You can't tell without access to the sharer's computer, but it's reasonable to assume the most likely number.
The 1.00000etc number in my GP post is off because I forgot that the initial uploader didn't have to download the file. The real number is slightly under 1.0, at 1-(1/N), where N is the total number of complete downloads across the entire network. I imagine a typical file will have a number like 0.9999. Multiply that by the value of the track and then multiply that product to get actual losses if you make the laughable assumption that every copy equals a lost sale.
Yet, even that's going to be off when you look at radically differing upload speeds and things like that.
I still think this is a possible defense avenue. I don't think you can actually prove (in the typical case with multiple up- and down-loaders) that a particular computer ever sends a complete copy of a file to any other single computer. But I don't know the tech deeply enough, hence my questions.
It's been a loophole in the past for shipping weapons around - ship the parts (which is legal) and the recipient assembles them. And since a part of an MP3 file is unusable without the rest, I'd think a similar defense might work here.