The market is inefficient. So are the techniques for beating it.

David Merkel summarises the complexity that stops markets being efficient.
Efficient market theories are 'highly stylised' models characterised by perfect information, the absence of transaction costs and rational participants. In reality, the opposite is true. The sources of complexity that undermine efficient markets include: Technology, which is costly and imperfectly distributed, opacity, which prevents investors from understanding the risks they are taking, leverage, which is discontinuous and violent when the prospect of bankruptcy looms, and many more…

You can't beat the market using quantitative techniques
I'm not sure what to make of this article in Scientific American. Statistical physicist H Eugen Stanley says mathematical models haven't yet beaten the market and can't predict the future, but can predict risk of 'something really bad' happening and that information is better than nothing. But avoiding blow-outs is half the performance battle so shouldn't risk-based models beat the market? Which of the well known models has he tested and found wanting? The Magic Formula? Piotroski?

After five and half years of lacklustre performance the MFI Diary gives up on the Magic Formula, a quantitative system
Since February 2006 MFI Diary's MFI Index is down 5.5%. The market is down 8.5% prompting the diarist to abandon the formulaic approach.
Mark Carter thinks he knows what's wrong: Nobody knows what the Magic Formula is because its author, Joel Greenblatt, was vague about some of the details. The Return on Capital calculation underestimates the capital a company needs and exaggerates profitability.

Geoff Gannon favours the same method of calculating the earnings yield I do, 10 year average return on equity / price to book value
A point of difference: He uses price to tangible book value, I include intangibles but check to see how significant they are (though that might be about to change for some companies).
A similarity: He requires a minimum 10%, so do I.