We update and extend our previous work reconstructing the potential of a quintessence field from current observational data. We extend the cosmological data set to include new supernova data, plus information from the cosmic microwave background and from baryon acoustic oscillations. We extend the modeling by considering Padé approximant expansions as well as Taylor series, and by using observations to assess the viability of the tracker hypothesis. We find that parameter constraints have improved by a factor of 2, with a strengthening of the preference of the cosmological constant over evolving quintessence models. Present data show some signs, though inconclusive, of favoring tracker models over nontracker models under our assumptions.