Aug

24

I read a very entertaining article in Scientific American yesterday: "Can Math Beat Financial Markets ".

Stefan Jovanovich writes: 

Check out the author's home page.

Gary Rogan writes: 

Looking at the page brings up the age-old question: can the same techniques that are used in natural science to study the universe that doesn't change from being studied and where the fundamental rules don't change at all be applied to a system where nothing prevents most rules (other than that old "human nature") from being changed at any time? He cut his teeth on never-changing, but how will it play with ever-changing? Somehow fractal coastlines are just not the same as fractal security price charts. Fat tails and non-gaussian distributions are great, but then what? We can evidently estimate the risk of something happening he says. But can we???

Ralph Vince adds: 

Yes, when he speaks of their raison d'etre "To quantify risk" I think, "Huh? How? VAR? Just HOW DO they 'Quantify Risk.'

If alluding to using it ("algorithmic trading," in the context of some sort of proce prediction a la quant, I assume he means modelling prices using models based on SDEs. Again, "Huh? How? Does this guy know what he is talking about?")

Quant-dom, traces it's roots to pricing of various securities — warrants, options, spread on futures/forwards, backwardations, and the pricing of the plethora of derivative creatures who climbed out of the ooze of ercent decades. This is NOT predicting markets, or "Quantifying Risk," (the latter, clearly, has utterly failed using their conventional models).

The entire article smacks to me of something dumbed down to the point of being useless and silly


Comments

Name

Email

Website

Speak your mind

4 Comments so far

  1. Arthur on August 24, 2011 9:43 pm

    Jim Simmons been doing it for years on consistent basis.

  2. Alec Misra on August 25, 2011 1:14 pm

    Jeff and Stefan, thanks for the links. One of the articles on the authors website inspired the following ideas for what they are worth:

    http://sicsemperliberalis.wordpress.com/2011/08/25/a-richter-scale-for-markets/

  3. Xox Ma on August 25, 2011 10:25 pm

    I read one of the papers published in May-2011 by the author in a generally respectable Physics Journal called Physics Review. It is called “Quantifying and Modeling Long-Range Cross-Correlations in Multiple Time Series with Applications to World Stock Indices”

    Authors try to apply some PCA and GARCH to find the “global equity index” and note that “Of the 48 indices, ten have a correlation smaller than 0.1 between the global factor, corresponding to the indices for Iceland, Malta, Nigeria, Kenya, Israel, Oman, Qatar, Pakistan, Sri Lanka, and Mongolia. Hence, unlike most countries, the economies of these ten countries are more independent of the world economy.”

    1. How can anybody in 2011 say that Iceland is independent of the world economy?
    2. Nigeria’s biggest industry is oil, but Shell is not listed there. Their second biggest industry is email spam - no listed spam companies either.

    etc etc… Mongolia!

    Being blind to the obvious facts of life seems to be the hallmark of distinguished professors… I don’t hold it against Prof. Stanley, but I also don’t want to spend any more of my time trying to find any value in his pontifications.

  4. douglas roberts dimick on August 26, 2011 9:44 pm

    X Marks The Spot

    That what may be a determinative point for understanding the underlying fallacies of both, multitime series a la vector autoregression (VAR) — as well as other such mathematical applications — and indexing (or batching) for purposes of cross-correlations of electronic exchange markets is found within a single query…

    How can you add something up that does not add up?

    Hence we track back to JW’s article as well as XM’s comment here. Also Phil currently approaches the issue from a slightly different perspective…

    http://www.dailyspeculations.com/wordpress/?p=6702

    As may be accomplished in legal analysis (or legal reasoning), one may approach such issues by reverting to questioning the definition itself, to wit:

    “All the variables in a VAR are treated symmetrically by including for each variable an equation explaining its evolution based on its own lags and the lags of all the other variables in the model. Based on this feature, Christopher Sims advocates the use of VAR models as a theory-free method to estimate economic relationships, thus being an alternative to the “incredible identification restrictions” in structural models[1].”

    http://en.wikipedia.org/wiki/Vector_autoregression

    Theory free?

    Are there not one or more underlying sets of theories being assumed when defining the symmetries for characterizing multitime-series to include the vectoring of autoregressions?

    Or does theory become anointed as being doctrine or law or doctrinal law if not physical law because a group of well financed academics find applications for merging intellectual disciplines (to include physics) into commerce?

    I am looking forward to a Sundance Film Festival submission when so pondering. See the trailer at…

    http://www.margincallmovie.com/

    During my limited and unremarkable experience within these worlds of money and math, I find the essence of this business, that business which appears to be labeled as “the markets” in terms of industry and profession (if one still cares to apply that term to what people do therein), is embodied within one axiomatic statement…

    “There are three ways to make a living in this business: be fist, be smarter, or cheat…”

    as so stated by the movie’s character played by Jeremy Irons (see trailer).

    Enough said.

    dr

Archives

Resources & Links

Search