no just came across it, while looking into machine learning. http://www.financial-hacker.com/hackers-tools-zorro-and-r/
But do we need to analyse machine learning Sir? price changes are not random and markets are being manipulated and it's not that hard to predict next price movements. At least, I find forex trading, pound in particular, not hard to trade. Imho, what they talk about algo traders in public isn't really true. Otherwise, there is no way I could close all my trades in green last week without a single day loss.
The studies are showing the smallest timeframes have predictable behavior, and the most efficient timeframe with the lowest cost, is the daily chart.
I don't believe they carried out a meaningful study Sir and as far as know, there is no single study that has examined the behaviour of market makers in small timeframes with actual data. Mostly used a proxy which doesn't capture the actual behaviour of market makers. Only my study in the whole world has used actual data and I am trying to get it published some time next year Sir. There is no such thing as an efficient time frame if you are referring to random walk theory. All markets are being manipulated Sir.
Some of the stuff on that site is really disconnected from the reality (e.g. the HFT post and the options post). I think that happens whenever a software person decides to get involved in this business, unfortunately.