Gary's point is right on -- once you've proven a design, USE IT. I suspect that getting a dozen rovers for $1 billion is exaggerated, since launch costs alone would eat up 2/3 of that (esp since future oppositions won't be so favorable), and there are operational expenses in addition to production costs. But four or five? Seems likely.

One quibble: assuming the same limits held for future MERs, they'd still be limited to a pretty narrow band in the tropics. Sun would be too low for a polar mission. We need nukes!

One historical note: sometimes the test articles are pretty damn close to being launchable. There was a third Viking lander built, and it was referred to as a "flight spare" rather than a "test article". I read a book by a long-time former JPL mission manager, and as he described it the thing was pretty much operational. After Viking's success in 1976 he fought to get money to send the third one, and even had done basic design work on sticking some motorized treads onto the landing legs. Yes, a mobile Viking! Would have been very crude compard to MER, much less mobility and autonomy, but still cool.

Anyway, no cash forthcoming, and so an operational Viking lander now sits in the Smithsonian rather than on Mars where it belongs.