As someone who works in a company with 10,000+ professional engineers*: yes, yes they do make mistakes. Despite what you seem to think, they are actually human beings. Good processes understand this and have multiple layers to catch and correct mistakes. But it's hard to fight limited data. Plenty of engineering decisions are still made based on numbers in photocopied tables from two or three studies from 50 years ago, each having a dozen or two data points. It's surprising how limited some of the data is that these Professional Engineers are basing their decisions on.

*Edit: okay, I don't know exactly how many are officially PEs vs. junior engineers or something, but at least that many are "PE-track".