Well, to be honest I don't know if it was a stress failure or a sudden overload of something caught in the spokes that caused the damage to the wheel that I've got. But later models came with machined troughs between the nipple seats, which may well be from trying to keep the weight constant while strengthening the most stressed parts.

How the overall life calculation was planned out from Mavic I have no idea, but judging by forum posts it seems like it can go either way.
There's a decent Ebay trade in single and small pack spokes - which wouldn't exist if rims consistently failed before spokes.
There's a decent trade in freehub bodies bearings, which, given the cost of replacement rims and a full set of spokes vs new wheel, wouldn't happen either if spoke/rim failure through fatigue was a common enough occurrence.
My theory - not that I have much invested in it - is that it seems like (some) Mavic wheels are marginal enough to be heavily influenced by rider and ride style. People who ride "light" - not necessarily short distances - can get good mileage out of them, while other riders seems to be able to thrash them in short order.
Add to that assumption that many of them get sold to "weekend warrior" types who maybe not manage much of a mileage and you'll have life predictions all over the place.