Scholarly estimates and opinions of the sailing performance of ancient Pacific canoes vary widely. This paper measures performance by testing real sails in a wind tunnel and hulls in a towing tank. The sails were three East Polynesian Oceanic spritsails of late eighteenth century type, held by the British Museum, collected from New Zealand, Tahiti and Hawaii/Marquesas, which conform to the first historical records. Also tested was a hypothetical generic ancestral sail, and the Māori sail was tested in different ways to accommodate different views. Tests of hull form found that upwind sailing performance improved as underwater hull profile changed from U-shape to V-shape and some archaeological hulls can be assigned to this scale. Velocity prediction programs (VPPs) were calculated for a range of different canoes and simulated voyages by the fourteenth century AD archaeological canoe (waka) found at Anaweka, New Zealand retraced real voyages made by the experimental Polynesian replica canoe Hōkūle'a between 1980 and 2000, in the same recorded weather. Both canoes could average speeds of up to four knots and sail upwind at 75°to the true wind angle (TWA), as proposed by Lewis and Finney. The paper identifies a package of technological innovations involved in the settlement of East Polynesia following the "long pause" in Pacific settlement in West Polynesia. Two innovations previously suggested by linguistics were the Oceanic spritsail and the double canoe, and a third was the development of complex composite planked hulls and V-shaped underwater hull forms. East Polynesian canoes were capable of two-way voyaging and some migrations were planned, as in the case of New Zealand.