Has anyone attempted to mathematically explain the "noise" surrounding typically linear approaches to trading? I've just begun inquiry into these matters and am thus open to all opinions on this subject.
Q Noise Fluctuations in the market that can confuse one's interpretation of market direction. Noise may make you think it's the bellow bull market when it's actually the roar of a bear. --- http://www.taxopedia.com/terms/n/noise.asp UQ http://www.trading-ideas.com/TATutorials/tat061402-1.asp http://www.jimwyckoff.com/articles/traders/kaufman.html http://www.eunite.org/earchives/tech_reports/Fuzzy_Exception_Learning_to_Detect_Noise_Trading.htm
Q Nonlinear Ehlers Filters The signals you deal with every day often can be described statistically. For example, human speech has noise-like statistics. The speech process is nonstationary because it changes from moment to moment. Even though speech has noise-like characteristics, it obviously carries information. Price data resembles speech in statistical characteristics; it is both noise-like and nonstationary. One of the main problems you encounter in trading when using technical analysis is that you must attempt to restore signals that often are nonstationary and also corrupted by noise. When dealing with nonstationary signals with sharp transitions or when dealing with impulsive noise, linear filtering techniques give poor results. http://www.traders.com/Documentation/FEEDbk_docs/Archive/042001/Abstracts_new/Ehlers/Ehlers.html UQ
Hi, Here we go again. The Ehlers thingie is here again. A couple of months ago we had a lengthy discussion on the merits of the "Ehlers" products. Not a single traceable reference seems to be found in the mass of his "rocket science" publications on credible backtesting of these theories, except for some paid for certification. I don't know of a single poster who after programming Ehlers formulae came up with a positive backtest! We got a lot of talented people though on ET! I referred to this as "scientism", the hullabaloo application of rigorous mathematical concepts to fields where they do not apply. The techniques proposed by Ehlers are borrowed from systems and control theory and are successfully applied to many problems. Before you can do this though, many conditions have to be satisfied. As to markets, what is noise? This is an interesting question. Many high powered academics have entertained us many years about this topic. What do we know about this to go on by now? Let's first try to line up what (we think) we know already about this before going any further. If you compare this with an inertial guidance problem, engineers know darn well how to model the noise mathematically before they would think about applying the theories that Ehlers is trying to peddle for markets to the credulous. Keep on posting about "noise" in markets. I am trying to learn from your posts, but don't post nonsense. nononsense
No one except me ... and all the laboratories in the world making research in quantitative finance . This is the fundamental question formulated in the book "Martingales and Stock Market" from a french financial mathematical researcher : "Noise can make illusion. THE question is to know if <i>from the observation of an evolution one can deduce both the deterministic dynamic and noise amplitude.</i> (in italic by the author)". I say yes but you would be obliged to take the red pill see "true truths about Fibonacci" thread http://www.elitetrader.com/vb/showthread.php?s=&threadid=19259&perpage=6&pagenumber=9 that's why nearly none of researchers will ever make this discovery as long as they won't have the courage hee hee ! Not even much noise : precisions of forecasts of the new highs had only a few points error with perfect market timing.
Just thoughts: Two aspects of noise: Input data/information, and Output signals. Most (if not all) TA based traders, usually having limited risk capital, would be kind of noise traders; unless otherwise 100% FA based and without referring to any TA signals (which case is rare). Basically noise (false signals) could be inevitable/unaviodable, therefore most likely an 100% winning rate for all trades could not be attained. Whatever (how effective) approaches to improve a system's signal-to-noise ratio would naturally and marginally increase time lag to respond to overall (including legitimate) signals; it's simply a trade-off issue.
Whatever ? yes within a paradigm but not obligatory within ANOTHER PARADIGM. You don't believe it ? Let's take for example the so called Heisenberg Uncertainty Principle that the huge majority of people think that it is a SURE LAW whereas it cannot be as says the physician John Casti (in "Would be worlds" http://www.amazon.com/exec/obidos/s...2286954-1255355 John L. Casti member of the Faculty of the Santa Fe Institute which is one of the most famous center of financial research using chaos theory and artificial intelligence) "The theorems of Goedel, Turing, and Chaitin are limitations on our ability to know in the world of mathematics. The same limitation applies to statements such as the celebrated Heisenberg Uncertainty Principle in quantum theory, which at first glance appears to refer to an inherent limitation on our ability to measure certain quantities in the physical world. But a more careful examination shows that Heisenberg uncertainty is actually a limitation imposed by certain mathematical formulations of quantum theory, and may or may not be an intrinsic limitation in the structure of the real world itself." The paradigm used in financial scientific research today is not the good one so scientists can search and search for one million hundred years they will never find the true reality but only its shade since they keep to wear the same sunglasses ! It's not a question of being a genious but of being honest (in the sense intellectual integrity not influenced by current overwhelming paradigm - that's why at the time of Einstein's relativity discovery they didn't dare to give Nobel Prize for that but for his brownian motion theory) in front of facts especially those that create financial paradoxes. Efficiency Principle is the counterpart in Finance than Heisenberg Uncertainty Principle in Physics.