Instead it uses fixed.Int26_6. Could have used floats but then we would need a new point datatype and we already have utility functions for fixed point math