Someone brought up an interesting point in another thread that, at least according to them, about 98% of all manuscripts submitted to publishers are complete crap and have no right being there. Having done tech editing for 5+ years I can completely relate. I've definitely seen my fair share of "OMGWTF!?" type submissions that are so bad it's a felony to burn them because you'd be releasing toxic pollution on an epic scale.

:P (lol, just kidding guys, don't everyone suddenly grab pitch forks and start attacking the castle. hehe)
But seriously, in all abject honesty, of all the manuscripts submitted to most publishing houses, how many do you truly believe are viable manuscripts? IE, manuscripts that are of sufficient quality to be worthy of publishing. And I'm not saying that anyone should run around saying "all manuscripts, except mine" or anything like that. I'm simply looking for an objective, realistic estimate. Would the 98% figure be accurate (ie, 98% of manuscripts are either insufficient or complete junk vs 2% that are quality, or at least good enough to publish), or is that too high, too low, or just right?
I'm interested in hearing your opinions. And please, keep it civil, as I'm more than certain that some comments in a thread like this could easily step on some toes, so play nice.