New historical Forex models | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Markov Models and Forex 33 replies

Anyone used ARMA + GARCH models in forex trading? 14 replies

Historical chart to track historical news events. 0 replies

Risk Models 0 replies

MT4 VWB Models 1-4 8 replies

Attachments: New historical Forex models
Exit Attachments

New historical Forex models

  • #41
  • Mar 11, 2016 7:11am Mar 11, 2016 7:11am
Quoting Matt098
Thanks to everyone for supporting the topic. If I understand correctly, all of you using indicators in real trading. FXEZ was able to raise the indicator performance to 67% as he said. Good result, if so. How far each of you reached by indicators? How deeply you analyzed the past? There is an opinion that more or less successful analysis accounted for at least 5 years of M15. But at the same time that the market is fickle, and doesn't make sense to analyze the history more than year. What do you think?
Personally I think it makes sense to take more than a year, still a year too small part & so many fundamental things have happened over the last year for example.

Thread is pretty interesting, I would say at least every trader once thinks about, but I'd like some specifics.
You just do research or you are already traded on found coincidences?
What were the results of trade? Thx
  • #42
  • Mar 13, 2016 10:05am Mar 13, 2016 10:05am
Quoting Matt098
{quote} U did a good job. Why 40 bars? This number is the method of selection?
The length of the sequence and the number of nearest neighbors are free parameters. The sequence shall not be too short to get something which might have an interest. Too many neighbors make the prediction neutral (converges to the marginal expectance of zero return). Too few makes the variance too high.
Quoting Matt098
How far each of you reached by indicators? How deeply you analyzed the past?
I went nowhere with anything that tries to predict the future or a turning point. The small edge I can find in any indicator is the trend. For instance I tried to improve the RSI by detrending the price first. Then I tested all the possible signals I could imagine (a top above the 50, zero-crossing, cross 20 upward...). As soon as you organize the results depending on the trend you realize that the edge comes from the trend only. I never found any indi making money by shorting in a uptrend. Even when a signal has an average return of zero it is skewed in the direction of the trend.
Quoting Matt098
There is an opinion that more or less successful analysis accounted for at least 5 years of M15. But at the same time that the market is fickle, and doesn't make sense to analyze the history more than year. What do you think?
To me what is important is to test in various regimes. Long term trend, choppy market, range, oscillating up trend, crash, bubble... You can't find all the cases in just a year of data. 10 years of the same price action is useless as well.
No greed. No fear. Just maths.
  • #43
  • Mar 14, 2016 7:56am Mar 14, 2016 7:56am
Quoting Kite
{quote} Personally I think it makes sense to take more than a year, still a year too small part & so many fundamental things have happened over the last year for example. Thread is pretty interesting, I would say at least every trader once thinks about, but I'd like some specifics. You just do research or you are already traded on found coincidences? What were the results of trade? Thx
I have my own account trading on.
Quite good results. But it has nothing to do with this topic. If interested, we can discuss it. But as long as the thread is not goes to the discussion of other line, I still would like to say the following. I always wanted to learn as much as possible deeply a theory of historical reiteration and coincidence. Through analysis I aim to achieve 100% result.
I'm looking for quality ideas of people like me, who believe the market is not just a chaotic movement. Who say that history repeats. And who may have already made some steps in determining a reiteration in history in order to predict future movements.

