Zigzag indicator and neural nets - page 8
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
SK - the purpose of my testing was to check the indicator in dynamic mode, but not to try to get the maximum profit using it, so I did not try to improve the Expert Advisor. The result of checking the dynamic characteristics of the indicator is very good, from my point of view, it is mediocre in terms of using it together with the Expert Advisor, the drawdown is too large, I would not use such an Expert Advisor for real trading.
But I am absolutely sure in one thing, this indicator will work equally in any market, no matter what timeframe I would check. It follows from the principle of its construction and is also confirmed by my numerous observations of its work during the year on a demo account.
I don't want to spend resources on downloading large quotes history for the sake of checking something that I do not doubt, moreover I am not planning to use this indicator separately, but together with an expert system that will make forecasts - the results will be incommensurable with what I have now.
If i'm not sure how to use it, i'll try to build a simulator with first class indicators and then compile the indicators with MQL5.
If i've got some experience in writing indicators and Expert Advisors in MQL, but i haven't got any further than NSDT from Ward Group and realizing them in MT4 is very difficult (for me, maybe i'll find out more).
available here: http://neuroforex.jino-net.ru/index.php?option=com_remository&Itemid=28&func=select&id=1
SK - the purpose of my testing was to check the indicator in dynamic mode, but not to try to get the maximum profit using it, so I did not try to improve the Expert Advisor. The result of checking the dynamic characteristics of the indicator is very good, from my point of view, it is mediocre in terms of using it together with the Expert Advisor, the drawdown is too large, I would not use such an Expert Advisor for real trading.
But I am absolutely sure in one thing, this indicator will work equally in any market, no matter what timeframe I would check. It follows from the principle of its construction and is also confirmed by my numerous observations of its work during the year on a demo account.
I do not want to spend resources on downloading a large history of quotes for the sake of checking something that I do not doubt, moreover I do not plan to use this indicator separately, but together with an expert system that will make forecasts - the results will be incommensurable with what I have now.
I expect to get a profit factor of not less than 25, and as for the benchmarks - there is no limit to perfection, and with the growth of computer power, and I am constantly faced with a lack of resources, it is possible to create more and more efficient systems. There are many ideas, not enough time, and often not enough programming experience to implement them all.
Regarding evaluation criteria, I believe the most important, along with the accuracy of predictions, is the stability of the system, it must work the same way now and in 10 years regardless of changes in the market, and this is achieved by the ability of the system to retrain itself in real time and to adapt quickly to the changing situation. Now my system retrains every 5 minutes and recalculates forecasts every minute with new bar, if I had bigger random access memory and speed I would do retraining at every step along with calculations and forecast accuracy would increase a lot, but now I take only small part of informative inputs that I use for neural network training, otherwise computer gets stuck because of memory shortage. But these are purely technical problems, I hope they will be solved with time.
For the sake of interest I tried in the Strategy Tester again the Expert Advisor with the indicator which I tested above. However I disabled trailing stops and shifted them, they only interfered (though I could leave the stop open as it never triggered anyway), the result was somewhat better.
If i'm not sure how to use it i can say that the indicator is already compiled and i've already learned how to connect it to mt4, i've already written it on a shell.
If i've got some experience in writing indicators and Expert Advisors in MQL, but i haven't got any further than NSDT from Ward Group and realizing them in MT4 is very difficult (for me, maybe i'll find out more).
Re-training every 5 minutes and recalculating the forecasts every minute - isn't that too often? And your desire to further increase the frequency of retraining (and calculations) to improve prediction accuracy (on every tick, or what?) seems strange to me. I doubt that a really working system would benefit from retraining at a frequency that coincides with the incoming data frequency.
P.S. And pf>25 is not just a dream, but something beyond belief... Although with the ratio of profitable trades to unprofitable ones 5:1 and TP/SL = 5 it is quite achievable.
P.S. And pf>25 is not just a dream, but something beyond belief... Although with a ratio of profitable trades to unprofitable ones of 5:1 and TP/SL = 5 it is quite achievable.
You think that PF = 25 is impossible? And what value of PF seems acceptable to you? And what other values are necessary in your opinion? For example, what do you think the allowable drawdown should be?
SK. I consider a pf of no less than 3 to be acceptable. Permissible drawdown - not more than 5-6% in one trade and not more than 15-20% in a series of losing trades. There are also other parameters - Sharpe ratio (not less than 3-4), p.o. of a deal (depends on the working TF), evenness of distribution of deals by time and profit.
A stable TS with pf>5 is expensive, to say nothing of 25... I'm not saying that such is impossible, but developing such a TS would mean that the author has caught Foreh by the tail.
A stable TS with pf>5 is expensive; let alone 25... I'm not saying that one is impossible, but developing such a TS would mean that the author has caught Foreh by the tail.
Agreed.
The cheapest gold is an alloy with a gold content of 0.375 to 0.875 (mass production).
Alloys with a gold content above 0.99 are somewhat more expensive.
The production of alloys with a gold content of more than 0.999 requires special laboratory conditions (special furnaces, special furnace linings, etc.).
The production costs of material with a gold content of more than four nines far exceed the market value of the gold itself.
Obtaining purer gold is to a large extent limited by the possibilities of modern science and technology (requires special conditions: deep vacuum, plasma, etc.) and is only within the reach of the national programme.
I think PF = 5 is more than enough. That approximates 0.84 profitable trades out of the total number of trades.
Obtaining PF = 25 - is a task, the solution of which will require considerable and costly effort comparable to 5-7 nines in gold:). If it were possible, then the application of such a solution to Forex would mean childish pranks and squandering of the national wealth. And this general solution should be applied to weather forecasting - it would bring an undoubted benefit to mankind and immeasurable profits to the author of the solution.
I think another indicator is significant, we can call it Regularity. This is the amount of profit made per unit of time, at a constant order value (e.g. $1000). This parameter is measured as [ $/day ].
One thing is if the trading system shows P = 10, i.e. an average of $10 in profit per day (over a period of time, for example during the year) at a cost of each order of $1000. It is another matter if P = 3. This figure is important, because all other things being equal (for example, FF = 5, 10% drawdown), the system with different P will give different results.
It is clear that not all of TCs work at a constant value of orders. For such TS, we can introduce an indicator calculated for the changing order values. This indicator (Variable or Percentage Regularity) can be calculated similarly: OR = P/S/S, where P is profit, S is order price, and Q is time. The unit of measure for this indicator is [ %/day ].
On the basis of Regularity, we can calculate the trading system stability. To do so, we should analyze how P changes during the period being tested. For example, if P varies from -5 to 15, then the system has low stability. If the range of P variation is narrow (e.g. from 7 to 12), such TS is more stable.