You shouldn't believe me
In fact, I don't consider myself necessarily good at it, just mightily intrigued. That said, I do get regular feedback that I don't..exactly suck at it, either. I'll admit it's nice to hear that your hunches are spot on, and moreso, your approach in using that information useful and can even be healing to others.
Meanwhile, this is how I go about things: I base my conclusions on the patterns I see, that I recognise and have explored before. And have verified before. I tend to ask...personal questions, especially when I'm one on one talking to someone. And people happily share and talk about themselves when they're in a safe, understanding environment. Patterns in their behaviour (bodylanguage, the way they responded to certain situations, how they slant and skew their own personal story, what they avoid, what they're drawn to, what the effect of that particular environment was and what the likely damage/benefit of that situation was), and so on) emerge as they talk about their past, their hopes and their fears and so on, effectively providing you with a personality map to what they value over what in life and why. And when you observe them later on in a social environment, the triggers for those behaviours become a pattern on their own you start to notice.
Once you observe that about a 100 times, you have a pretty accurate map of the likely reasons behind a trigger that leads to a specific behaviour, narrowing it down. When you then take those options and ask the necessary questions to verify your hunch...voila, the answer
Iow, the answer is 'Ive asked, repeatedly, and was curious enough to backtrack the answers, for years.' And from that, I can have a reasonable idea of what is likely going through someone's head. Add to that my own experiences and patterns and behaviours, allowing me access to empathic information from that, which I then can adjust for their own values and background information (as that will greatly vary their experience) and I can do an estimate of how the situation must be impacting them, giving me a good starting point for what they likely need right now, and as you approach them with that good starting point, you can hone in and fine-tune based on the way they're responding to you at that precise moment.
It's never perfect, and it's never completely correct...kinda like a flight simulator. But it sure comes in handy.
Granted, it can also make people incredibly uncomfortable, paranoid and even fearful that you seem to have all this information about them - usually those are the ones that are not comfortable with vulnerability and that haven't really had that one on one with, but who are displaying very recognisable patterns you've encountered before in others. It helps to either make it clear you have no harmful intent with that info, or to keep from acting on that intel - which..admittedly can be hard as it becomes second nature.
In short - it's just like any other skill you can hone. It has its benefits and downsides. You'll never be perfect at it, so you'll never stop learning, or being wrong. But you can become a lot better and a lot more aware of it, providing you with infinitely valuable intel on how to approach certain situations - like any skill.
I might as well ask the same question about the people that muck about with spring theory, tbh. I can trust that they actually have a skill that I don't and spent a lot of time refining it, therefore taking their word for the fact that string theory actually has value, or dismiss it as wishful thinking. Neither would really affect them or their expertise, and the same is true here.