More about the predictability

Predictability (P) is obtained by calculating the correlation between the past predictions and the actual asset movement for each discrete time period. The algorithm then averages the results of all the predictions, giving more weight to more recent performances. As the machine keeps learning and amassing more data, the values of (P) generally increase.

Predictability is measured on a scale ranging from negative 1 to positive 1; this metric is an adaptation of the Pearson correlation coefficient.
P=-1 means the actual market moved in the opposite direction than the algorithm predicted.
P=0 means that there is no correlation between the prediction and the actual market movement.
P=1 means that there is perfect correlation between the actual market movement and its predicted movement.

Any value of P above zero indicates a positive predictability, the higher the better.

Reading the Prediction

It is recommended that investors consider both the signal strength and predictability, as a highly predictable stock that barely moves and an unpredictable stock that is projected to move drastically both make unattractive investments. It is important to note that longer-term forecasts (1-month or more) tend to have higher predictabilities as the algorithm can more easily spot long-term trends.

Every day the algorithm generates a heat map demonstrating the overall direction of the market. For example, a bull market would be indicated when green “buy” predictions at the top of the table dominate the red “sell” predictions at the bottom.