I am too late to offer my own speculation, but upon reading this thread I would like to offer a couple of collateral observations:
1) At one point, inferences were drawn about the state of battery charge based on voltage. Doesn't work. In theory, for a lead acid battery, cell voltage is a rough surrogate for state of charge if -- AND ONLY IF -- the battery has been isolated from all sources of charge and load for many hours. In modern vehicles, even with the key off, the battery is under a load to sustain the dashboard and electronics; these are not big loads, but they are enough to render terminal voltage entirely unreliable as a measure of state of charge. There are only two ways to tell the state of battery charge: direct specific gravity measurement of the electrolyte (works only for flooded cells) and sophisticated bi-directional summing ammeters (not found in cars).
2) Most of the time when a battery dies, the battery is a victim, not a culprit. That is to say, the battery almost never dies because it is worn out, but rather because it has been abused (usually by excessively discharging, which results in warped and shorted plates).
3) The life of a lead acid battery is measured in the number and depth of the discharge/recharge cycles it suffers. For good deep cycles batteries, the battery should last for about 500-750 deep (i.e., to the 60% charged state) cycles. For starting batteries, one should be able to get more than a couple of thousand cycles.
4) In fact, in an automobile application, the battery gets very little load. It is used primarily to start the engine. Most starters draw on the order of 150-250 Amps, but they do so for only 5 or 10 seconds (or less). This works out to less than 1 Amp-Hour (or about 1-2% of the battery's nominal capacity), and this minimal draw is almost immediately replaced by the alternator (assuming the alternator and regulator are functioning properly). After that, the alternator/regulator serve as a battery eliminator and carry all of the car's DC loads (again, assuming they are working correctly).
5) Alternators are consumptive devices; that is to say, so long as they are spinning, they are using up some of their useful life. High capacity alternators in marine applications are routinely rebuilt every three or four years. In most automotive applications, the alternator will begin to fail long before the battery does (but as the alternator fails, it will begin to take the battery down with it, because it will begin to shift to the battery load carrying duties that the auto battery was not designed to carry). Likewise, the regulators in most auto alternators are inexpensive, fixed-voltage, cannot-be-repaired devices that are not likely to survive more than three or four years. (The regulator lives inside the alternator case, where ambient temperatures routinely exceed 200F.)
6) As a side note, an automotive alternator/regulator setup is not designed to recharge -- and cannot recharge -- a substantially discharged battery. Such regulators operate on a fixed voltage that is slightly above the "float" voltage of the battery (say 13.6 or 13.8 VDC). At that voltage, the battery's internal resistance will quickly reach the point where it will not accept any more current from the battery. The only kind of charger than can practically recharge a materially discharged lead acid storage battery is one that can take the battery to, and then sustain it at, a voltage in the range of "gassing" (mid-14's for most battery lead-acid chemistries) for the right amount of time. These multi-stage, computer-controlled, shunt-regulated chargers are not found in automobiles.
7) For what it might be worth, I have been consulted a whole bunch of times about "dead battery" cases involving fire apparatus, fire command vehicles (usually Ford SUVs), and police cruisers. In every single case, the culprit was either the alternator or the regulator.
For those couple of folks who may still be awake: sorry for the long post.