We study the quality of secondary school track assignment decisions in the Netherlands, using a regression discontinuity design. In 6th grade, primary school teachers assign each student to a secondary school track. If a student scores above a track-specific cutoff on the standardized end-of-primary education test, the teacher can upwardly revise this assignment. By comparing students just left and right of these cutoffs, we find that between 50-90% of the students are "trapped in track": these students are on the high track after four years, only if they started on the high track in first year. The remaining (minority of) students are "always low": they are always on the low track after four years, independently of where they started. These proportions hold for students near the cutoffs that shift from the low to the high track in first year by scoring above the cutoff. Hence, for a majority of these students the initial (unrevised) track assignment decision is too low. The results replicate across most of the secondary school tracks, from the vocational to the academic tracks, and stand out against an education system with a lot of upward and downward track mobility.