Algorithmic ''centrifuge'' - 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
This is a genetic optimization algorithm. Only it doesn't usually parse which block which parameter belongs to which block.
ps: everything you can think of has already been invented a long time ago.
ps2: the centrifuge has its rightful place next to the kernel and the engine.
Optimisation of the resulting something, the selection of effective parameters, will be decisive. When, at what point, there will be an intermediate mini-optimization, a local selection of parameters of an indicator or a bunch of indicators. Some block.
For example, who will decide which parameters to leave for the new element of the trading system? Many systems are based on one brush. On the intersection with the rate. MA(SMA,9.0), MA(SMA,104.37), MA(SMA,28.35), MA(SMA,14.0). Trading strategy is based on intersection of a rate with MA(SMA), but trading systems and their mini-ideology are a little different. We changed the period and the shift of the MA and decided on these parameters. Of course, I may be wrong in saying that these parameters should be exactly those.
Besides, the "writing" of the currency pair changes from one instrument to another. Somewhere 28.35 will do, but not on the other instrument. And if you change the timeframe of one symbol, the parameters will be different. Neither 28.35, nor 104.37 are good enough. The session has changed, the volatility has risen, and we need more parameters.
Now the volatility changes - the minimum stop loss changes automatically. This is another block. Trading. From the selected risk the maximum possible safe lot is changed. I have a material about it on my website. Working lot, less than the maximum, when and by what criteria the working lot will be selected. By whom or by what. Optimal, effective parameters are constantly jumping and changing.
Trading systems on Stochastic 5,3,3 on one instrument on M1 and on D1 are different. It is clear that to buy cheaper and sell more expensive. But the approaches in the real successful trading are different. The theory says: there are waves and there are waves. Whatever. Stochastic can be applied. And practice. As you start to really trade. Everything is different on the same stochastic 5,3,3. You may not see a positive result there and there, but for completely different reasons. In theory and on history everything is beautiful.
Perhaps I was optimizing-optimizing, selecting local effective parameters and then selecting and selecting the effective parameters common to the trading strategy design. While selecting them, the parameters disappeared. The news and its consequences. The strong trend on D1 has changed to flat - pro-trading. Rates have changed. Currency intervention began. Or just sailed away. A strategy without the selection of positively effective parameters is nothing. I have just made the strategy blocks, we should assess their effectiveness by fact or by history, at least.
One ideology is making profit on the flat. If we want to make profit on the trend - that is another one.
This is the third or which one. It is necessary to create a database of strategies, grid strategies and their variants, martin - and their variants. And so on and so forth. The database will show which indicators are used the same way and for what purposes and in what positions. Here already blocks will be visible.
In short, the one who will compile a database of strategies with a breakdown of used elements of those and fundus of analysis and criteria of their use - he is a genius. Can get it out of the drawer. :)
I don't think the idea is fantastic. It all boils down to the optimisation that everyone loves here, only more complicated.
It is not only the values of the system parameters that need to be selected, but also the system parameters themselves. Indicators are taken as a sample of the system parameters. Otherwise everything is the same.
1. First, we select the system parameters (assembling of indicators).
2. We select values for the entry points (for the selected indicators assembly).
3. Selecting values for the order parameters - stops and lot.
We obtain the output trading strategy.
You can make it even simpler to go through SEEDs for MathRandom:-)
It's about the same as taking arbitrary indicators from a pack and optimising their parameters and relations.
Peter - next year trade with some cents and minimum lot at least. There will be no such bad ideas.
---
There is another tip: use trend indices on 3 different timeframes, volatile indices on different timeframes and 1 zigzag. This is supposed to be a system
Carefully optimised to make the comp work and have visibility of the process. And most importantly - martingale them afterwards (who doesn't care if it's open that way) :-)
You can make it even simpler to go through SEEDs for MathRandom :-)
It's almost the same as taking any indicators from a pack and optimizing their parameters and relations.
Peter, next year trade with some cents and with a minimum lot. There will be no such outright stupid ideas.
---
There is another tip: use trend, volatile indices of different timeframes and 1 zigzag on 3 different timeframes. It's supposed to be a system
To optimise carefully, so that the computer works and the process is visible. And most importantly - martingale them afterwards (who doesn't give a damn if it is open in the right direction) :-).
The genetic algorithm starts like that, and then it goes through the combinations of groups of parameters.
...
The concept is that EACH indicator is ONE PARAMETER.
All indicators are the OVERALL PARAMETERS OF THE TRADING SYSTEM.
All indicator calculations are reduced to one parameter, which is used as a part of common input/output signal.
Several indicators constitute ONE SIGNAL.
Configuration of indicator values selected under the Signal is searched by a genetic algorithm.
The configuration of the order parameters values is also searched using the optimization method.
What is a Parameter?)
What is a Value?
What is System?
What is Optimisation?
Do you know?))
Now think of an Indicator as a formula for calculation of ONE Parameter that is part of a Trading Signal.
Several Indicators - several parameters - ONE Signal.
The entire trading system is a complex of parameters: Input and Output Signal parameters, Order and Stop parameters.
Trying parameters of the System and their values in the process of optimization, we obtain the Strategy.
The concept is that EACH indicator is ONE PARAMETER.
All indicators are a GENERAL SELECTION OF TRADING SYSTEM PARAMETERS.
All indicator calculations are reduced to one parameter, which is used as a part of common input/output signal.
Several indicators constitute ONE SIGNAL.
Configuration of indicator values selected under the Signal is searched by a genetic algorithm.
The configuration of the values of the order parameters is also searched using the optimization method.
Why? It has already been done. If the stochastic is so-and-so and so-and-so, then the Boolean variable = True. If the SAR indicator is so-and-so in such and such conditions - another Boolean = Troue. etc. 5-6-8 such variables. At the end of the day, it turns out that the Expert Advisor has become very quiet. It does not make deals. Or makes one trade of 10 pips per week. There is no convergence of such logical variables, or it happens very rarely. And trends and flotsam go by. Martingale is already making money. But we are afraid to use it. The reality is like this. It is being done already.
About optimization. Any optimization is carried out using a genetic algorithm. It's not the genetic algorithm itself that searches for something. Optimization is performed using the genetic algorithm. Which gives a time gain in optimization.
...
About optimization. Any optimisation is more appropriately done using a genetic algorithm. It's not the genetic algorithm itself that searches for something. Optimization is performed using genetic algorithm. Which gives a time gain in optimization.
If I understand GA correctly, it narrows down the search area for values during Optimisation.
For example:
There are Parameters A,B,C. Their range of possible values is 4.5 billion.
There is Parameter X, which varies from values of Parameters A,B,C. However, a pattern of changes is not revealed.
Task: bring Parameter X to value Y by enumerating values of A,B,C.
Two variants: (1) direct brute force and (2) genetic algorithm.
The second variant effectively narrows the field of search for the required values.