I think any metrics would be almost meaningless unless someone were testing with banners that were absolutely identical except that one was hardcore and one was softcore... and that is usually not the case.

If you test 5 different banners for the same site, you'll get 5 wildly different conversion rate... and further, you may even find that banner 1 outperforms banner 2 using traffic source A, while banner 2 outpulls banner 1 using traffic source B.

Not trying to be unhelpful, but I hear all the time somebody say "Oh, non-hardcore doesn't sell" or "Oh, it never sells as well" or whatever, and it has so much more to do with the quality of the models, banners, and traffic that it gets reallly difficult to compare.

If anyone does have any actual apples-to-apples comparision data, I'd love to know more, because traffic metrics have always fascinated me.