The Results Are In! New Test Results!!!

Compar critique is well founded, but that doesn't mean that this is a wasted effort. Just as you can't prove (at least at this point) that this will work in competitive environment, Compar can't prove otherwise. Further testing is required, and I suggest the following - it requires more work and you must use X number of domains (new sites) rather than just pages off a single domain.
Replace current kwds with some that are not completely unique , but also not too competitive. In this way you will have some other results interspersed with your own. You are testing for two things
1. Repeatability - if your results are in same relative order then you are probably correct in your conclusions. This still doesn't mean that the conclusions are significant.
2. Significance - if your results are all "bunched up" together - there is no significance - it means that if you had just one site, and changed these factors sequentially, you would still rank the same. If they are interleaved with (a lot) of other sites, then there is significance, because one you your methods actually outranked other sites (that are not yours)

Practical problems with this - sandbox (maybe avoidable if it is a low competition term).
Anybody else linking to you (messing up your backlink count)- probably not a big chance.
Not all new sites crawled at the same time - they probably will be if there aren' t too many - still watch for it (new page boost etc...). Obviously you need to link to them from a single page of yours.

Good luck - I wanted to do something like this - but there is never enough time.
To the harsh critics - c'mon - this is interesting! Be constructive. When there are too many variables you isolate , repeat, automate ...

 

Good and informative. Thanks for the post.

 

I agree! But it is necessary to point out that the results found are in no way 100% conclusive. I think it's a great test though and I would be interested in reading any future results.

 

Great test. Thanks for the info

 

Took me long enough to respond to this, but if you really need a question like that answered you may need to do some reading on forums progression ignited from community contributions.

In testing for the community as the OP did, not only has he created an example for others to see and conduct their own experiments based off of with, but he's also contributed information that will excellerate further discussion of the matter.