AI will get better/smarter. it just needs more data, more "experience". people make the mistake of considering today's "boxes" and thinking that they are static - they are not. Signed, Hal
and just wait till you see the fireworks when they break loose of those artifical constraints on human nature. I saw this coming when I designed my first system in 1992.
Eckhardt, poor chap, doesn't understand the scope of TA. A scapler looks for turning points in price action - that's all. Doesn't matter if it's a robot, mass psychology or a news driven event that turns price. His grasp of maths is do doubt fantastic, but I suspect it comes at the expense of a inferior grasp of the mechanics involved in TA at it's best.
It is quite simple, there will always be pull backs, bounes and break outs. We just exploit them, does not matter who drives them.
This random walk stuff is nonsense. OK lets see what happens if I begin trading General Electric stock 2 January 1962 and stop trading 18 July 2006. This mechanical system has these three rules: 1) Buy when price first increases above the highest price of the prior 400 sessions. 2) Sell when price first decreases below the lowest price of the prior 400 sessions. 3) Size positions so that the number of dollars to risk stop is 10 percent of account equity. Number of trades 5 Total profit $ 2944958 Profit after subtracting $ 10.00 commission & slippage per transaction: $ 2944858 Risk is 10.00 per cent of equity. Drawdown is 0.0211 (2.11 per cent). Cumulative Annual Growth Rate (CAGR) is 66.28 per cent. CAGR / Drawdown is 31.48 Instanteously Compounding Annual Growth Rate (ICAGR) is 7.69 per cent. Annually Compounding Annual Growth Rate (ACAGR) is 7.99 per cent. Information Ratio is 0.52 Initial capital is $ 100000 Long trades only. Growth rates are calculated after subtracting commission & slippage. ===== About 66 percent growth each year and the greatest drawdown is about 2 percent.
Why would you trade purely technical with such long term variables? 400 period MA, that's one and a half year rite? I assume your time frame is daily, have you check the system is making $ in the resent years?
I just chose 400 days for no particular reason. Other values might show "better" performance. I trade to make money. I do not trade to make myself feel productive, or useful, or to feel that I am a productive member of society. Sometimes long term trading methods are the best choice for me. Your preference may differ. One thing I noticed during my long lifetime is there is a big difference between productive work and appearing to be productive by running around a lot.
The entry really isn't about random walk. The entry has to do with the saturation of programmed trading. Your post does show a historically profitable system. However there is nothing to say that the system will remain profitable. That's random walk theory.
A very successful NYMEX trader Mark Fisher (owner of MBF Clearing) made millions over a simple observation which he later developed into a solid strategy. The observation is that market open is the high or low of a day 20% of time. His strategy is outlined in his book âThe logical Traderâ. Market are not that random, but it takes while to realize it, and much longer and harder to consistently profit from it. redduke
The first main statement in Marks book is based on a mistake in understanding of the behaviour of random walks and is incorrect. Market behaviour is much closer to random than most traders appreciate and is getting closer every year.