+Charles Akins wrote a post called, Standards Turned on Their Heads. The post is about the problem with running established settings. Settings that people are very knowledgeable about and love. And how difficult it is to run a game where expectations, either by the GM or the players, can hamper someone's creative process. Not to mention players calling out inconsistencies. I wrote a long comment that didn't quite satisfy my gaming brain.
The first time I ran into this was back in the 80s. Dragonlance. You had a couple of camps, those the read the books and those that had not. Those that read the books basically wanted to reenact the books. The railroad style of play the Dragonlance adventures are notorious for only encouraged this. You play the characters in the books and go through the adventures they went through in the books. So this camp wanted to play the PCs as closely as the books as possible.
Those who did not read the books did give two shits in a basket about Tanis being a good leader, Strum being an honorable man and they all wanted to nail Goldmoon. The railroad nature of the adventures drove these guys nuts. I have no idea how many times I would hear "But Tasslehoof would act like that" and other variations.
Add that all together and its a tough way to have a good time adventuring because of the expectations of the setting interfered with the natural creative outlet of the game.
The few times I GMed where I used a well known setting were Camelot and Thieves World. Love those two settings and many people know them well. To avoid the problem of people's expectations of 'this is how it's supposed to be', I simply moved the timelines ahead of where the stories ended. The players found themselves in familiar settings, with familiar names and legends, but the story framework was not there. I could use the setting, use the power of the name and the emotion it evoked without being constrained by what was supposed to happen. The players know the Camelot when Arthur was around, but put the timeline ten years after his death, things change. Players get more invested because they want to see what Camelot looks like now and what happened to it and so on. It proved to be very efficetive.
With Thieves World you almost have the same situation, not nearly as known as Camelot and there was no central driving figuring of the mythos, yet there is a lot there to pay with. Because it deals specifically with a city, Sanctuary, the city really is the biggest character in the books. With Thieves World I had a little more latitude with the timeline. This time I started the players a few years before the first book. And before we started I told them this would be their chapter in the book. They went out hunting familiar names and places. And while they interacted with some of the book people they discovered their own storyline within Sanctuary and stopped worrying about Shadowspawn and Tempest and Durbo. And focused on their own thing.
Using popular settings is difficult. Especially if you have well versed players. God forbid if I tried to run a Game of Thrones game with my crew. They know it way better than me. But if you can manipulate the setting to remain familiar, but yet making it your own, you've got a great chance of having a memorable game.