Does anyone know of any paper that describes a reproducible method of generating above normal returns in the mature western financial markets? Nope. Me neither.
[1] Fama and French, 1992. The Cross-Section of Expected Stock Returns. http://faculty.som.yale.edu/zhiwuchen/Investments/Fama-92.pd...
[2] Jegadeesh and Titman, 1993. Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency . https://www.jstor.org/stable/2328882
These don't use machine learning but they do answer your question.
> The strategy is able to nearly double the investment in less than 60 day period when run against real data trace.
http://cs229.stanford.edu/proj2015/029_report.pdf "Algorithmic Trading of Cryptocurrency Based on Twitter Sentiment Analysis"
http://journals.plos.org/plosone/article?id=10.1371/journal.... "When Bitcoin encounters information in an online forum: Using text mining to analyse user opinions and predict value fluctuation"
https://pdfs.semanticscholar.org/e065/3631b4a476abf5276a264f... "Automated Bitcoin Trading via Machine Learning Algorithms"
From the paper "Bayesian regression and Bitcoin".
I think are more relevant question is: This is clearly not the case in the real world - so why does it appear to be like that?
But thanks for the links; I will enjoy reading through them.