I'm not so familiar with this area but my guess is that if you turned used noarchive, Google would not cache the page at all and therefore would not be able to use the text in your page as keywords for search results. So most sites therefore did not use noarchive because it improved discoverability/SEO to allow Google to cache your site. This is just a guess though and what I always assumed to be the case. This seems to be the case though because the cached versions would often contain the entire article for free, which makes no sense unless they were doing it for SEO. For example you could use this trick to read any nikkei article.
I didn't mean to say that noarchive is only a google thing. My only point was that my assumption is that until now if Google didn't cache your site its contents would not be used for SEO.
