Neither system is infallible, and both have to use some guesswork on top of sensor data to dry a load appropriately.
If you want some insight, reading patents is a great resource. For example, read some of the 1950s auto-dry patents - that's when the basic systems that were common in the 60s-70s were really born, and the descriptions are the most succinct. Everything sounds so rosy and magical - wrap it up folks, the problem is solved.
Then skip ahead to the 1980s-1990s, and read the patents on the equivalent modern systems. They will describe the new systems and their merits, going into great detail about what was so wrong with the 1950s-1970s designs, and the complex systems and tables of data it took to improve upon those results.
Remember too that "dryness" will always be subjective. Maytag struggled with this a lot in the 1970s - erring on the side of under-drying to avoid going over, but receiving frequent customer complaints for incomplete drying as a result. Both their dryness thresholds as well as machine settings and labeling were adapted over the years, in response to customer feedback, and to ultimately move more of the "automatic" system's control and configurability back into the hands of the operator.