Ok, so now that I've drooled over the contents of the Castle Ravenloft game, how does it actually play? Short answer is: pretty well.
Right off the bat, it passes one of my biggest tests: It plays fast. A given session takes about an hour to play. As an old guy with a kid, this is a big deal for me. It means weeknight games (sometimes even, *gasp*, multiple game) are a possibility. We've knocked out 4 games so far, both of them in pairs, something that was absolutely delightful.
At a high level, the game plays a lot like a 4e dungeon crawl, with the dungeon building out randomly and each tile equating roughly to an encounter. You play recognizable 4e characters at first level with powers represented as cards, with numeric values like AC streamlined but recognizable. Over the course of play you'll encounter monsters, traps and random weirdness, all attempting to kill you while you attempt to achieve some objective based on the scenario you're playing.
The scenario's are the rub. Over and above the dozen or so in the box, there are more of them online - both official and fan created - and they provide the real variation in play. They also provide a lot of the challenge. Of the four scenarios we've played so far, three were cakewalks - straight tactical romps. The fourth very nearly wiped us, primarily because it added extra considerations that kept us from playing optimally.
That's awesome. One problem I often have with cooperative games is that they can easily turn into puzzles to be solved. Yes, there might be some randomness that could hose you, but given the right kind of approach, a game like Pandemic can be solved more than played. Adding extra limitations, especially dynamic ones, really cuts into that and brings back the game. That said, I'm a little fearful of how the scenarios will hold up to repeated play. I know I don't really want to duplicate one until I've tried them all, but given how fast games play, it's going to be a problem eventually.
In the end, I can't really speak to how well CR works as a pure boardgame, but as a streamlined D&Dish experience, it's great. The random element introduces enough tough choices to keep things interesting, and some very clever mechanics really keep you on your toes. For example, the map is made up of tiles (which in turn are composed of squares). Player movement and adjacency are determined by squares, but pretty much everything else is measured in terms of the tiles. This leads to some fun exploits (fights on tile borders) but it also has a big behavioral impact.
See, the sequence of the game is pretty much 1) Player acts 2) Player reveals a new tile and monster 3) Bad things happen. This means that you pretty much get rushed by monsters ALL THE TIME, and you depend on the next player to deal with the monster you revealed. This means that if you want to be really efficient, you stay together, except there's a catch. Most bad things happen to everyone on a tile, so if you group up to best fight monsters, you're more vulnerable. This tradeoff means you need to stay on your toes, staying close enough together to cover each other yet not so close to all get caught in the fireball (This is also why events that move players around can really mess with you).
This revealed something surprising to me. There are several decks in the game, one for monsters, one for loot, one for events, then one for the dungeon tiles themselves. The Monster and loot ones work roughly as you'd expect, but the other two have interesting subtleties. First, while the monsters seem like the most obvious threat (especially in the case of villains - boss monsters of certain scenarios), it's the event deck that really drives things. No one event really overwhelms things, but it will kill you by inches. It provides a constant drumbeat of menace that really sets the tone of the game.
The dungeon tiles are something I didn't put much weight on at first. They're pretty and clever in their interlocking, creating lovely dungeons, but at first they seem like a timer. You have an objective tile somewhere in the deck, usually 9-12 cards down, and all you need to do is churn through the tile deck to get there, right? So I thought, until that one game that went badly - the shape of the dungeon in that one ended up having a big impact of play, forcing us into tactical decisions at times and at other time giving us advantages we could leverage against the bad guys. It was, in part, a function of the scenario, but it revealed possibilities to me that I will keep my eyes open for.
Like most such games, things are nominally equally difficult dependent on the number of players because the rate of opposition action is tied to player action, but I suspect more players makes life a bit easier. More synergies, better ability to cover one another as well as other advantages. It's not a huge thing, but it seems noticeable. Time will tell.
Anyway, in case it's not obvious, I enjoy the game a lot, and I totally feel I got my money's worth out of it. It's not flawless, but most of my complaints are minor, fixable thing (difficulty to distinguish minis, slightly flimsy cards, a serious need for an official FAQ and a good example of play), with no obvious dealbreakers, except possibly replayability. We shall see.
So, go play it. But when you do, make sure to read the cards carefully. It's really important to be aware of what is or isn't an attack as well as what phase things happen in. This is an area where your 4e assumptions may hurt you, so take the extra minute to actually read, not just assume it's what you expect. We made a few mistakes like this, and each one of them hurt us.