RChung

There are ways that you can check accuracy (and precision) by yourself but, depending on what your needs are, they range from relatively simple to quite complex. These methods are well-understood if not well-known. Most power meter companies don't actually want you to know how to check accuracy or precision.

At the easiest end, you can do a "static" calibration check. This involves getting some known weights. You don't really need gram accuracy but I work on a university campus and I had a friend who worked in a lab so he weighed some of my barbell weights on a scale and we marked their mass on the plates with a sharpie. You could do the same with the Post Office or UPS or FedEx. They have to get their scales checked and calibrated regularly so you know they're accurate.

Then you put your bike in a sturdy stand or on a trainer, hang your weights from the pedal or crank arm and, depending on whether your power meter is a pedal, crank, crank spider, or rear hub type, do a little calculation for the torque. Then read that off your head unit. Try different combos of mass to check that it reads accurately across the entire range of weights you have.

A dynamic check is a little (a lot) more involved but it's the same principle: you set up a situation with known drag force and then check to see if the power meter is reading properly. Because it's dynamic there are more things you have to keep track of, but you can still do it at home if you're motivated enough. I've done it but it's enough of a hassle that I normally just do static checks (unless there's a good enough reason to go through the hassle).

It's really not a mystery. People who say it's impossible (and thus conclude that all power meters are equally accurate/inaccurate) just haven't thought about it.