Our models output attractiveness score for every cryptocurrency traded on exchange. The higher the score is, the more likely is the asset to outperform the broad market in the future. Contrary, the lower the score is, the more likely is the asset to underperform broad market in the future. Therefore, our systems buy assets with high attractiveness scores and short the assets with low attractiveness scores, taking into account expected return, risk, transaction cost and AuM. We construct a portfolio that is market neutral, and profit solely from the predictable patterns in investors' decision-making.