So I just played an interesting game involving Mountebanks, Gardens, and Ironworks.

http://dominion.isotropic.org/gamelog/201212/06/game-20121206-085141-5d0931c8.html I lost the Garden split 3-5, but won due to my larger deck size, fueled by my Ironworks. Both of us sort of ignored Lighthouse, as the Curse+Copper from Mountebank isn't that terrible in a Gardens deck anyway. I get so into this mentality that at one point, I voluntarily take the Curse+Copper even though I have a Curse to discard. If my opponent had done the same at some point, he would have won! This got me thinking as to when this is actually a good move. Let's consider the last reshuffle of the game, so the curse and copper never get played. Is this worth is from just a VP perspective?

Supposing that you have no idea how many cards are in your deck, adding 2 cards gives you a 1/5 chance that you'll cross into the next decile of cards and get more points from your Gardens. So, the benefit is dependent on how many Gardens you have.

With 3 gardens, you have a 1/5 chance of +2VP, 4/5 chance of -1VP = -0.4VP

With 4 gardens, you have a 1/5 chance of +3VP, 4/5 chance of -1VP = -0.2VP

With 5 gardens, you have a 1/5 chance of +4VP, 4/5 chance of -1VP = 0VP

With 6 gardens, you have a 1/5 chance of +5VP, 4/5 chance of -1VP = +0.2VP

So, with 6 or more gardens, it's beneficial to take the copper and curse. At 5, it's an even split. 4 or less, probably not worth it. That being said, we all know that Dominion doesn't work on averages. If you do get the 2 cards you need to go from a deck size of 38 to 40, it could just mean the win!