The underlying assumption here is that when transferred into R or Julia, with an equivalent level of development effort and testing, that there will be no/fewer

The underlying assumption here is that when transferred into R or Julia, with an equivalent level of development effort and testing, that there will be no/fewer bugs. I'm not entirely sure that's true. I think we are comparing something that is knocked out in an hour in excel to someone spending 4-10x of the time doing it in Python / Julia, but I don't think that's a like-for-like comparison.