I'm sure everyone here has heard of PPR - in a game without or with unreliable +Buy it's dangerous to get the second-to-last Province because the other player might have more VP than you blah blah blah. There's also that tactic of trashing all but one opening Estate so you can win on an even Province split.
I'd like to step in here because this is a common mis-statement of PPR that can lead to some really disastrous decisions if applied literally. A good, concise statement of the Penultimate Province Guideline is: If there are two Provinces left, and +buy is hard to come by, and either deck could hit 8 any time, and
you are currently behind but could get ahead-or-tied by buying anything but Province, then don't buy the Province. So if you are losing by one or two: buy a Duchy. If you're behind by three
as second player, buy a Duchy.
Otherwise,
get the Province. I have seen countless games in which someone is winning by two points, and buys a Duchy because there are only two Provinces left, citing PPR. This is a disastrously poor decision when the next turn is only good for another Duchy, while opponent gets both remaining Provinces for the win. If the leading player had just bought the province, extending their lead to 8 points, the game would be much, much more secure. Similarly I have seen players who are losing by four points pick up a Duchy. PPR does not tell you to do this! If you're already behind by four,
you need both Provinces to win! So, buy one of them now while you still can!
Anyway, this is only tangentially related to your more general "greening traps" question, but I try to prevent these misunderstandings of PPR.
On the topic of your original question: similar things happen in games with (for example) Remodel, where either player has the "option" to remove a Province from the supply by remodeling a Province into another Province. I
recall a game I played against Temron about a month ago, where he had out-done me in the engine-building phase, ending up with a much better deck than me. We could each reliably hit single Province, so it looked very bad for me, but I bought the first Province, and with six Provinces left in the supply I bought one and milled another, leaving only four. This altered the pace of the game enough that I could come out ahead.
Now for some more abstract theory-stuff about the general case of such situations: this is all an interesting situation related to the theoretical "game" of
Nim, which is covered in some math and computer-science classes. There are N objects and you take turns removing between 1 and M of them, and the goal is to be the player who takes the last object. In terms of the Dominion game I mentioned, N is 6 and M is 2: you want to take the last Province because then you can do some other stuff on the same turn, like get Duchies or Forge your Tactician or whatever. You use your trashing flexibility to get you on the winning side of that Nim divide.
In fact, ordinary PPR is related to Nim too: take N=2 and M=1. You want the last Province, but you need to be ahead before reducing N to 1 so that your opponent can't take the last one and still win. This would even extend out to N=8: if players built perfectly-reliable decks before even beginning to green, you would never want to buy the first Province!
