detecting and spectrally characterizing at least one habitable Earth-like planet and an expectation value of ∼3 planets found, within the mission lifetime and ΔV budgets, even in the worst-case scenario (η ⊕ ¼ 0:1 and ε ¼ 100 zodis for every target). This achievement requires about 1 yr of integration time spread over the 5 yr mission, leaving the remainder of the telescope time for UV-NIR general astrophysics. Cost and technical feasibility considerations point to a "sweet spot" in starshade design near a 50 m starshade effective diameter, with 12 or 16 petals, at a distance of 70,000-100,000 km from the telescope.