Originally Posted by
.Ix
I did some math regarding the second arbor, and it appears that the problem with it is the fact that it pushes our "mulligan" percentage to nearly 50%. With 14 "forests" and 4 dead cards (2 hoof 2 arbor, let's face it, no one likes these in hand), we have 48.27% of either getting no lands, or getting at least 1 of 4 dead cards. That's terrible for a deck with no way to fix its hand. That said, running one less dead card only drops the probability to 41.03%, which is still really awful. It's just not worth the jump IMO. These probabilities don't even count the chance to draw the dead cards after the opening hand, which obviously increases with more dead cards in the deck. The chance to draw other bad openers like 4-5 land hands, or hands full of spells and no creatures is also not accounted for in these probabilities.