In the near future I will publish new examples below. Thanks for your feedback. Success in trading
  • #44
  • Mar 15, 2016 4:42am Mar 15, 2016 4:42am
Quoting Matt098
{quote} In the near future I will publish new examples below. Thanks for your feedback. Success in trading
Matt098, hi. Are you realy earn on forex?
I have impression that you just fantasize about your theory...
  • #45
  • Mar 15, 2016 8:56am Mar 15, 2016 8:56am
Quoting AlexFF
{quote} Are you realy earn on forex? I have impression that you just fantasize about your theory...
My theory gives me really good result. I earn in part by the fact that find my analysts, but above all I have my own trading system.
  • #46
  • Mar 18, 2016 6:25am Mar 18, 2016 6:25am
Quoting Matt098
{quote} I have my own account trading on. Quite good results. But it has nothing to do with this topic. If interested, we can discuss it. But as long as the thread is not goes to the discussion of other line, I still would like to say the following. I always wanted to learn as much as possible deeply a theory of historical reiteration and coincidence. Through analysis I aim to achieve 100% result. I'm looking for quality ideas of people like me, who believe the market is not just a chaotic movement. Who say that history repeats. And who may have...
Achieve 100% result is not possible, i think.
Reiteration may be accidental, that won't give any future results, what do you think?
Many experienced traders will agree with me probably. It's smart to develop your own theories, rather than follow existing.
But personally, I'd have dealt with it in spare time from trade
  • #47
  • Mar 23, 2016 3:31am Mar 23, 2016 3:31am
Quoting Matt098
{quote} Thanks for your opinion. Market is a reflection of the psychological crowd actions. If to find match patterns in the past, we can use them in future. Patterns are not always 100% visually similar, admit the possibility of deviation up to 10%. Number of candles in pattern may also differ by 10%. Further if we compare them in the future and the past, they may be worked out 100%. Other words is possible to guess entry and exit almost to a point. So we can take out the whole movement, but sometimes happens pattern works on half or takes place...
Hi, Matt 098.
It is important to consider how much time working out at the pattern of history.
The use candlestick patterns to trade should be a condition of their frequent recurrence. If the history of 10-15 years a pattern has been perfected just 5-20 times, it's just a coincidence.
Regularly there are robots that are encoded in the patterns of recurrence. Their main mistake - rare patterns that the authors are trying to optimize the profit in the tester. And to increase the total number of deals - to increase the number of such "rare" patterns. As a result of such robots are profitable 1-3 months after optimization, and then very quickly lose all the money.
  • #48
  • Mar 23, 2016 6:38am Mar 23, 2016 6:38am
Quoting Sergey5
{quote} Regularly there are robots that are encoded in the patterns of recurrence. Their main mistake - rare patterns that the authors are trying to optimize the profit in the tester. And to increase the total number of deals - to increase the number of such "rare" patterns. As a result of such robots are profitable 1-3 months after optimization, and then very quickly lose all the money.
I agree with you. Do you think it possible to optimize robots so that they made fewer errors, or weren't give too big errors?
I like this topic. Indeed, whole forex built on recurrence, all market theories built on recurrence. For example Elliott waves, they too are calculated on the prediction based on fact that certain wave will necessarily go in the right direction.
The problem is that no one will talk about the entry and exit points so that eventually all the theories are designed for the analysis of history.
  • #49
  • Mar 23, 2016 7:57am Mar 23, 2016 7:57am
Quoting Grrreen
{quote} I agree with you. Do you think it possible to optimize robots so that they made fewer errors, or weren't give too big errors?
That's a very difficult question. It must be very clearly understood that we optimize it. After optimization of the tests should be done with a deviation from the specified settings and view the results. To avoid strong deviations. It is necessary to take a closer look at the graph of balance, so that it was stable. And yet a lot of different things.

Repeatability of events on the chart should not be taken as dogma. For example, the last 2-3 years stopped working model absorption candle on the daily charts. Because all traders see that the candle is formed such. And beginning to go to this position. As a result, the price is changing faster and stronger than normal, and completely selects the full potential of the movement in the first candle.
The reasons are many details to talk about them for a long time.

But as you can see, and the formation of new patterns that did not exist before. In the classic technical analysis.
But this is a big and long topic of discussion ...
  • #50
  • Mar 24, 2016 9:34am Mar 24, 2016 9:34am
Quoting Sergey5
{quote} But this is a big and long topic of discussion ...
Sure!
I am confused by the fact everyone knows everything, knows how it works or doesn't work.
But still trying to develop and looking for (I mean the author).
We must of course pay tribute to the desire to find patterns still, but in my opinion their searching takes too much. The market is changing faster ...
  • #51
  • Mar 24, 2016 11:27am Mar 24, 2016 11:27am
Quoting Grrreen
{quote} Sure! I am confused by the fact everyone knows everything, knows how it works or doesn't work. But still trying to develop and looking for (I mean the author). We must of course pay tribute to the desire to find patterns still, but in my opinion their searching takes too much. The market is changing faster ...
But what if find the Holy Grail?
But seriously, there are still some patterns, looking at that most traders take positions and starts a lot of pressure on the price.
For example, the breaking of the strong trend lines.
Here is my current deal, which I opened a month ago at breaking through of a trend. Screen attachment.
Now the price in TP area.
Attached Image (click to enlarge)
Click to Enlarge Name: GBPCADWeekly.png Size: 32 KB
  • #52
  • Mar 25, 2016 2:46am Mar 25, 2016 2:46am
Quoting Matt098
{quote} My theory gives me really good result. I earn in part by the fact that find my analysts, but above all I have my own trading system.
Well, you do realize these are only words, where is the proof? We need real arguments, not just a beautiful story about how it all works
  • #53
  • Mar 25, 2016 3:05am Mar 25, 2016 3:05am
Quoting Sergey5
{quote} For example, the breaking of the strong trend lines. Here is my current deal, which I opened a month ago at breaking through of a trend. Screen attachment. Now the price in TP area. {image}
Correct deal at the right time

