Using the custom data from Quandl, I've selected 2 inversely correlated assets to build a simple strategy where the algorithm buys the asset that has performed the best in the last 9 days, and closes any other open positions.

The assets are from the Indian market: The NIFTY index and the USD/INR currency pair.

Besides of having an interesting Sharpe ratio and nice returns, I've tested the resilience of the results by shifting the rotating window from 3 days up to a 30 day window to optimize the variable for the best performance:

The resulting sharpe is fairly robust regardless of the precise value of the rotating window period.

I'm absolutely sure the strategy could perform way better; e.g. by refreshing the number of days it should consider for the calculation window, the criteria of buying and selling, and increasing the pool of assets to choose (maybe recalculating the correlation between assets, and choosing the strongest ones?), etc.