The search engine page can be used for validation, just allow people pressing the back button on the page to tell you whether the results were useful or not.
What was being proposed was a way of decentralising the crawling. I tried to demonstrate with some examples why that could not work: you'd end up with an extremely inefficient index. What you're proposing does not solve any of those problems. Sure, you'll get a weak signal about page quality, but far too late in the pipeline to inform the decentralized crawling and indexing.
But further, you are not really thinking through how one would abuse this kind of a feature. If doing seo, I wouldn't forge a page to have content that make it be returned for irrelevant searches. Instead I would forge some high quality pages to show up as having backlinks to my page, and boost its pagerank. Or to demote the page of people I dislike, I'd forge it to have results that make it not show up on any searches. Your heuristic would not work there: if there's no clicks in the first place, there can't be any bounces.
But further, you are not really thinking through how one would abuse this kind of a feature. If doing seo, I wouldn't forge a page to have content that make it be returned for irrelevant searches. Instead I would forge some high quality pages to show up as having backlinks to my page, and boost its pagerank. Or to demote the page of people I dislike, I'd forge it to have results that make it not show up on any searches. Your heuristic would not work there: if there's no clicks in the first place, there can't be any bounces.