Have you heard of Didier Sornette?

opmtrader,

That's interesting. Would you mind sharing what you have understood from his work? All I can gather is that he is saying that the bear isn't over by far. ok that's good to know!

But seriously, what else is this guy saying? I'm just curious. I don't think I can really use it for trading per se but it is a refreshing perspective.

TIA

 

I do not find this "prediction" unique, early or original. Half the junk mail I receive from cycle cranks have made the same comparisons and predictions for years. I saw a fantastic comparison between the naz and the early 80s biotech bubble and the nikkei, right at the top and I even saw the followup tracking the sell off. It allowed me to make the best trade of my career. So while this guy may be smart, I think he needs to show a little longevity.

I think stockcharts.com has a least a couple of chartists who have been making this claim for a while. Decisionpoint had links to this stuff. Mark Boucher has said these things. Lots of news letters. I just can't get enthused for some guy with some unique equations. Let us put him to the test for a few years.

 

I've read this stuff. As jem says there is nothing conceptually new here, there's a lot of cycle research out there already, and the so called predictions are neither accurate nor frequent enough to build a profitable trading strategy. 2 trade signals each century is not going to make money to anybody except maybe to Methuselah.

I don't care much about the math part either, in my opinion it's hopeless to capture market dynamics adequately. But I have to recognize it's an improvement in comparison to the dumb autoregressions and SDE's they teach in academic finance classes.

What ticks me off about this guy is his condescendence to TA people. This guy will earn his right to be arrogant only when he has trade results to show. It's the same thing about Mandelbrot, no trading record that I know of, but nevertheless he feels entitled to call Elliott a peripathetic accountant or somthing like that, why, because he won a Wolf prize and Elliott didn't?

The problem is that he won the Wolf prize for "inventing" fractals, and that's clearly not the truth. Fractals had been around since forever, Gann already knew the concept to give an example. So maybe the idiot is Mandelbrot first for winning a prize for something he didn't do and second for being unable to make money with his own stuff.

Finally, somebody implied that DS was delving deeper than Gann. That's a gross overstatement of DS achievements to say the least.

 

Babak, jem, and buzzy2,

I respect your opinions and your criticisms have much validity. First and foremost you are correct in saying that it is difficult to create a working strategy from Sornette's work. Second it is true that the frequency of these monumental crashes, and following anti bubble behavior, is not frequent enough to satisfy the trading needs of most of us daytraders but their frequency seems to be increasing in our volatile world. The best way I can describe the underlying concept, and remember I am just a fan and not an expert, is that financial crashes exhibit the same types of patterns or behaviors found in many other natural phenomena. The same type of minor perturbations that set off earthquakes or kick off full blown fracture in a metal stress test or set off an avalanche, etc are the same things that can trigger an unstable market into a crash. Seemingly independent agents (in this case traders) will actually have an impact on one another, through information transfer, conversation, influence, and of course the feedback loop that is the markets. Out of this sea of supposedly independent agents can arise a polarization of opinion and action, an “emergence” if you will, that can set off a crash. These things are not supposed to be possible in the normal framework of EMH, Gaussian distribution, random walk, etc.

What is more interesting than examining the actual crash period is going through the equations Sornette uses to establish the behaviors of traders during the buildup before the crash. The underlying idea that I took away from it was that at the point of the crash the return expected from a rational trader/agent becomes nearly infinite to compensate for the immense risks he is running. At this point the entire system becomes unsustainable and crashes.

To really grasp the potential application of this work for yourself you have to develop your own novel approach for detecting the same pre-crash signal that Mr. Sornette is. With this system in place you should be able to theoretically predict crashes and receive a number of other equally important pieces of information if you are creative. If you read his papers you will find that his method was able to predict many crashes within an accuracy of 3 days even months beforehand.

Now I don’t think this guy has the infamous “Line to God” or anything concerning predicting each and every market move, but I do think that the underlying premise, that during great manias and subsequent crashes the markets resemble other natural systems, is right on target.

 

opm I think we are in agreement. I am just concerned about survivorship bias in his methods. Let us see it work going forward. Real time with real calls. or real results. This board does not let regular traders make farfetched claims and I see no reason to subject a phd to any lower standard. It is all in the real time profit and loss statement. My crash detector is probably better than his I just did not make a large enough bet nor did I hold the position long enough. (See what I mean you do not believe me and yet I too have a post graduate degree). I'll let you know part of it, inverted yield curves---market p/es are at historical highs/ and on average for a few weeks the market closes near its lows. Let us call this the Jem market crash indicator. In twenty years we can put it to the test). The bottom is reached when we reach pre bubble prices.

In the mean time we will scan markets for blow off tops and a tightening of monetary policy and markets closing on their lows. Now we can also look for cycle translations on all of the time frames we look at. With periodic rallies that last 20 to 50 percent in time and price. Again, until we get back to pre bubble prices then we can see consolidation as the public losses interest and the smart money starts accumulating.

I could write the same thing twenty different ways adn if I had a PHD in math I could program this stuff and claim, I am the man, but you should not believe me.

All I am saying is that this stuff about agents and chaos theory is just new words for the same old stuff until proven otherwise.

 

I did not know that he is at UCLA now, I will have to pay him a visit then. I studied some of his papers in mathematical physics (some variants of nonlinear Schroedinger equation) at the time the nonlinear SE was of interest to me, but I am not so familiar with his work in complex systems. His work in this area is part of a larger field that is called econophysics. You can find much more about it by doing a Google search on this topic. Some of my friends work in this field. It is one of the hottest fields in physics these days.

 

I disagree. The laws of economics are as universal as the laws of nature. The human nature does not change, the greed and fear have been around for quite a time by now and will stay with us for some more time to come. The same is true with the law of supply and demand, the most fundamental law of economics.

However science has some intrinsic limitations. For instance, this type of equations may exhibit a chaotic behavior and so we may not be able to predict things with a desirable accuracy well in advance. The same problem concerns the weather predictions where the problem of chaos was discovered (the Lorentz equations with a famous Lorentz attractor). Because of that we cannot predict the weather well in advance even in principle.

 

Interesting article from thestreet.com, apparently this guy is getting more and more publicity as his book is published.

You might be interested in reading this as it has some specific 'predictions' and outlook for the near/intermediate future of the market:

http://www.thestreet.com/funds/supermodels/10063045.html

 

I've got an academic paper from a journal Physica (quants) saying price behaviour just before crashes becomes more and more smooth in a daily horizon but less and less orderly - erratic in hourly and minute data. This kind of "buzzing" under a calm surface preceeded both 87 and some subsequent mini-crashes. The author is a guy named Zbigniev R Struzik

 

Yes it's true