The detail that makes this post is that the page already printed the catalogue prices a few pixels below the estimate, and the two disagreed. A per-case rule priced pressing at $3,300 to $9,900 when the cheapest semi-automatic press listed was $24,975, and a press does not come in a smaller size just because you make less wine. Using the listings as a floor at the chosen automation level is a tidy fix.
It reminded me of a piece we published at Ascendra Ventures about the "70% of transformations fail" figure. It started as an unscientific estimate in a 1993 book and kept getting repeated, while the measured data underneath was more useful: Bain found 90% of the value comes from less than 5% of the roles, and McKinsey found even successful transformations capture only about two-thirds of the value. Your calculator had the same problem in miniature, a rule of thumb sitting right next to real data and winning.
Are you planning to show buyers which stages are set by the catalogue floor and which still come from the per-case rule?