The placebo and nocebo effects highlight the importance of expectations in modulating pain perception, but in every day life we don't need an external source of information to form expectations about pain. The brain can learn to predict pain in a more fundamental way, simply by experiencing fluctuating, non-random streams of noxious inputs, and extracting their temporal regularities. This process is called statistical learning. Here we address two key open questions: (1) does statistical learning modulate pain perception? and (2) is it different in people with chronic musculoskeletal pain? In a first experiment, we asked 27 participants to both rate and predict pain intensity levels in sequences of fluctuating heat pain. Using a computational approach, we show that probabilistic expectations and confidence were used to weight pain perception and prediction. Given that statistical learning involves supramodal processes, we developed an online, stock market game to assess the ability to explicitly predict volatile and stochastic time series, probing the most fundamental components of statistical learning. The game was played by 56 chronic back pain and 55 healthy participants. We show that back pain participants learn the statistics of the sequence more slowly than controls. In conclusion, this study shows that statistical learning shapes pain experience and can be disrupted in common chronic pain conditions, opening a new path of research into the brain mechanisms of pain regulation