Fred asked me to post this after we had a side conversation about plug-in battery checkers vs. coulomb counters.
A plug-in battery checker cannot be a coulomb counter. The only way to count coulombs is to have an ammeter inserted in the battery lead as it's charging and discharging. In your laptop or phone this "ammeter" is an itty bitty resistor and some circuitry on the battery monitoring chip.
It's impossible for a plug-in checker to do that, because it doesn't fly along with the plane. So it has to just be looking at voltages and making guesses. If it's showing you percent charge, it's because someone's boss said "marketing wants us to show charge". If that someone was good, they said "but, but, but..." before their boss said "just do it fer c****sakes!" and they just did it.
LiPo cells are -- for batteries -- weird in that they have a pronounced voltage vs. charge characteristic when there is no current being drawn from them -- almost all other cell chemistries will show the "full charge voltage" even when they're nearly dead. So the old RC battery testers would actually draw current to get an accurate-ish reading. LiPo testers can just look at voltages and at least get a sorta-kinda good idea.
It is theoretically possible for a tester to try to measure more than just voltage (as, for instance, by drawing current like the old NiCd testers). But given their price and size, and the ones that I've taken apart and looked at, I'm certain that the little plug-in testers just measure cell voltage. Personally, if I had one that told me state of charge I'd just switch it over to telling me voltage and I'd make my own conclusions.
As to deducing state of charge from how much you put back in to the battery -- another way that LiPo cells are peculiar in the battery world is that for almost every bit of charge you put in to the battery, you get that same charge out. Few electrons manage to bypass the charge/discharge process. Which is a really tediously long engineering-nerd way of saying that yes, monitoring the charge that goes into the battery is a really good way of telling what came out, assuming that it was fully charged to start, and fully charged at the end.
Gonna stress that last point: yes, monitoring the charge that goes into the battery is a really good way of telling what came out. If you start with a fully-charged battery, go fly, then fully charge the battery, the mA-hours that the charger puts in to the battery is almost exactly the same as the your motor took out. Unless we want to start equipping our airplanes with ESCs that have coulomb counters, this is the only good way to get that information.