Electrical resistivity tomography (ERT) allows for fast, non-destructive, and efficient sediment characterization and geotechnical process monitoring in the field as well as in laboratory applications. Besides the spatial distribution of resistivity, specimen geometry and electrode configuration determine the electrical potential distribution and the ensuing spatial resolution in the tomogram. We examine potential and current density distribution in various ERT system configurations using both experimental and numerical methods and explore optimal electrode configurations for cylindrical cells. Results show that optimal ERT configurations must take into consideration the required spatial resolution, sensitivity to anomalies, signal strength, and shunting effects along the cell perimeter. The system characteristics are defined in terms of electrode width Welec and length Lelec, cell diameter Dcell, and the distance from the electrode plane to conductive end-plates δ. The following dimensionless ratios emerge as guidelines for system pre-design: W_elec/D_cell≈π/2n, L_elec /D _cell ≈0.4, δ/D _cell ≥1, where n is the number of electrodes around the perimeter.