Hey Mike, this is super interesting. Great post. I would be interested to know how draft time corresponds to word count. If longer drafts tend to contain more words then they also have a higher ceiling.
If that’s the case, may be worth looking at draft time versus percentage completion. Although you can’t measure the exact quantity of words read (yet), you could make that a function of actual time spent reading / predicted total time to read.
Also, I realize now that this post is over 2 years old, so perhaps you’ve already gone through all of this!