Dreamers used to believe they might predict financial prices exactly, but settled down to believe in predictions that are "mildly" in error yet allow riskless portfolios. The actual variation of prices is "wild" in many ways, but close to the speaker's latest model, which will be sketched.
That model centers on the concept of multifractal time. No longer a recent sensation, fractals/multifractals are, in many fields, not only the best models, but the only sensible ones. The internet, a recent technology, is also a recent addition to an already long list of cases (of which will be sketched), where fractality holds over a very broad range of scales, as measured by the ratio from the largest to the smallest. These cases range from the works of nature to the works of man.