Econophysicist Predicts Date of Chinese Stock Market Collapse (Next Week) | Hacker News Reader

Econophysicist Predicts Date of Chinese Stock Market Collapse (Next Week)

technologyreview.com

23 pointsAnon8423 comments

I predict the Shanghai Composite Index will NOT burst between the 17th and 27th July. If my prediction is correct I will happily predict lottery numbers for anyone who asks.

I predict there is a 50% chance you are either right or wrong.

I predict there is a 100% chance that he is either right or wrong.

Such kind of predictions have a funny impact. Imagine if people start believing the prediction 100%, they would start selling out or short the market. And those actions collectively would result in a crash.

Self-fulfilling prophecies, they're called.

All the more reason to make sure that everyone hears you.

Perhaps that's what he intends... Maybe the reason he isn't giving the reason he is predicting the collapse is because the reason is to see what effect his prediction has on the market.

The article says economics (well, graphs of stock prices) are scale free. Is this true? Seems hard to prove, can anyone point to any research?

I have some books on complex networks and some on computational finance, it's common knowledge that such graphs are scale invariant (and not hard to show, because you just measure relations of ups and downs and volatility a.s.o.and see that they are independent from the absolute numbers on the graph scale). Many people like the works of Benoit Mandelbrot on fractals in finance. Google these keywords.

Can you recommend any?

http://www.amazon.com/Mis-behavior-Markets-Benoit-Mandelbrot...

A fun exercise is to hook a randomwalk up to a pretty charting package, show the resulting charts to a finance person, and watch them ask "What stock is that?" I did that all the time when I was working at a financial software startup. (Mostly because I was working on our charting library and needed fake data to test...)

Benoit Mandlebrot said that in the sixties.

But the scale-free quality does not imply predictability. In fact, I suspect it implies the opposite.

Indeed it is true, except for very short periods of time where it begins to break down. In quantitative finance one can model the market or an individual stock as a Brownian motion with a long term upwards drift. The results look startlingly similar to what we're used to seeing on actual stock charts.

Econophysicist?

Does he study the supply and demand for pendula and inclined planes?

Econophysicists, roughly speaking, model economic systems using techniques from statistical mechanics; not nearly as silly an idea as it might seem at first blush.

a) 'twas a joke, sir. :)

and b) the idea may not be silly but the name sure is. By this nomenclature, all current economists should more properly be called "Econostatisticians" or "Econopsychologists" (to quote Charlie Munger speaking about the silliness of the term "Behavioral Economics": "If economics isn't behavioral, then what the hell is it?!?")

No offense meant, of course. I just think it's a bit humorous. He's an economist who uses methods from physics, that's all - no need to make up a new name for it (let alone a name that sounds even dopier than the previous name).

Yeah, sorry about missing the intent of the joke :) I'm cynical enough to think people'll just take pot-shots at scientists playing outside their perceived core expertises just on principle. Not that I'm a scientist turned startupper or anything.

Anyway, it's a pretentious name, but I think it's a bit like "systems biology" (biology done by physicists). If physicists do it, they like inventing a new name.

Of course, there's masses of that kind of thing, and I guess I'm a bit more sensitive to it than most; the research I did when I was still a researcher could have been called quantum chemistry, solid state physics, materials science, or what it was actually called (theoretical mineralogy) and still have been the same damn thing! (Predicting how glass works.)

"systems biology" (biology done by physicists). If physicists do it, they like inventing a new name.

Or biophysics, nanobiology, "human dynamics", etc... sexy names make for sexy media attention (and grants).

<joke> Also, we like to think of it not as "X done by physicists" but rather "X done right" </joke>

Oddly enough, a lot of econophysics went into justifying the recent bubble and the creators of heat-equation-inspired Black-Scholes equations were directly responsible for the LTCM collapse in 1998.

Econophysics doesn't have a "not snake oil" feel to it...

Reading "When Genius Failed" or whatever that book was that discussed the fall of LTCM, I thought LTCM failed because they started doing things their original mathematical models didn't cover, and assuming that Russia wouldn't default on its debt. They started doing straight up directional trading and other things that weren't safe or something to that effect.

I just have to comment here.

There's a journalist style that loves to blame failure on some personal quality rather than on the structure of the system itself. Certainly, when it fell apart, I'm sure if one looked hard enough, one could find a "mistake" that LTCM made other than following an unsustainable system. But the invariant is that escalating leverage and escalating speculation is ultimately unsustainable, ultimately identical to a Ponzi scheme and will fail sooner or later - I mean, I'm sure you could find the "mistake" that brought Madoff down.

There's a total solar eclipse over China on the 22nd, exactly midway between the forecast dates. Quite a coincidence, if that's what it is.

In any event the prediction is NOT for the market to tank within the quoted timeframe ... rather that the market will PEAK between 7/17 and 7/27. This would be followed by a period of weakness possibly leading to a crash. If the experience of the USA in 1929 and 1987 is any guide, you would look for a crash low 55 days later (9/15) or possibly 89 days later (10/19).

Then again, maybe not.

Reply on news.ycombinator.com