A most interesting topic, and it made me take another look at this Caracalla Tet where the top left quadrant of the reverse is apparently reproduced on the lower right of the coin.

Initially I had assumed that the double strike on the reverse had been made from the same die. On a closer look it appears that a second die was involved. The wing does appear to be of a slightly different design, but the main difference is the lettering in way of the wreath. In the top left quadrant the P and the X are on either side of the wreath, and in the lower right quadrant the P and the X are before the wreath.
Quote:
A range of experimental archeology has been performed over the years to understand how Roman coins were produced. From this we can conclude that the generally the obverse die was engraved and set in an anvil and the reverse die was set into a punch. Blanks were created and placed on the obverse die in the anvil and the punch was struck with a hammer to impact the images from the dies onto the blank. That is all fairly straight forward. It has been noted in these experiments that if you have a single pair of dies and you start trying to operate at speed that the reverse die starts to overheat and crack or mushroom. These experiments led to further experiments with different strike rates etc. which let to an experiment where two reverse dies were mounted on a yoke and these dies were alternated between strikes. This let to the die being able to cool between strikes sufficiently that the cracking and mushrooming lessened significantly.
This explanation from @maridvnvm shows how the areas could differ, so thanks @maridvnvm for the research highlighting this.
