I just ran a test and thought I would share some results. My site has 3 ads on it; a 728x90 banner at the top of the page, a 728x90 banner about 1/3 of the way down and then a 160x600 in the right rail.
The top banner has always had a really low CTR and RPM, but because the pageviews are so high it still brings in decent amount of revenue. But I recently removed the top low CTR banner hoping a couple things would happen.
I've read many people suggest, including AdSense account managers, that the first ad on your pages has the potential to be the most valuable. And I was essentially "wasting" the top ad revenue potential on a location that was my lowest CTR. So I was expecting that my 2nd ad location, which now because my 1st ad spot, would now see a higher CPC and RPM. This ad location has an pretty decent CTR - much better than the top of the page banner. Well I've seen absolutely no improvement on this good ad spot which is now the 1st ad. I have about 5 different types of pages (and hundreds of each style) and no pages have shown any CPC or RPM improvement by removing a very low 1st ad CTR and RPM.
Also, now I have fewer ads and saw no improvement, so there's goes the theory that less ads means higher paying CPC and RPM's. Sure, my site is only 1 site and if 50 people ran the same test they might see improvement, but I was surprised there wasn't ANY improvement.
I currently do not have any ads labeled like I guess I should <div id=”ad1″>. That said, previously I had my worst performing ad coded highest up on my pages. Now I have the best ad coded highest up on my page and am seeing no CPC/RPM increase. Personally, I'm not convinced the algorithm currently works as perhaps it once did. And my guess is that it is more of a best practice than a hard and fast rule.