There shouldn't be a problem with using the guid, in this case the guid and origLink are the same, though the guid has the "permalink=false" attribute, which usually doesn't matter, but I try to not use the guid when it uses that directive. But, it comes down to personal taste, tomahtoe, tomaytoe.
Your LEP regular expression should put only the ID itself into field {0}; so your LR should probably include the link upto the ID, if its filling the entire html link into field {0}, the regex engine is being nice to ya.
My regex only grabs what's directly around the digits just because I try leave as much room as possible for site changes--if the link changes at all your regex won't match, mine isn't much more flexible, but either works--it's more a matter of taste.
Your LR leaves a bunch of gunk at the bottom of the entry ("More from Newsweek Health"), so just make sure to adjust your CEP regular expression to account for the extra gunk. I left mine open-ended so that it'd be a little more flexible in case the source html changed at all--but there's nothing wrong with putting something solid on the trailing part of your CEP regular expression.
You do need to include a CEP, when there's a <title> tag in the html that's sent to htmldoc, it'll make a title (even though the cmdline specifies "no title"). The title overrides the filename in the PRS-500's display, so in your "book" listing, it'll show as the contents of the title tag instead of "rss-Newsweek Cover".
There's only one article in the feed at a time, so one article is valid, though I'm going to attribute any errors in my message to it being late--the entry that I published last night to Geekraver's server should be correct. I'm testing a date format fix ATM, so my copy may be parsing dates that V23 isn't.