We can see a similar process in the stock market. If you find an Alpha signal that suggests that a stock will have positive future returns, the assumption of future higher returns means that the price today is lower then the expected future price. In other words, its current market price is cheap according to the signal. If you buy it now at this price, and if your hypothesis was accurate, you’ll be expecting the price to rise in the future. Now, what if lots of other investors also find the same signal and also decide to buy the stock now? The buying activity will push the stock price higher. So, by the time you get around to buying yourself, the price has already risen. If you buy it like everyone else, you may find that it doesn’t produce the positive return you were hoping for. This is an example of the efficient market hypothesis, which states that assets are fairly priced based on publicly available information. The factor is no longer helping to drive the mean return of our investments. Instead, the factor is triggering the market price to move because it is known and being used for trade decisions by many other investors. So this factor is now driving the movement of the stock. Every time the factor signal changes, the market price also changes soon after. We now say that the factor is a risk factor. So in summary, good Alpha factors enable us to find mispricings and seek a competitive edge in the market, which drives the mean return of a portfolio. When Alpha factors become two well-known, then they end up triggering the market to move as a changes, which drives the variance of our portfolio’s return, but no longer the mean of the return. So now it’s a risk factor. Market participants historically have had to use their judgment and experience to decide which factors to use as Alpha factors and which factors to use as risk factors. Can we use AI to do this instead? Yes we can. We’ll go in depth into Alpha factor combination in term two.