Does anyone have statistical results that compare the performance of different algorithms used to predict forex market?

I'm basically interested in ANN, Markov chains, Time series analysis and any other statistical / Soft computing methods. Some research papers say it may be possible to predict the market just by looking at the past data (trend analysis I think ,) but I can't find a paper which compares different approaches used by researchers. If any one can help me on this area I would be really grateful.