2003MNRAS.343.1131T

We analyse three years (1998-2000) of Optical Gravitational Lensing Experiment (OGLE) observations of microlensing events to place limits on the abundance of planets with a planet-to-star mass ratio q = 10^–3 at distances ∼1-4 au from their host stars, i.e. `cool Jupiters'. We fit a total of 145 events using a maximum-likelihood fit that adjusts six parameters. Each data point on the light curve allows us to exclude planets close to the two images of the source appearing on opposite sides of the Einstein ring of the lens star. We proceed to compute detection probability maps for each event, using Δχ² threshold values of 25, 60 and 100, and combine the results from all events to place global constraints. Our selection criteria returned five candidate events for a planet with mass ratio q=10^–3. Only two of these remained as plausible candidates after three were rejected due to poor data quality at the time of the anomalies. Our results suggest that less than 21n per cent of the lens stars have Jupiter-mass planets orbiting them at an orbital radius of 1 < a < 4 au. Here n ≤ 2 is the number of planet anomaly candidates that are actually due to planets. The data sets presented here were obtained from the DoPhot analysis of the events available at the OGLE website. The main conclusion of this work is that observing time is more efficiently allocated by observing many events with sampling intervals that produce non-overlapping detection zones than using intensive sampling on a small number of events.