The greater topic is player expectations. If you're playing a heroic fantasy game, developers will generally assume you want to play some sort of fantasy hero along the known tropes. Since we're talking about good vs. evil stories, I'm coming from the assumption that the described scenario isn't a random encounter, but the beginning of a campaign written without a specific group of individual players in mind, like a game company would have to produce. This is what makes it tricky. Of course, if both heroes and villains are only passing through, you need to come up with some adventure hook to involve them. I'm assuming some generally benevolent force is in power, not able to protect the farms on the border. After getting the players involved, the usual way of narrating a 'good' campaign would be investigating where the goblins came from, why they attacked and how to stop it from happening again. This doesn't mean there can't be plot twists, but it's a classical backdrop that would work for both hack'n'slay and political drama. Especially when playing with alignments, hero players in heroic fantasy are much more likely to work with the existing good or neutral-aligned structures (in 2e, paladins would risk forever losing their class abilites and become regular fighters if they didn't), like the local noble who hired them, while evil will be forced into either a submissive or a disruptive role. If you don't keep this in mind, these disruptions will lead to interruptions, because you hadn't planned for an unexpected outcome of a scenario. At a gaming table this will, in the best case, lead to improvisation and lots of documenting and re-planning work afterwards. In the worst case, the game has to be stopped. As a player in video games you would either hit brick walls with your actions or suffer from a lack of immersion because the system can't cope with them and has weird NPC reactions or triggers not working. Yes, the player of a good-aligned character can surprise you as well, but their goals will still resonate with the good outcome, instead of a number of potential individual "evil" goals.
In a broader context, it's also not really a question of harsh good vs. evil morale. It's about preparing for disruptive players. In Shadowrun, you essentially play throwaway agents working for corporations against other corporations. You break in, steal, betray, kill the guards and hopefully escape with the loot. But still, most players will follow the middle of the road = meet the guy, accept the job, break in, get double-crossed, survive against all odds, escape with the booty and get the Karma points. If you sell it to the guy who double-crossed you, or kill him to sell it to the competitor, or blackmail the company you stole it from - those are variants of the same playthrough. A player toxic shaman, however, would just use the stolen key cards to get onto the compound, but would then abandon the run to sabotage the waste processing of the compound which causes an environmental catastrophe in the area, drastically increasing the player shaman's powers, but laying waste to your carefully designed setting. Instead of worrying about police response times in case of PCs triggering the alarm, you now have an ecological disaster scenario at hand, with many NPCs rushing home to save loved ones or being evacuated. And you can't move on to adventure two, because the players never stole the loot.
...and that's what I'm reading from a lot of the negative feedback for evil playthroughs. Players expecting they could commit a very specific sort of evil, like burning the mill with Barcus attached, or whatever, and then get disappointed because the game wouldn't have it. I wanted to blow up the loos for grease mephits and extra-loot, but not even smokepowder barrels did any damage. There are complaints about cutscenes or NPC feedback not acknowledging certain deeds or said deeds not having the expected repercussions. There are similar gaps for the Githyanki playthrough where you're sometimes treated like a human Tav, since being a second full Githyanki would have required an entire additional variant story arc with cutscenes, NPCs reacting differently, and potentially even more endings. You won't hear the player of Joey Tav complaining, because they got their heroic tale.
Then the last problem is that the official continuity of a franchise will probably canonise the "good" ending if they continue the story, meaning any disruptive actions are erased from memory, leading to more disappointment.
What I'm taking from this is that it's not actually a majority of players preferring darker settings, but a vocal minority of players who like the disruptive style and often get disappointed, because it's hard to prepare for such players if you don't personally know them. It's not about not seeing enough blood and gore, it's about being able to take a lanceboard and playing a different game on it.
