I can't get my head off of this subject. I'm just not really seeing the 9M or 10M figure anywhere.
For anyone else who cares, this comment thread has an abundance of information. Tango, MGL, Swartz, and others commenting on a pretty good study.
http://tangotiger.com/index.php/site/comments/how-much-do-wins-cost
I also found this 1-2 year old Swartz study that looks like it found about ~6M in 2011 with a trend of roughly ~6% inflation. So say 7.25M this offseason. This corroborates the more recent study from the first post.
http://www.fangraphs.com/blogs/a-retrospective-look-at-the-price-of-a-win-part-two/
More than anything though, I've found the question to be extremely complicated. It's obvious that the cost/win in free agency isn't the cost/win in all of baseball. The most efficient way to add wins is through league min and low arb players. The Rays pay like 2 million bucks per win, or something like that. The Blue Jays probably paid a near free agency rate for all of their wins above replacement last season. A win near the 90 win mark is worth a lot more than a win anywhere else, so value is very relative. $/WAR value calculations might be more dubious than informative in some circumstances. I'm wary of the entire exercise.