It's an interesting discussion. Everyone seems to state that validation matters, though at the end of the day there's not really a whole lot of evidence to back that up. At least from my personal experience.

I've seen sites with all round similar audiences, similar types of content and even similar styles though one may or may not validate better than the other it made no difference in terms of SEO.

In fact, a site I once worked on put a fair amount of effort into SEO, wasn't a great deal but the thought was still there. In the end, more focus was put onto user experience and even though the design itself never validated perfectly the end result was better. As is said in business it's easier to retain a customer than to obtain a customer.

I am not saying it's good to ignore validation, not at all. In fact I plan on validating KWWH one of these days. But the argument remains whether or not it's critical or even factored into SEO results or have an outcome on a sites popularity. I personally believe that if you can validate a page, validate it. You're not losing out on anything by doing so, except for perhaps the opportunity cost which could have been used on creating a new feature, etc. I guess the concept of "do it once, do it right" will always be there anyway.

Food for thought...