Here is the truth: My trading is improving. The reason why it is improving is because I am learning to deal with my emotions better. However, I am not consistent yet. That, like most everything else in life, takes time. However, my reasoning for doing this is not to try and find some holy grail. It is just to understand the dynamics of markets and, if it does happen, discover something that I can apply as an edge. Why would this be stupid? If, after all is said and done, I am no worse off than when I started. However, I see nothing wrong with trying to apply computers to "assist" in trading (not replace the trader, but to assist in his efforts to obtain maximum profits and consistency).
first, any automatic parameter optimizer can do that, not just NN-s. second, the problem with NN-s, GA-s, etc. is that if you don't have a deep understanding, what you are doing, you will end up with useless curve-fitted systems. simply specifying "here's the T&S, profits are good, losses are bad, no go give me a system" to a NN would be like asking for the meaning of life from it -- it will be totally stumped. we, being programmers, started with very similar attitude, yet quickly found it to ba a dead end. - jaan
You want to understand the dynamics of the markets? Sit in front of a 1 minute and a 5 minute ES chart, without any ma's, indicators, s/r lines, BB's or anything and just watch how the price moves without having any opinion on why it is happening. Just watch and observe. Then, one day it just pops out in front of you why she does what she does. It works. But, the harder you try, the longer it takes. Learn price action. Learn to trade. Good Trading, H
I respect where you are coming from -- but let's face it, how many people are really making money trading ES? First of all, you aren't talking stocks -- you can't use the diversity of stocks that stock trading gives you. You are trading an INDEX, which means you are trading one of the most efficient markets in existence. Between the US bond market and the S&P index market, it is as efficient as it can get. So, in essence, my battle is with EMH -- since I know EMH is false, it is my obligation to find ways around the "almost efficient" market that is contained within the index movements.
Aphie... Aphie... Aphie... I can see where you're at. I've started Gann Analysis to understand the dynamics of the market, trying to "be" the holy grail. I don't care now. I've started programming stuff to find a matrice of the market and create a holy-grail system based on that. I don't care now. Look. The market has a mind of its own so that's why there's order in the market. The market itself doesn't have an opinion, thought and a bias, that's why it stays dynamic, not static. It doesn't favor anything so that's why it's fair and reasonable. Everything we try to quantify is just a glimpse and only a part of what the market has done. Remember, it's only a part of the whole past. All the analysis we make is just a bias of the trader, not the market's bias. Also, animals bases it's action on instinct. It can never understand nature. Humans bases their actions on both their instinct and their mind. We've got 1/2 of what it takes to understand nature. How about a computer? It doesn't have a mind or an instinct, it's only cluster of a human bias. Can you teach a computer to have a mind? Probably not, sometime in the future it maybe able to act like a person but it's not a person still, it's a cluster of human aspects. Blunt and never dynamic. Logical dynamics is not the dynamics of the nature. Look at Chaos Theory, it's got fractals and all the other stuff which logically explains the dynamics of nature but the actual nature's dynamics is more dynamic than logical dynamics. Chaos Theory in itself is a bias so it can't go close. I trade ES and NQ. I'm not in the most section.... LOL
Anyone who thinks my quest is stupid, just PM me a scanned copy of your most recent account statement trading ES. Thanks.
I hope you get over the quest soon and concentrate on your trader side. Remember, the market doesn't make you trade. You trade because of you. System(mechanical) or discretionary.
http://www.nytimes.com/2002/11/15/technology/15SUPE.html NEC Machine Again Leads Speed Ranks SAN FRANCISCO, Nov. 14 â The twice-a-year ranking of the world's fastest computers, to be announced Friday at a supercomputer industry conference in Baltimore, will include two new entrants based on personal computers, including one cluster of machines running the Linux operating system. The Top500 List, which is to be released by computer scientists based at the University of Tennessee and the University of Mannheim in Germany, will again be dominated by the Earth Simulator, made by the NEC Corporation of Japan. The Japanese machine shook up the computing world when its performance results were announced in April by Jack Dongarra, one of the computer scientists who, since 1993, have been tracking the speed of the world's fastest machines. American scientists see the powerful Earth Simulator as evidence that the United States has been spending disproportionately on computers that are primarily aimed at weapons research. Before the Japanese accomplishment this spring, the United States had dominated the top computing spot for many years. Despite the new Japanese dominance, seven of the world's fastest computers are installed in the United States. France and Britain also have machines that are among the top 10. In the new ranking, the ASCI Q supercomputer at Los Alamos National Laboratory in New Mexico occupies the No. 2 and 3 positions, because of its segmented design. Additionally, two supercomputers based on I.B.M.'s Power4 processor have joined the list. The ASCI Red supercomputer, based on the Intel microprocessor and housed at Sandia National Laboratory in New Mexico, which had been ranked the world's fastest seven times since 1997, has now dropped from the top 10. The Linux-based supercomputer is the NetworkX/Quadrics machine at Lawrence Livermore National Laboratories in California. It ranked fifth, and it is based on Intel microprocessors used in the personal computer world. Separately, Cray announced today a new supercomputer design, the X1, which the company said could potentially reach 52.4 teraflops, or a trillion mathematical calculations per second. The Japanese machine is ranked at 35.8 teraflops. Cray, which once dominated supercomputing, does not have a machine in the top 10. The full list will appear online at www.top500.org. In the current ranking, the United States kept its top position as a supercomputer user, with 46 percent of all systems installed, and as a producer, with 91 percent of all systems manufactured. Despite the performance of the Earth Simulator, Japan ranks third in installed supercomputing performance and in number of systems, behind Europe. To rank the machines, the scientists require computer sites to submit their performance running a measurement known as Linpack. Linpack is a collection of programs that analyze and solve a series of difficult mathematical problems.
Chaos theory is only minimally helpful in understanding the market because the market is only semi-chaotic. But chaos theory does help with other complex problems such as measuring the flow of water through a pipe, the rise of smoke plumes, waves crashing against the shore, or just the weather. But you are correct in that you do not need "logical dynamics" to succeed in beating the markets. All you need is an edge. The rest is probability and risk. Give Alphe some credit... he's trying to find that edge.