In the animation we see an epicycle that has "half" the radius of the deferent being rolled. The epicycle contains a point on the circle and its trace is shown as a dotted curve as it rotates. The shape that is traced has a name, nephroid, although most of the possible traces are not named. The small circle actually makes three complete revolutions from start to finish. The two vectors, $\mathbf{u}$ and $\mathbf{v}$ point to the other two places where the epicycle has just completed a full revolution. You can try it by toggling the animation to a stop right when the epicycle is tangent to one of the vectors. If you do that, you will see that point $B$ is back on the east side of the epicycle, indicating a completed revolution.
In the next animation, we have an epicycle that has 1/3 the radius of the deferent. It will roll around the deferent with 3 revolutions, but will make a 4th revolution by the time it returns. This can be see clearly by toggling the motion when the epicycle (small circle) is tangent to any of the 3 vectors and confirming that the epicycle has made a complete revolution.
Besides rolling the circle around the outside of the fixed circle, we could also roll it around the inside. Doing that will produce a curve called an astroid (not asteroid). Although, if the number of cusps is reduced to two, the curve produced is commonly called a line segment. Even then, it isn't boring! If the point at $B$ were extended out of the plane, say toward the viewer, and the two circles were provided with gear teeth, then we would have changed circular motion into reciprocal linear motion.