If you can die before saving in a video game intro, then yes. It's like that.
I think the purpose of the encounter is to be a tense "we are so out of our league" moment. But the PCs can die. Their actions count. If they ignore it, the dragon kills a bunch of people and then gets bored. If they aren't smart and safe about how they drive it off, they die. It's just the dragon doesn't unleash its full might.
Thaumaturge.
I admit that the concept is good, but you'll note that the OP didn't criticize the concept of the module, but it's implementation. Starting a campaign with a tense "we are so out of our league" moment is a good thing that I highly recommend and have been recommending on these boards for like the past 10 years. But it's not the general concept in question here that is the problem, but the implementation. The particular implementation here kludges two scenarios together in a way that is lesser than both, and which misses the critical details of doing either well.
Those two scenarios are:
a) "We are so out of our league"
and
b) "It's a stalemate."
In "We are so out of our league", the PC's don't dare face the threat. It's way out of their league and if it unleashes its full might, the PC's are toast. There are three critical aspects of this scenario. First, that the threat is so unimaginably overwhelming that it is absolutely clear to everyone at the table that it is suicide to 'stand and fight'. For example, you are first level and its clearly a colossal elder wrym dragon, with a CR 20+ above anything you could actually face. Second, that it isn't paying attention to them, and there is virtually nothing the PC's could do to draw its attention on them specifically. The threat could be mindless, or is observed at a distance, or generally believably has better things to do and that is conveyed to the players in a meaningful fashion. And thirdly, and most importantly, that despite the fact that the threat is beyond the means of the players to deal with, there is still something meaningful and exciting for them to do. That means that for any "We are so out of our league" scenario, the PCs need to be able to acquire new meaningful goals in the context. So, those could be, "Hey, we can't fight that thing, but at least we can rescue the orphans from the burning building.", or "Hey, we can't fight that thing, but at least we can loot the abandoned shops.", or "Hey, we can't fight that thing, but at least we need to escape the firestorm that is consuming the city.", or "Hey, we can't fight that thing, but look - minions!" In "We are so out of our league", you have what is the equivalent of a disaster movie going on. "Ok, we can't stop a tsunami, volcanic explosion, or Godzilla, but at least we have something heroic to do at the moment, if only escape with our lives."
But this scenario doesn't play out cleanly like that. Instead, because the PC's are expected to take the initiative in driving away Godzilla and getting Godzilla's attention, what we have is more like the scenario, "It's a Stalemate!"
In "It's a Stalemate", the intention is to introduce a reoccurring NPC villain with the PC's playing the role of foils so that each will then have a vendetta against the other. There are two critical features of "It's a Stalemate". First is that both sides can place the other in check, so that neither side can continue. One example is that the PC's place the villain in position which it must retreat because otherwise the NPC will be overwhelmed, but conversely where the PC's are unable to pursue either because they'd be overwhelmed or because the NPC simply has a mode of travel they don't. The reverse is also possible, but harder to pull off consistently, with the PC's being able to successfully retreat while the NPC dare not follow. The second and harder feature to pull off is that the end result must be both sides manage a partial victory, succeed within a limited goal, and logically had no better strategy available to them. And therein is the problem with presenting the dragon encounter as, "It's a stalemate." Because unless the PC's have provably survived the worst the Dragon can dish out and given as good as they've taken, why should the Dragon allow himself to be driven off?
The examples of how this has actually played out justify the concerns with the encounter design. If it is "We are so out of our league", attacking the dragon is the "wrong" move. Yet there is nothing for the characters to do if they misunderstand the intentions and don't attack the dragon and no impact on how the script is suppose to work. Choosing to treat the scenario as "We are so out of our league" is rational, but within the story doesn't "fail forward" gracefully. But conversely, while having a dragon that can just fly away ought to make, "It's a stalemate!" a viable scenario, the actual implementation is that the dragon flies away "just because" rather than because it was left by the PC's with no choice.
For, "We're so out of our league", a bigger dragon and more fallout is required. Plenty of scenarios around a great wyrm attacking a town can be imagined. The fun is in dealing with the indirect results of the dragon's attack - dealing with looting cultists, calming panicking mobs, assisting the innocent to safely, putting out fires, rescuing the injured, and generally just escaping the catastrophe. For, "It's a stalemate", I'd probably build the scenario around the PC's rallying local defenses and becoming the focal point for more widespread resistance. I'd give the PC's a 'stronghold' that they could defend and a group of innocents to be the bulwark for, with NPC allies that could be rallied to be in a support role - the retired fighter, the town cleric, the city hedge mage, the dwarf merchant. By fortifying the stronghold, say a fortified temple, and dealing with a series of scenarios as the Dragon tries different ways of overcoming defenses, the PC's eventually thwart and slow the dragon's depredations while surviving the best it can deliver. Eventually, the 'cavalry arrives' in the form of reinforcements, and between that and the PC's, the dragon decides to leave with a partial victory. This is a bit trickier to pull off because there is the possibility of the PC's not counter-punching the way you imagine and you end up in a full failure scenario, but doable. You also have to guard against the PC's choosing to be passive and excusing themselves from the story. That is to say, if the NPC's in the stronghold our your quest givers, or if the justification for the PC's being on the quest by someone is solely "Thanks to your heroism", you better have a fall back position.
Well scripted encounters seem like they arose naturally. Poorly scripted encounters have quite likely results where utterly illogical things are happening because 'plot'.
