… your estimates are implicit by creating (hopefully) similar sized slices.
Exactly. I don’t doubt that Joel’s method is just as effective, but by sizing similarly we don’t have to pay the cost of designing up front nor giving hour-granular estimates to our development tasks, both of which seem pretty expensive. In a recent episode of the Agile for Humans podcast Troy Magennis said something like “if two measures are similarly useful, use the cheaper one”. :-)
I wish I had some data on how similar “similar sized slices” need to be. My intuition says that it doesn’t matter as long as the slicing heuristic is well defined, but it would be nice to have empirical evidence to back that up.