I don't think any of the new ideas completely flopped, but not all of them may end up as keepers either. Here's my sense of what worked and what didn't work as well.
Enemy in the temple: No one liked this very much. The AI we came up with was clearly rough, but it was more the idea of having to have an AI was unpopular -- you want players focusing on how to plan their own best route through the temple, not having to worry about policing the rules by which the enemy explores. In a game with no GM, enemy movement just didn't seem to fit, and everyone agreed the time track was an appropriate abstraction.
Enemy outside the temple: I liked this quite a bit more. We played that you add an enemy when you move into a city; you reduce the enemy progress track by the number of enemy in the city at the end of your turn; and when there are 5 enemy or more in a city, you roll a d6 and the enemy either digs for the temple (1-2), kidnaps the theme card in the city (3-4), or activates an "agent" (5-6) (which simply controls the speed at which the enemy explores the temple). I think it worked pretty well -- you were worried about the enemy becoming too big in a particular city, but it didn't lock the game board down like the old system.
Lead cards: Neutral to negative. I don't think there was any specific excitement about this system, whereby you're told where you can find someone. But the implementation was at least part of the problem -- it was sometimes the case that you'd draw a lead to the same city you were in. And they weren't easy to get; facing a challenge to get a lead, or asking a theme card for a lead, were useful but not heavily used by the players. So as a result, it felt like you were just supposed to chase down whatever lead you were initially dealt, which lost some of the route planning that originally made the game fun. Maybe if, instead, every theme card has an associated whereabouts, and you need to get a lead to be authorized to look at it, this could be an acceptable compromise between the previous version, where all cards are always available, and the new system, where no cards are out at the start.
Encounter cards: Neutral to postive. These had descriptions of a scene, rather than an actual image, so players had to use their imagination, but they were all good sports about this, and most everyone saw some potential in the system. Concerns about knowing the "right" way to beat each card emerged, and the observation was made that there's no difference between the various skills -- they're just different numbers, but there's no difference between a "fight", an "escape", etc. A very interesting suggestion emerged to switch this to a group vote dynamic, a la Dixit, whereby you'd all look at the card, the active player would select his "response", and the inactive players would select what they think the solution ought to be, and if they match, the player passes -- or some permutation of that idea. There are some issues to be thought through but I think it could work, the question is simply whether there's enough to make it interesting.
Clues instead of check marks: Neutral. I think it was nice to be able to get a concrete clue from everyone you interview, but on the other hand, once you reveal that card X gives a level 3 clue, everyone else pretty much just wants to go visit that card to get a level 3 clue. This does what I wanted -- it creates convergence at a particular location, which boosts the enemy's strength by giving the enemy die rolls and progress track movement; but that doesn't create a disincentive to the other players to go there to get the best clue (*). The previous version made it harder to get better clues, this version makes it more statisically unlikely, and I think I like the previous version better, because it introduces more challenge to the decision of how much effort to expend to get the better clues?
(*) A couple of players suggested that a city that is full of enemy is locked down, and can't be entered by the players, period. And probably, that players have some ability to pull enemy out of the location they're in.
Theme cards as allies: Neutral to positive. Most players liked this thematically; Steve suggested separating the theme cards, which give info, from a separate category of ally cards, which give boosts to your stats.
Turn structure: Neutral. Turns were pretty quick, but on the other hand, there wasn't quite enough for the players to do; I think we'd like to see a little more action. Steve specifically suggested wanting there to be things you could be doing even in a turn where you won't going for information. I think this is a fruitful direction for further thought. But there are two important considerations that have to guide such thinking. The first is that the game can't tolerate much more complexity, and the second is that any system in the game should ultimately relate back to the information hunt in some way.
With respect to the last point, I think the "problem" is that there just isn't that much information in the game -- 3 clues with 3 levels each = 9 pieces of information, and you can possibly skip over the lower ones if you luck out or pay attention. I don't think we can add much more information than this, though -- breaking a clue into more than 3 levels would be difficult and/or would render the lower-level clues nearly worthless. So, the action in the game is really about adding obstacles to make the info hunt more challenging or time-consuming or interesting. The variable difficulty of visiting each location used to be the way that we did this, the current version makes it hard to find the people that give clues, but maybe there are additional ways to achieve this that we haven't considered yet.
Overall, it was an informative session that I hope will lead to some promising next steps for the design.