John Hempton has a great post on the difficulty of "beating the market" even if one possesses superior insight or knowledge. I would just add the following:

It is common among commentators to conflate two very different assertions:

  1. It is extremely difficult to "beat the market"
  2. Markets are efficient.

The common error is to assume that 1 proves 2 which is most definitely not the case. Shleifer and Vishny discussed this in their paper on the "Limits to Arbitrage".

Timing is extremely important especially when taking a short position for many reasons:

  1. Short equity is a short volatility position i.e. limited upside, unlimited downside. At the very least, margining requirements in the interim "inefficient" period may kill you before the market corrects. Long equity positions can atleast be left alone if liquidity permits and if the position is not leveraged.
  2. Principal-agent problem: Fund managers need to get the timing spot on when they are taking on contrarian positions. Else, they will be fired by their investors long before the market corrects.
  3. Even if one is not an agent, simple uncertainty means that we're never certain about our judgement. The longer the market refuses to come around to our viewpoint, the less certain we become and the more tempted we are to liquidate.

As John mentions, timing the market requires not only holding the contrarian view but knowing when this view will dissipate through the market. "Wisdom of the Crowds" explanations of the market require that the uninformed majority hold sufficiently diverse opinions. Given that betting on Kodak's demise is tantamount to betting on a technological paradigm shift, the "crowd" by definition is not diverse and wedded to the old paradigm.

This is essentially the reason most contrarian investors are long-only long-term investors. This is the style of investing that Jack Treynor called betting on "slow travelling ideas". The always excellent Michael Mauboussin has a discussion on wisdom of crowds and slow travelling ideas here and here.