See my analogy with gene-manipulated food, in my earlier posts in this thread.
One test is no test. Ten tests are no test. Even a thousand tests are no test. A few million tests... now, we're talking about
some kind of evidence! However, on the level of a few million tests Starforce was counter-proven. There's no point about more arguments or stupid contests from the developers. Case closed, period!
With my analogy, the manufacturing of gene-manipulated food is too cheap to stop only because a few ten people out of every million die because of eating it. Wow, talk about justification!
[Edit] Oh, and as for test systems, fuck the virgin ones. Whoops!

I mean, systems used day by day are the ideal test systems. Such a low-level system, that Starforce is, has to be able to cope with unusual settings, drivers, components, whatever. Which is why it should be very careful about what it does. Which it isn't...