The designers will have to ensure that the personal story does not depend on the world events and vice versa. The two streams can be related, but never dependent. An example would be the personal story informs the player of bandits doing bad things in a certain area. If the second player goes to that area and the poison has already been cleaned up, that player might find:
a) the first player did not seek out the bandits so the poisoning has started again
b) the bandits have resorted to outright attack on the farmers
c) the farmers are forming up to hunt down the bandits
etc
So the player participates in the event(s) where ever they are up to, but it should still make sense in relation to the personal story.

What happens to the next player who has the same or similar personal story? They get to the village with nothing going on – until the event somehow resets. No permanence there, unless Anet plans on having unique PSes for every single character created by every single player out there. None of the events can have permanent conclusions if they want to satisfy the achievers out there, because part of being the achiever is being able to experience all the content created.

Doesn’t matter if the next player gets an event of his own to solve. The achiever will still want to be able to solve the poison problem himself as well.

If you stopped the irrigation, another player might join you later on to fight off the bandits. That player didn’t know how the poison got into the water supply, and he might never know unless a bunch of events and variables are randomly triggered so that the bandits are once again trying to destroy the village. Probably, by that time, the two players are already on another event ;)

It also throws a small wrench in the wiki’s trying to say definitively, what is the best to do, and what to ignore.
People will still look up guides, but that happens on EVERY SINGLE video game in existence. So long as events are not too reliant on a single person being able to screw it up, to the point where others get the ‘elitism’ that infects other MMO’s, it will keep people from feeling like they have to research it before they play it (like WoW’s instances).

If on the other hand they have a lot of interacting variables and on a larger scale e.g. centaurs rise up, form an army, start threatening nearby farmers, which reduces food supply to a nearby city, which increases demand for food, which increases guarded merchant trips to the city, which interdicts the recent increase in brigands, who would otherwise have poisoned a well, which would have left the city with fewer guards, then *deep breath*…

you have the opportunity both for self-correction in the case of player incompetence/disinterest/laziness, and there are so many variables that players will actually find it quite hard to predict the results of their actions.

We will have to wait and see whether GW2′s system is complex ‘enough’.

Whether an event is worth the trouble for pure Achievers is determined solely by the whether the combined reward is greater than that of non-stop monster killing at another location. This is tweakable via the event reward, and should probably be based on real-world results and, at least as a first approximation, on the amount of ‘downtime’ the event forces on the player (e.g. how much of the time is spent fighting versus moving?)

Best of all, GW2 will probably be like GW1 where they can not only patch the server without downtime, but change a lot of parameters live, without anyone even having to notice! So balancing that over time might be easier than you’d expect.

BTW, an interesting point I haven’t seen mentioned: the experience reward will probably depend on the event difficulty/downtime percentage, but the ratio of gold to karma will probably vary more for roleplay/coherence reasons; i.e. if you save a poor farmer’s house from ogres you’ll get more karma, if you save a rich village of merchants from bandit raiders you’ll get more gold.