harrytrader wrote: Current academics' stochastic process models are non-EMH (and subsume Technical Analysis). Just browse Quantitative Finance for example, or how technical and fundamental traders effect the stochastic process of the market - http://www.santafe.edu/~jdf/papers/jebo.pdf Are any of these models correct or even good for anything? One thing I am convinced of: the most complex stochastic models (including HMMs) are developed deep in the bowels of hedge fund firms and are not published, for the obvious reason.
Hi again Harry, so basically you say "traditional quants" or whatever you might call them cannot correctly judge about "efficient vs. inefficient" because the underlying thesis/assumption about market behaviour & pdfs is wrong? Then let me pose the question: Can there be a "correct" market model? If such a one would exist, wouldn't the market soon become "dead"? Ok, obviously it woulde depend on how many market participants would discover it & thrive from it.... If I look at your posts & model I just wonder why the market should follow a deterministic path, are pdfs stable enough to allow for this? I think not, but maybe you found a way to switch analysis underlying pdfs accordingly & in advance...that's the only way to profit from a model - know when it is a square wave & predict using cycles or just following the trend... best & tia for any comments to the above! cosmic
The most representative of Santa Fe is J. Doine I like him very much he is very rigorous guy not a cheater but he is far from predicting stock market, he is just dectecting small anomalies but there's nothing astonishing that there are small anomalies since Normal Law is not the norm even in process industries see "Normal distributions are not the norm." http://www.elitetrader.com/vb/showthread.php?s=&threadid=25943&highlight=normal+law So small anomalies are not questioning EMH because if there weren't any small anomalies there wouldn't be no arbitrage and arbitrage is mechanism given today by EMH propagandist to explain how efficiency can operate in concrete : what EMH means is that these small anomalies can of course appear but would then disappear rapidly and that is rather true in practice. Since small anomalies can't contradict EMH, it's not worth focusing on them so I would rather focus on what is called "long term memory effect" (non-linear dynamics theory) which can have two possible causes: either stochastic, either deterministic - I will come back to that later. More complex models (using computer's simulation using or not artificial agents with or without neural network) are rather toys models : they can simulate an artificial market to explain some phenomenas that ressemble the true stock market (for example herd behavior) but they are only surrogate of the true market: they are far from being able to predict the real market. It's like building a robot with artificial intelligence to try to explain a human behavior but it won't be able to predict the behavior of a human in particular only human in general. I have already recommended in the past to read "Would Be World" from John Casti - who belongs also to Santa Fe - which gives an introduction to non-linear dynamics (I prefer this term to the fashionable but misnamed expression "chaos theory") He also introduce the basics of different kind of modelling: predictive (Newton's law for example), explanatory (qualitative model) and prescriptive (interaction of factors). His book contains a stock market simulation: it clearly exhibits the El Farol problem: there are so many participants and strategies that it seems impossible to build any predictive model. That's why many scientific researchers just gave up the possibility of the existence of any determinism in stock market and consider that all TA that is founded on such determinism like Dow Theory, Elliott and Gann are only pure FADs. I say that they are wrong because their premisces that the agents drive the market is false. The first to say that is an Economic Nobel Prize Maurice Allais - affiliated to the same school than Milton Friedman - but he never developped a model for that.
OK enough for EMH as for me, next I will talk about Dow's (or rather Hamilton who transcript Dow's theory since Dow himself has never written about his theory) SEMI TRUE LIES .
Yes I call this maximum inefficiency Potential <IMG SRC=http://www.elitetrader.com/vb/attachment.php?s=&postid=410292>
Examples on last days: Optimal Potential - or Max Inefficiency if you prefer - was forecasted at 10368.8 on 12/01 in real on 13/01 it made 10368 Optimal Potential - or Max Inefficiency if you prefer - was forecasted at 10594.6 on 14/01 in real on 15/01 it made 10592.44 (for more precision see illustration here http://www.elitetrader.com/vb/attachment.php?s=&postid=410039 on with comment here http://www.elitetrader.com/vb/showthread.php?s=&threadid=27118&perpage=6&pagenumber=2) <IMG SRC=http://www.elitetrader.com/vb/attachment.php?s=&postid=410293>
See above: everyday the appearance of market changes that is to say it is impossible to find a function that fits this appearance in advance. You can curve-fit each PAST day but you cannot fit (predict) FUTURE day if you only take appearance into account. So it is not by trying to curve-fit this MOVING APPEARANCE that you will be able to predict the market. What doesn't change is the UNDERLYING STRUCTURE: for that it is possible to find a set of functions or equations to model it (except if the market is really driven by the crowd due to El Faroll problem you would give up straight away because it is just then impossible. But as I said above this premisce adopted by the overwhelming majority has been taken as false in my original hypothesis). But this structure is INVISIBLE you have to build a THEORY to SEE IT and then confront it with REALITY see Poincarré on "Science and Hypothesis" (http://spartan.ac.brocku.ca/~lward/Poincare/Poincare_1905_01.html summarised here http://www.utm.edu/research/iep/p/poincare.htm: Experience suggests scientific theories; but experience does not justify them. Experience alone is unable to falsify a theory, for the theory often corrects the experience. Experience is judged according to a theory. A central aim of science is prediction. ) or remember Plato and the Cave Allegory : Skepticism, or the Problem of Knowledge [Plato's allegory of the Cave] "Imagine the condition of men living in a sort of cavernous chamber underground, with an entrance open to the light and a long passage all down the cave. Here they have been from childhood, chained by the leg and also by the neck, so that they cannot move and can see only what is in front of them...At some distance higher up is the light of a fire burning behind them; and between the prisoners and the fire is a...parapet...like the screen at a puppet show, which hides the performers while they show their puppets over the top...Now behind this parapet imagine persons carrying along various artificial objects, including figures of men and animals...which project above the parapet...Prisoners so confined would have seen nothing of themselves or of one another, except the shadows thrown by the firelight on the wall of the cave facing them...And they would have seen as little of the objects carried past...Now if they could talk to one another, would they not suppose that their words referred only to those pasing shadows that they saw?...In every way, then, such prisoners would recognize as reality nothing but the shadows of those artificial objects." http://www.island.net/~dbruiger/doc_html/6.html - English
The weird thing about markets is that you can mathematically frame them but forecasting them is a different matter due to the variability present within. Look at economic models, they obvioulsy work, but no one (apart from Gordon if we are to believe a politician) can accurately predict them.
"..the market's price is equal to the fundamental price." Absolutely not. The conjunction of supply and demand at any one moment is too hugely hugely imperfect to give a so-called fundamental price at that moment .. thats got to be self evident to any economist or indeed to anybody If on the other hand one said that a general and current price level or a price range over a currently running period was the fundamental price then yeah ok one could go with that because as well its harder then to suggest an alternative.. although you could still come up with a different theory.