The weak form is basically true: if Alcoa, say, announces a better than expected quarterly profit, Alcoa's stock price will very rapidly reflect that, as will its bond prices, and so forth. It’s not like the market doesn’t get around to reading the newspapers until the weekend or whatever. There are professional arbitrageurs who make sure that all markets stay in sync.
But that’s pretty trivial.
What has always been missing from discussions of the Efficient Markets Hypothesis is the question of the quality of the interpretation of the publicly available information. In 1999, for instance, the markets interpreted the information available about Pets.com as indicating that it would generate a lot of positive cash flow over the years, and thus its stock was worth a lot of money.
Well, their interpretation was wrong.
In 2006, the markets interpreted the publicly available information to come to the interpretation that those subprime borrowers in the Sand States would, on the whole, pay back their borrowings.
Now, subprimes were less of a pure failure of interpretation of available information than Pets.com, since with subprime there was a lot of fraud and fiduciary negligence all up and down the line, from the borrowers to the investment banks.
But the basic problem for the Efficient Markets Hypothesis was the same: the markets misinterpreted the information available. There was plenty of information available to point out that the people borrowing a half million dollars with zero down payments in California in 2006 were unlikely to ever be able to earn enough money to pay it back, nor were they likely to be able to unload their hot potato to an even Greater Fool, because who wants to pay a huge amount of money to live in a neighborhood full of poor deadbeats?