No, it's entirely possible for a poor connection to behave as a resistance rather than a completely open circuit. Consider it a function of the surface area making contact between, say, a spring terminal and a battery. At some point enough area is in contact that the resistance is small enough to be negligible, and any more contact area makes no difference to anything. But as the contact area decreases (perhaps by accumulation of corrosion or other crud between the metal surfaces), the resistance increases proportionally.
Particularly irksome, in a murphy's law sort of way, is that the voltage lost due to this contact resistance will depend on the current being drawn. So from the device's perspective, the battery voltage can look fine, until it actually needs to draw some current, at which point the perceived voltage suddenly becomes Not Fine.
(Resistance within the battery itself has the same effect - you'll be familiar with knackered batteries appearing to have a sensible voltage, which then collapses as soon as you try to use them.)
I reckon removing and re-fitting the batteries scraped some oxide off something. Though there remains a slight possibility of a software issue causing it to stop sampling the battery voltage for some reason.
