The perfect mechanical trading system. - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
1. there are people I know, even myself, who close their ears when they hear even a hint of FA. And (paradoxically), these guys are catching profits. So... I think you have to understand that a trader is the same system. And the trader who uses technical indicators is the same (un)artificial intellect that uses technical analysis indicators, whether macd or rsi. The very phrase "neural networks" should identify people and machines. It should be clear by the end of the year 2006. For example pensioners will never agree that machines can be smarter than humans, for man is the pinnacle of superiority, in the image of what each of them/us understands and represents differently.
2. fundamental, did someone at the top say? No problem! You can use economic indicators. Who is stopping them? It would be fashionable to build a macroeconomics model ..... of the world...
3. Personally, I have been in the AI field for 5 years. Of course I've been training grids. (Interested? 161130815). Giving errors. Strongly. But it seems to me that using a system on the NS, instead of trivial MA, something more attractive.
s.w. well this is solely mayo opinion. and + pardon the slang (today saw the rector's order for mayo enrollment in postgraduate studies, so glad)
Hello programmers and coders, philosophers and pragmatists :) I propose to develop the idea of creating a subj.
Personally, I am very interested in this topic! Ready to take part in every way!
About selftuning parameters: with basic everything is clear, but what to take as the basis for additional parameters, ie, what indicators, levels, channels, or what?
I had such an idea:
- put several indicators on the chart (e.g. RSI, Stoch, CCI, MACD, etc.) pick up "approximately" the values of these indicators;
- then, look at the history for approximate price reversals (i.e., where it is clearly visible "here to buy, here to sell");
- then write down the values of all indicators in these points, for buy and sell, in an array or in a file;
- further, in the Expert Advisor - check it (taking into account the deviation of indicator values in percentage from the ideal values), ie, for example, in the array, the RSI value for buy turned out to be 20, then, if the percentage of triggering is 10, then the buy will trigger from 18 to 22, as well as with all other indicators;
- Dale, you can (or should:) also add to the check crossing indicators of different levels or their signal lines;
I myself have not tested anything (although I started to write an experimental Expert Advisor, I have not got results yet), so I can't say anything will work or not.
In my opinion, this is a cardinally different way from the initial idea (by the way, the topic where arrows are attached to "good" places in charts has already been discussed (I don't know how it ended) where the Expert Advisor was calculating the best parameters using these arrows) ... In general, it was so complicated that it didn't lead to anything. And we start from the simplest. THE SIMPLEST ONE. We'll take any indicator, the simplest one, and use it. RSI and stochastics are too primitive :) I mean, the range of parameters is not great. But we can take the classic MACD as a basis.
But we are also programmed in our DNA. But we learn. That's why they are neural networks. In turn, artificial neural networks are not trained by seniors. And instead of binary logic, let's use fuzzy logic, like in our heads, shall we? You'll run into the Turing test later. Don't you dare believe in that nonsense. It's absurd to create a mechanism that's as dumb as a human being.
njel, and pensioners are right by the way. While the same neural networks are being programmed by us... I have just started to get into this business. My background is that of nuclear physicist, but now I'm more of an electronicist than programmer, so I'm learning it all :) What I strongly dislike about neural networks is that they are too slow to learn. Trends (not trends, but that elusive thing that makes previously successful systems fail) can be shorter than the time it takes a network to learn. This confuses me a great deal...I.e. A neural network, if it can be used, can only be used on sufficiently robust features. Here again I'm referring to the coded patterns in'Self Learning EXPERT'. I have a slightly different thought in relation to this. Encode not in binary, but in ternary balanced number system (computer Setun and Setut-70). If the price went up, add 1. If it went down, then -1. If the price did not change during the certain period of time, then 0. It seems to me that such a scheme provides very good information preprocessing for a neural network. As for the fundamental, I myself got interested in news not so long ago. I had an epiphany when I accidentally looked on H4 and higher charts (previously M5 was my habitat). How amazingly stable trends are there, they last for weeks! How many news are released in a couple of weeks! And still they don't influence the trend ... It seems the market only hears what it WANTS to hear. And trends are more its sentiments and expectations which are not so easy to dislodge.
Because trends on the daily and above are the results of stagnation and growth of the economy, which is known to happen in cycles and does not depend much on the current news.
eugenk1
But we're also programmed in DNA. But we learn. That's why they're neural networks. In turn, artificial neural networks are not trained by retirees. And instead of binary logic, let's use fuzzy logic, like in our heads, shall we? You'll run into the Turing test later. Don't you dare believe in that nonsense. It's absurd to create a mechanism that's as dumb as a human being.
I second that. The MACD is optimal.
Well, now we have to write the code :)