OK, I bite. Here's my version of that:
http://www.slowtiger.de/examples/pendel.html
It's just an improvised movement, but within a planned setup. As you see in the file (
http://www.slowtiger.de/examples/pendel.anme.zip), I had the string and the medal as separate layers, the medal was layer bound because it's rigid. The bone setup may look strange, but I did that on purpose. A chain of bones starts at the medal
upwards along the string. From its root bone there's one bone for the medal. The origin point of that bone layer I set to the pivot point of the pendulum, the origin of the medal to its pivot as well.
I started animation with selecting all bones and create rotation and position keys, as well as a rotation key for the bone layer. This was the resting position of the pendulum, I pushed it to the end of the timeline and worked backwards from it.
First I created the overall pendulum swing. Using onion skinning for the extremes, I set each swing further outside than the last one.
Then I copied the resting position keys just one frame before the first swing extreme. I grabbed the root bone and translated its position, then did the same some frames before that to create the start oposition of the fall. The string of course became un-attached from its pivot point now. So I selected all bones minus medal bone and rotated them together - this way you get smooth curves in a chain. With careful rotating the root bone I was able to position the end of the string at the pivot point again. I repeated that for all frames within the medal fall. As you will notice, the end of the string isn't really fixed, but that doesn't matter, it looks good enough and could coverd somehow if the pivot point were visible. I fixed the medal rotation so it fell in nearly a straight line.
In the first swing I needed the medal to jump up again, so I repeated what I did in the last paragraph, only creating an S curve in the string, just for variety.
Since this looked nice, I created a swing of the medal alone, with its bone rotated a bit, but with a different frequency than the overall swing. Last thing I did was an oscillating medal layer scaling, again with a different frequency, to simulate the medal's turning at the end of the string.