Tweakguides did some research on this, and crunched the numbers. The article (which is very long, but very worth reading), is here: http://www.tweakguides.com/Piracy_1.html
As far as methodology goes, I think that's covered here: http://www.tweakguides.com/Piracy_4.html
But, in brief, for the console market, the numbers for a popular game (they used Call of Duty: Modern Warfare 2) in the console market was 1 pirated copy for every six copies sold. For the PC game market over the same period, it was 10 to 12 pirated copies downloaded for every legitimate copy sold (both download and CD/DVD).
Loss is a bit more complicated, but measurable. The most measurable part is the freeloader effect - pirate copies using authentication servers and technical support (and some of the supporting data is the number of copies authenticating on the servers vs. numbers sold, which can be over 10:1 at times), and that's costs of time, server bandwidth, etc.
As far as lost sales, that requires an understanding of the game market. You have hard core fans, who will buy the game no matter what - that's a small core of guaranteed sales. You have hardcore pirates, who will never pay for the game, so there's no point in even trying to market to them. The lost sales take place in a third group - call them "casual gamers." This group is interested in playing the game, but they aren't particular on how they get it. So, if they can't download it, they'll buy it. Most of the sales for a game tend to be in this third group. So, the longer the piracy can be held off, the more sales occur in this group. Zero day piracy wipes any sales from this group off the map.
Now, again, this makes it very difficult to measure exact numbers of lost sales, but you can get a sense through trends. So, if a game had 400,000 sales with piracy held off for two weeks, and the sequel to that game had 50,000 sales with zero-day piracy, you can draw some general conclusions about lost sales, as well as the sizes of the various market groups.