well technically the pages are relevant, as with XML feeds you can not place keywords in there that ae not relavant to that page, wherethere the keyword is in the page content or not.
The XML feeds are reviewed by Yahoo "human" editors that check the relevancy of all keywords added to the page, to make sure that this is not abused.
I realize that. I'm not talking about spamdexing. Since it's my heartfelt belief that on-page optimization is good for spiders and
people, I believe that people who search on keyword1 keyword2 keyword3, and click through to a page on the SERP, are at least somewhat satisfied when they scan the page and find a heading (or some text that represents a heading) containing at least some of those keywords. Synonyms are fine, but not as good, at least when it comes to this one specific purpose. That is to say, a page about "dog" is more relevant to "dog" if the word "dog" is on it than a page with the word "canine" on it, all other things being equal. It more immediately satisfies the request of the searcher.
So my point is just that pages that are deemed relevant because they're about the queried phrase, but don't contain the queried phrase, are getting a boost that the user is unaware of. That's also potentially true of a page that has a great many links with the queried phrase in the anchor text, even if the phrase isn't on the page, but that's not a deal that the site owner made with the search engine -- it's taking advantage of a weakness in the algo, and as such is only partly the responsibility of the search engine. It's something they need to fix, rather than something they're getting paid to allow.
So sure, the pages are relevant. The editors make sure of that, and I'm certainly not claiming they're dishonest in that role. I'm just saying that the pages themselves are not as alogrithmically relevant as pages they're appearing above, and it's strictly because of the exchange of money. That should be made clear to the user.