Hi Aaron,
Thanks for the file! (The forum is picky with files, all c4d files need to zipped).
Please have a look at this file (Preview mp4 included)
https://www.amazon.com/clouddrive/share/73zGUUGIQqqq59GtibXwHsG9zb31IK2LhbxreFlUcwb
There are a few things that I like to address.
The first is, that the Object manager runs from top to bottom, in terms of processing, except the Priority numbers and levels are set. The lower the number, the earlier something is calculated. This also means higher numbers (or layers, e.g., Expression [3] or Generators [4]) are processed later, and those results are not available for lower priorities on a given frame.
If something is not available freshly from that frame, the previously stored information is used.
In this way, one can create a very long delay chain. The more dependencies a scene has, the more likely those problems show up.
Since many objects have hard-wired Priorities, this can be difficult to manage.
It gets even worse if the Xpresso nodes are not sorted out. Typically most of them can get the information in time (frame), but the X-Manager in the Xpresso Editor should be sorted from top to bottom as well to avoid problems.
I often use the Layer Manager to give colors and names to objects, when I get something like this to sort out. To “see” the flow in the whole scene.
With that sorted out, and the problem persists, things need to be baked, cached, or streamlined as much as possible.
One major thing is not to take information from an object that took it from another already. Go to the earliest source, like the fishing line from the top of the pole to the Water. One end is given in the Pole Ring already, so no need to have an extra step in it. I cached the Pole, and that cached the influence of the Deformer. It can be re-cached for changes, but the given information is available form the entering of the frame.
I merged the Pole (not the handle) into a single polygon object. This allowed me to set the Deformer as a child (less trouble!). From here, I took the four-point values easily “mapped” to a spline.
So, sorting things out, getting clear about the flow, producing (cache or bake [PLA eventually]) data early on, and taking source not second-hand information seems to work.
My best wishes.