Quote
Well, you do realize these are only words, where is the proof? We need real arguments, not just a beautiful story about how it all works.

You're a little rude I think. Matt098 develops excellent theory. If it gives results, this will give all of us hope to think about the market in a different way.

But to be honest in some way you're right. I also wanted to ask the author about the results? You haven't post new charts long. You're talking about your trading system, in addition to development?
  • #54
  • Mar 25, 2016 3:14am Mar 25, 2016 3:14am
Quoting Kite
{quote} Well, you do realize these are only words, where is the proof? We need real arguments, not just a beautiful story about how it all works
I believe him, the theory is very interesting, I'd advise you to reinforce all of it by alligator. Alligator very correctly shows the trend. I'd like you to continue your experiments and introduce to us. Thanks
  • #55
  • Mar 25, 2016 6:56am Mar 25, 2016 6:56am
Quoting LawKin
{quote} I believe him, the theory is very interesting, I'd advise you to reinforce all of it by alligator. Alligator very correctly shows the trend. I'd like you to continue your experiments and introduce to us. Thanks
Thanks for support. As for alligator I've been thinking about it. There can be many variations of using this indicator. All of them will be gradually included in the development.
  • #56
  • Mar 25, 2016 7:20am Mar 25, 2016 7:20am
Quoting Sergey5
{quote} Hi, Matt 098. It is important to consider how much time working out at the pattern of history. The use candlestick patterns to trade should be a condition of their frequent recurrence. If the history of 10-15 years a pattern has been perfected just 5-20 times, it's just a coincidence. Regularly there are robots that are encoded in the patterns of recurrence. Their main mistake - rare patterns that the authors are trying to optimize the profit in the tester. And to increase the total number of deals - to increase the number of such "rare"...
Hi Sergey5
We're here trying to figure out a relatively complicated models of the analysis that could occur in the past and how they move in the future. And most importantly how to eliminate unnecessary coincidence. Advisors optimization is generally a separate huge topic. I know many successful programmers write their testers because standart MT4 tester conceals a lot of surprises, a normal user has no idea about.
  • #57
  • Mar 30, 2016 5:11am Mar 30, 2016 5:11am
Here write about the alligator, but the alligator is more about trend indicator. Perhaps you need to separate the market on trend and correction. To correction bollinger more appropriate. This is the good indicator. Mixed with the alligator is powerful thing.
  • #58
  • Mar 30, 2016 5:55am Mar 30, 2016 5:55am
Quoting SantaNunez
Here write about the alligator, but the alligator is more about trend indicator. Perhaps you need to separate the market on trend and correction. To correction bollinger more appropriate. This is the good indicator. Mixed with the alligator is powerful thing.
Thank you for notes. Really Bollinger works well in horizontal movements of the market. We will work in this direction.
  • #59
  • Mar 30, 2016 6:27am Mar 30, 2016 6:27am
I think you ignored my question. When you show the results of your trading. How much do you earn? $100-200 per month?
  • #60
  • Mar 30, 2016 8:18am Mar 30, 2016 8:18am
Quoting Kite
I think you ignored my question. When you show the results of your trading. How much do you earn? $100-200 per month?
Specially for you made the screenshot. These are the results for last 10 orders. This profit is partial. My and my partners.
Attached Image (click to enlarge)
Click to Enlarge Name: screen.jpg Size: 404 KB
0 traders viewing now