We consider results from from a one year simulation in the Oslofjord area, produced by an 800- and 300- meter versions of the ROMS model. The two versions differ only in resolution and grid orientation. Our objective is to evaluate similarities and differences between the two version. To this end we study current patterns, kinetic energy levels and last, but not least, particle trajectories. By doing this we quantify the effect of an increased grid resolution, enabling us to assess what size resolution is required for simulations of the fjord dynamics. We find that a higher skill regarding the representation of bathymetry and coastal geometry in the model grid enables resolution of smaller, stronger and more confined eddies, causing the higher resolution model to be more energetic. Moreover, simulations in areas where the two model grids display diverging coastal geometry and bathymetry produce diverging particle trajectories. In contrast simulations in more open regions where the relative difference in bathymetry is smaller yield more similar trajectories. This conclusion is reached by decomposing the kinetic energy into a mean and eddy part, and inspecting the eddy patterns of the two. We find a correlation between the spatial location of the eddies and the trajectories taken. We conclude, that it is not the effect of resolution itself that is of major importance in the fjord, but rather the implications it brings in the form of enhanced bathymetry and coastal geometry. Some other more obvious findings are that a more natural coastal geometry preclude/permit certain trajectories. Moreover, the inclusion of islands in the model domain obstructs eddy generation, and possibly cause island wakes.