I scanned through this forum and didn't see mention of this, but I've been bothered by it since I started here: randomizing starting army placement can, in certain game types and given similar player skill, turn the game into 100% luck.
I know the game is fundamentally luck-based, but it's all about playing odds and taking risks. It shouldn't be about who had the better starting placement, and the board game way of setting up the board (taking turns placing armies, with the first to place an army also the last to take their turn) eliminates this issue.
On the site I've had games, particularly team games and no cards games, decided from round one by army placement. If team A has all of south america and africa, for example, while team B is scattered randomly across the other continents, AND team A goes first, there is practically no way for team B to win.
As for no cards games, which I used to like as they were a fine test of player skill, I've gotten turned off because too many of them come down to one player starting with all or most of a continent secured, and going early enough in sequence to secure it before they can be challenged.
So I'm not sure what the best solution would be, but what should be a fairly simple one would be to stop players/teams from getting too much of a single continent or too many adjacent armies in their starting position, and possibly gauging the starting position by whoever has the best placement and making sure that person doesn't go first.