I'd guess the percentage of approaches considered "quacky" which are later proven to work, is fairly constant over the ages. I'd also guess 80/20.