The problem is that right now there are very few "serious" sites (i.e. sites with significant content and where an effort has been made to SEO them) using IDN, so the approach of thinking up a keyword and testing it to see if IDN sites show up is like trying to find a needle in a haystack.
I know it sounds a bit circular, but you should first try and find a popular IDN site, then look to see what generic keyword the domain name contains, search on that keyword and see if the IDN site you identified comes up first in the results. That's the only way to do an even halfway "accurate" test at this early stage of the game...
If I had to guess - and it is a guess based on gut feel and experience - I would expect IDN to SEO better in the long term because of the "double helping of keywords effect".
For example, if I have a Japanese site about bookmarks and it's at "bookmarks.com" then links to it might look like this...
Code:
<a href="http://www.bookmarks.com/">ブックマーク</a>
On the other hand, if I owned the Japanese IDN form of "bookmarks" and I was confident that browsers could resolve it, I could link to the site as
Code:
<a href="http://www.ブックマーク.com/">ブックマーク</a>
and suddenly I've got the keyword twice into the link, without overly "stuffing" it.
Also, sites quite often post unsolicited links to other sites based on the domain only rather than some tidy description i.e. Code:
<a href="http://www.dnforum.com/">dnforum.com</a>
- if the domain itself was an IDN even that type of link would automatically contain the keyword!
Bookmarks