In finance there are so many trade-offs that this approach is most sensible. At some point, I was trying to get a project going while working with a partner PM. He was a perfectionist programmer and literally obsessed over things like "we gong to have a script sending emails in this format" etc. Guess what, after about a year, as we went live we discovered that the idea was non-viable (for non-algorithmic reasons). This was something that we should have found out in the early stages, maybe even by trying to trade the strategy manually.