I think I'm finally overcoming my previous issue of reading too hastily and making careless errors, but I'm still making these "over-thinking" errors. Sometimes I am just shocked at how easy the question was that they were asking versus what I interpreted. I feel like a deep, brute force knowledge of the material will get you very far (like ~260, which is what I am averaging on practice tests), but beyond that it may actually serve as a hindrance as you start making unnecessary connections to esoteric points that the average student has never heard of and the examiner doesn't expect you to know. For example, there was a question on one of the NBMEs essentially asking how do CD4 counts go down in HIV infection. I can tell you that CD8 and NK cells kill virally infected cells in my sleep, but in that moment I started thinking about random papers that I've read about pyroptosis and HIV "controller" phenotypes lol...
At this point, I basically need to muster up the ability to take a step back and distill the question down to the most common-sense interpretation and, if it's between 2 choices, actively ask myself "Which choice, if I didn't pick it and it up ended up being correct, would I be more upset about?" -- that often helps eliminate the "over-thinking" option.
I partly blame UWorld for this. While it is an amazing resource, it has a critical flaw in that it trains you to overthink. Overthinking is great for getting those questions that only 20% got right on UW, but it hurts you on NBMEs.