@EPinTexas,

A backdoor newspaper industry deal.  I've alluded to that way earlier in this thread.

Interesting that CNN reported today on this from the result side per se:
"The Online Publishers Association, a group of content producers comprising many of the Internet's largest properties (including CNN.com), estimates that the algorithm change shifted $1 billion in annual revenue."

see: http://money.cnn.com/2011/03/08/technology/google_algorithm_change/index.htm?section=money_topstories

Now how can a member association quickly whip a number out like this?  Follow the money I say, it's going to their publishing mafia.

Just look at the OPA's membership, hardly what I'd call online publishers.  Most of these dinosaurs are leftovers from the time when we stopped pressing information on stones:
https://64.13.250.16/current-members

About.com stands out and is slathered all over the results, highly.
Associated Press, didn't they threaten to go paywall on Google and Google cut a deal to pay them for their content?
Huffington Post?  People have aired their content farming history out already.

That membership is a collection of media monopolies who own many sites beyond what's there, which are high on search steroids right now.

How many of those companies have deals with Google and/or suits over content theft by Google?

Coincidence, maybe yet another one.  Certainly a clear pattern.

My hunch, Peter Norvig isn't in charge of Search Quality anymore and as a result, others have decided to  change from "manually crafted formulas" for search results and switched to a "machine learned model".

Manually crafted models are less susceptible to such catastrophic errors, so said Norvig.

If we have intentional distortion variables mixed in with the machine learning, like meeting a quota to pump traffic to the above media horde  clearly the machine learning has to be reporting to Google something is likely way wrong.

Iterative changes aren't going to fix this matter as machine language isn't a literal linear matter.  So Google is either going to need to roll back and retreat or continue to monkey with ML models and yo-yo everyone up and down until they feel they have accomplished something.

Right now what is massively appearing in search is useless (my experience searching for other things).  It's like 1996 search quality with a heap of steaming crap from social all around it.

I'll say it one more time, the targets of Google's attack, hit don't appear to be as penalized as small publishers of original content.   Stolen content trumps the original all over the place.

Instead of content farms, Google has created by defect or via intent much larger content plantations.

I want Google to own their problem.  So far what they've said pre-roll out and until now is incorrect, deceptive, mostly useless and without regard for the well being of the community or simply, their experiment in ML has failed.   Everyone can't fix what they don't know is broken, Google can though and refuses to.