Floats lose precision because of the way they are stored. Essentially, it is stored as a fraction instead of explicitly; Wikipedia's article offers very good explanation: https://en.wikipedia.org/wiki/Double-precision_floating-point_format.