However, the long-term future cannot be predicted for the same reasons that the weather can only be predicted three weeks out. The stock market is a nonlinear dynamic system as it contains positive and negative feedbacks. Positive feedback, such as when making a profit after investing in the stock market, causes people to invest money back into the stock market, leading to further purchases that increase the price. Highly complex systems are not always chaotic, but behave predictably for a time and then, seemingly randomly, transition into chaotic behavior. These types of systems can be mapped using simple chaotic systems that often exhibit patterns called strange attractors that demonstrate that the system jumps into different modes of behavior. Chaos in stock markets is caused by human trading psychology which is never completely rational due to many external factors. By analyzing statistical data, it is possible to find fractals which are infinitely complex patterns that are self-similar at different scales. These fractals are created by repeating a simple process over and over in a continuous loop, and due to the simplicity of fractals, they can be used to predict the short-term future. Long-term prediction is also virtually impossible, just like the weather, for similar reasons. The butterfly effect means that variables that seemingly have a minimal effect on the overall outcome of the stock market slowly have a larger effect on the outcome. Therefore, the short-term future of the stock market can be predicted using Lorenz attractors and fractals, however the lack of information causes long-term predictions to be virtually inadequate.