Never presume that because your software agrees in one or a scant few cases with your specific recipes, procedures, or set of circumstances, that it will thereby do likewise for all other recipes and/or sets of circumstances or procedures. Hint, it most assuredly won't. Correlation does not imply or equal causation.
This is important. One of the things that happens with software
at scale is that the more users you try and tailor it to, the less effective it starts to get.
The most accurate and useful software? Usually a spreadsheet made for one user. It is tailored specifically to the process and brewhouse of one person. Say you know 10 people who brew roughly the same as you. It will still be accurate to a point, as each brewer may require certain tweaks that affect the overall accuracy of the individual but yield solid results for the group.
Now go to 100. 1000. 10,000 even. You end up diluting things and generalizing to the point where people accept things will just be "in the ballpark". I've gone back to just making sheets for the folks over at the LOB forum and only because there is a smaller pool of people and the spreadsheets stay consistent across the small group and the low levels of variance between ingredients, process, etc.
I don't envy those have to have to market and sell their software products (in any sector) to a large number of people because it is generally hard to please everybody, and in brewing in particular, nearly impossible to keep things to everyone's liking the more users there are.