What makes the battery so strange to deal with is that's it's the only electrical device that deals with chemical reactions. Every other device's behavior can be accurately predicted at any time knowing voltage, resistance and current. But not the battery !
One of the strangest behaviors is the "surface charge" effect. Surface charge is when excess electrons pile up waiting to leave the battery. Sort of like commuters that jam the platform when they expect a train. If you simply look at the 100 people on the platform, then you might guess the entire station must have a total of 1000 people, when it's really only 120. Surface charge does the same thing by giving you an incorrect "picture" of what's going on.
And that's what "load testing" does. It drains off the excess voltage so that a more realistic view of the battery condition is obtained. There are standards for load testing, but you can set up your own standard using an old headlamp bulb. Try connecting the bulb to the battery for exactly (say for instance) 2 minutes. THEN take a voltage reading, and that reading will be far more accurate. The bulb acts as a "load" and burns off the excess energy. A "good" battery should end up no lower than 12.3V, but might have originally read 12.7V just off the charger. (Just be sure and use a high wattage bulb, not some dinky 5W tail lamp bulb.)
If you have a 12.7V battery just off the charger that dips to 12.0V after being "loaded" you know that's the exact same thing the starter motor is doing, so the battery is no good.
Hope this helps.
