When it comes to running books, "Training for the Uphill Athlete" is the Gold Standard. It's both descriptive and prescriptive, backed by research, informative, well-written, well-structured, and provides a wealth of data, including charts, tables, and illustrations.
Unfortunately, this book doesn't quite rise to that bar. It's written descriptively, but also does not provide prescriptions outside of training plans that make up 1/3 of the book. It is well structured, but also written from a high level overview that introduces running power to the reader as if to sell it. The book does not provide any references to back up any of the claims made.
However, the book provides a wealth of data metrics and definitions, and defines them very clearly. When taken in the whole with heart rate and pacing metrics, it's not difficult to see how valuable running metrics can be for the athlete.
One of those metrics, Efficiency Index (EI), is invented by the author, and I struggle to find its value. It's defined as pace in meters per minute divided by average power in watts. This metric is very similar to Running Efficiency (RE) which is defined as pace in meters per second divided by normalized power in watts per kilogram.
Because EI is not normalized, it can have values larger than 1.0, which is (should be) impossible for RE. So one run might show an EI of 0.75 while the other shows an EI of 2.13. What do these values mean? Jim Vance tries selling EI to the reader multiple times in the book after defining it, but it appears its only value is seeing if it increases for similar efforts. It's not all that useful as a post-run metric, and it's not all that useful comparing to other athletes, such as elites.
I do think I'll find myself coming back to the book for some definition specifics that are difficult to find online, so I appreciate it as a reference. Again, it was well-written. Aside from the couple of typos that the editors and publisher missed, I had no problem with the language or presentation of he concepts. However, as a mathematician, computer scientist, and post-run analysis nerd, I wish there would have been more data-driven examples with references.