I have read it somewhere, but I seem to have lost the link and I would like to confirm that I remember it correctly. So is it true that algorithmic trading strategies are basically of two types: 1. Modelling of stock markets: This type relies on modelling the market using mathematics and physics. Usually, high frequency trading algos are implemented this way. An example is candle stick analysis. 2. Data mining of stock markets: This type doesn't attempt to model the market, but rather learn from historical data using machine learning and artificial intelligence techniques. An example is support vector machines. Is this correct?