Is Sornette going to reset his anti-bubble, or was he just caught by a W bottom? http://www.ess.ucla.edu/faculty/sornette/prediction/index.asp#prediction
Maybe he'll be the new offensive corrodinator for their football team and apply the theory there. Makes about as much sense.
Huge problems with the big sports programs, and I can't imagine that alumni are happy with the way that Head Football Coach Karl Dorrell had his team prepared against Fresno State in their Bowl Game . . . Only 96 yards passing, and only 68 yards rushing! What a joke! UCLA Mens Basketball just lost six straight before defeating UW 80-75 last Saturday. Meanwhile, the CAL Bears are in sole-possession of the Pac-10 with 3 starting freshman, having knocked off Arizona recently, leading wire-to-wire. The Bears take on undefeated Stanford at Haas Pavilion in Berkeley this Saturday. In their last meeting down on The Farm, the Bears were down by 5 with one-minute remaining in the game. Look for a much tighter game this Saturday on the Bears homecourt!
Take note that he subscribes to the random theory. Also, he is using a purely mathmatical model with no fundamental basis. I notice many of these purely mathematical models seem to fail as the ramdom-walkers are constantly proven wrong. Also note that he is a geologist; no formal training in econometrics or finance. Perhaps he should stick to looking at rocks.
Don't rule Sornette out completely. Yes the S&P seems to be deviating from its anti-bubble type path. What you probably don't realize is his models calling the runup in China over the few three months. He hit that on the nose when the index was bottoming. Wish I had found that paper as soon as it was published. Ultimately the jury is still out on his overall record. If only we knew the results of the fund his research fuels. Something tells me he is doing just fine.
I think his statement "We have then fitted each of these 10 synthetic noisy clones of the S&P 500 by our log-periodic formula" gives the key. Looks like he may be doing some variation of curve fitting which I can't imagine will ever work consistently...