Floating points - how is this possible?

Floats are rounded to a certain precision., so i think that calculation introduces an extra decimal place, and that gets rounded.