A pretty quick and easy regression process, based on an admittedly limited sample of players who signed free agent-like contracts (read: not buying out their arbitration years, which generally comes cheaper) to be closers over the past few years. Things to know:
$(millions): Average annual value of the contract
WXRL: Wins added over a replacement-level pitcher. The methodology is similar to that which goes into the Win Expectancy charts that I posted over the course of the season.
We end up with an r value of 0.81, implying a strong correlation. By plugging Guardado's WXRL of 2.71 into the equation for X, we end up with a final figure of 4.19 - that is, the chart thinks that, based on his 2005 performance, Guardado should/will earn $4.19m per season under whatever contract he signs this winter.
(Note that there was a stronger correlation between $(millions) and WXRL than between something more standard, like $(millions) and Saves.)
I'm a little skeptical, but it's worth remembering that Eddie did pitch with a busted shoulder last year, and ended the season in something of a rut that could scare away potential suitors. He may end up getting more in free agency than his 2006 player option would otherwise provide, but I don't think it'll be by all that much.