Indeed, OS X goes down, and obviously none of us actually believe that. But not only does Statcounter report that clearly faulty number, but they have yet to fix the problem.

This happens all the time. When their numbers are clearly wrong, they don't care about the numbers enough to fix even the glaring problems, their sample is unsound, and their methodology is unpublished, why exactly are we supposed to give any of their numbers any credence?

What you've written is the first I've heard of a recent change to the Safari on OS X user-agent string, and I see no indication of it in my access logs. What's it supposed to be now? It seems a bit unlikely, and given Safari never ran on classic Mac OS, it seems like a company that's supposed to specialize in analytics should be able to handle it...