I kinda suspect there isn't an active community here, but whatever. here's my input...

Everyone knows the game's pretty old by now, and it's starting to show its age in the design. While it's not a bad design, there's limitations on it from being great.
Played it with some friends and we all kinda agreed the same thing. A lot things didn't feel impactful enough. While it's essentially a filler game ala card games like Loot... it's twice as expensive and oddly offers less strategic choice.

So, here's our test rules. I don't own the boardgame out of the group so I wasn't able to simulate extensive playtesting, but whatever. Anyways, it's 3 basic changes.

• Players get 3 actions per turn
  • With better games that have single-action turns (loot and skull, for example), the action you take in itself creates a consistent degree of stress in response to the other players. We didn't feel that placing a tile or moving 1 space consistently provided that level of stress. ... so we're trying it with multi-action turns now.
  • Idea being that if you're low on options you can just blow out some dungeon tiles to empty your hand get new tiles.
• Players may spend 1 action to refresh the hero ability. However, you can't do the heroic action twice in a turn
  • The hero abilities fall in a weird state of "way too weak to singularly throw the game" and "too annoying to be done every single turn".
  • So ... now you have to spend an action to refresh a used hero ability. Obvious exchange of action economy is intentional
• Drakon is now a persistent threat with limited AI...
  • First off, a titular character with a strong chance to not show up is underwhelming
  • Secondly, a titular character who's really just a tool for being mean to other people feels thematically wrong
  • So... this new subsystem seeks to address that...
• Drakon Change part A) Use anything at your disposal to mark a player as Drakon's target for her wrath. Just some sort of token to place in front of the player themselves.
• Drakon Change part B) As soon as the first player picks up a gold token, the mark is assigned to them. If another player gains more gold tokens than the marked player, they recieve the mark. If the marked player loses gold tokens and has fewer than the opponents, he gives the mark to the player with the most gold tokens (player's choice in the event of a tie)
• Drakon Change part C) When the marked player would take their turn, they must first move Drakon 2 spaces towards themselves. While the marked player gets to determine the path, it MUST close the distance between the player and Drakon (you can be smart enough to figure that out). /// and, if Drakon is not on the board yet, she is placed on the entry tile instead. (essentially giving all other players 1 turn to flee)
  • The reality is that there's an obvious winning strategy: set up a chamber loop of multiple coin rooms. Doing so with Drakon actively pursuing the player in the lead will put a short end to that.
  • Additionally, the system is designed to be a pretty straight forward system: whomever has the most gold tokens is targeted by a dragon that will chase after them at the start of their turn
  • Obviously it also means that Drakon will be ever present on the board.

So yeah, it's not perfect, but we feel it addresses several of the shortcomings.

Let me know what you think.