'Average' is the problem. We have adverts that are in the same industry and are targeted in the same way, yet they perform very differently.
From a statistical viewpoint you could say that 100,000 clicks would be a good amount of data to work with when trying to figure out the differences between the top 10 positions. In practice, we have looked at this data and found no real pattern apart from a very rough-fit increase in CTR.
We have huge anomolies with our Google data but find very clear trends on the data we collect once a user visits our sites. I would say that Google has a system that inadvertantly creates 'blips'. An example of this is an advert that rund as ave pos 4, this could be #3-5 or it could be #1-7, if it's the latter then the times it appears at #1 will skew the % upwards.
My experience shows that, in our industry a position 8 advert will get around 45% of the CTR that position 1 enjoys. Plot a line from 45% to 100% and you will have a VERY rough guide for our industry. However, we have seen very different results for client industries. It's all down to how easy it is to find the information that you advertise for.
The best approach is to analyse your own data and go from there as that is what you will know the most about, and that is where you can make calls on how much you can stand in terms of ROI. As the difference from #1 to #8 can be huge in terms of CPC, it can be a good tactice to broaden your advertising but keep bids low, however if you need to back up the ads with REAL content then you have to factor in the cost of creating that content. There are many factors that will be unique to you and your business.
No-one has an average business.