The similarity system | Page 186 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Similarity: A Practical Approach 859 replies

Similarity Trading - The ZZBB Method 215 replies

Rainbow similarity 256 replies

Similarity Stuff 17 replies

Directional Similarity 0 replies

Attachments: The similarity system
Exit Attachments

The similarity system

  • #3,701
  • Jul 21, 2013 3:59pm Jul 21, 2013 3:59pm
Quoting Eurusdd
{quote} I have VERY GOOD candidates for F and G but I am still modifying key parameters. That is one reason I am able to make GOOD CALLS that sometimes do not make sense - taking into consideration NEWS items and the general environment at the time the CALL was made.
So the F and G could be an even lagging indi, however we shall look for something more accurate, or find parameters making them more accurate?

There are approaches basing on pointing-out differences between brokers and require speedy machines to process orders in a way so thousands of small profits are achieved, but this approach is something more. Looks like written in the universe like construction of the leaf. It's amazing.

You mentioned developing some standalone (in relation to trading terminal) SW. Is that due to speed, limitations of MQL, or more sophisticated analysis and constructed F and G candidates?

I see you have a lot of fun with it. I like solving analitycal problems and I know this feeling. Must be great
Time matters most
  • #3,702
  • Jul 21, 2013 3:59pm Jul 21, 2013 3:59pm
Quoting candero
{quote} I have a function on 2 different timeframes, which almost surely agree on the same data set. For number guys and girls let me give you exact number - it is 96.085409252669039145907473309609 percent the signals of that function agree on M1/M5 timeframe. In other words 98% of signals from M5 are there on M1. I must be close. I damn hope so
You cannot just take any two functions and make a claim like this. According to the principle: F and G MUST satisfy a set of conditions! So, I don't know what you have right now!!!1
  • #3,703
  • Jul 21, 2013 4:04pm Jul 21, 2013 4:04pm
big spread for the opening
  • #3,704
  • Jul 21, 2013 4:06pm Jul 21, 2013 4:06pm
Quoting burnssss
you have algorithms that take into account up to the news so to speak. Their decisions are not affected by news. great job
From Stanley Kroll commodities book, "I can tell you what the news will be from the price better than you can tell me what the price will do based on the news."
Please keep this in mind when all the news is against your trade!

I just joined FF because of this thread. Been trading on and off since 1976 (paper trading soybeans at the age of 14) so I think I will be able to contribute something at some point. But you guys that are fast on the uptake make it hard for someone with a full time job to come up with something new before you do.

Will post some indicator ideas soon.
  • #3,705
  • Jul 21, 2013 4:08pm Jul 21, 2013 4:08pm
third function would suffice to complete the model.
  • #3,706
  • Jul 21, 2013 4:11pm Jul 21, 2013 4:11pm
Very interesting methodology, but the thread is getting impossible to read.
It's turning into a garbage dump of BS comments and miscellaneous charts.
Eurusdd's posts only account for 20% of the posts in this thread.
He should be doing the great majority of the posting and the rest of should be reading.
  • #3,707
  • Jul 21, 2013 4:17pm Jul 21, 2013 4:17pm
Quoting Eurusdd
{quote} You cannot just take any two functions and make a claim like this. According to the principle: F and G MUST satisfy a set of conditions! So, I don't know what you have right now!!!1

They do satisfy a conditon! And while doing it, they agree on M1 and M5 96% of the time. I must be close, you said it yourself

On a more serious note, I'll carry on testing, if I have a breakthrough of sorts I'll shout out loud.
  • #3,708
  • Jul 21, 2013 4:19pm Jul 21, 2013 4:19pm
Quoting billbss
Very interesting methodology, but the thread is getting impossible to read. It's turning into a garbage dump of BS comments and miscellaneous charts. Eurusdd's posts only account for 20% of the posts in this thread. He should be doing the great majority of the posting and the rest of should be reading.
Let people have fun!!! I don't have a problem with that!!! No comment is BS - yours inclusive!!!

That is the spirit here - if you don't like it; ignore the thread!!!
  • #3,709
  • Jul 21, 2013 4:21pm Jul 21, 2013 4:21pm
Quoting candero
{quote} They do satisfy a conditon! And while doing it, they agree on M1 and M5 96% of the time. I must be close, you said it yourself On a more serious note, I'll carry on testing, if I have a breakthrough of sorts I'll shout out loud.
Not just any set of conditions. Read the principle again Anyway - good luck!
  • #3,710
  • Jul 21, 2013 4:25pm Jul 21, 2013 4:25pm
Quoting billbss
Very interesting methodology, but the thread is getting impossible to read. It's turning into a garbage dump of BS comments and miscellaneous charts. Eurusdd's posts only account for 20% of the posts in this thread. He should be doing the great majority of the posting and the rest of should be reading.
Hate to sound argumentative in my second post, but I bet you that EURUSDD is thinking things he never thought he would be thinking, due to the feedback he has gotten here. And R's indicator program was modified due to feedback as well. Who is to say what is garbage and what will turn to gold until it is posted?
  • #3,711
  • Jul 21, 2013 4:35pm Jul 21, 2013 4:35pm
EURUSDD as I said has a consistent system, and the fact of making a mathematical expression in Forex which makes it cost effective because it makes me understand a person with great mathematical abilities. I've been quite a few years also in search of a mathematical expression that defines the market movement. EURUSDD data is a spatial form. By using different softwares MT4 is because I assume that needs to process new data continuously applying the mathematical expression, such as MATLAB. Again, the complexity of the model lies in the elaboration of algortimos that under certain conditions, which helps to increase the likelihood, explaining 97! rotation. Simply amazing.
  • #3,712
  • Jul 21, 2013 4:42pm Jul 21, 2013 4:42pm
The comments you make here are very important for me because I want to see if you understand the basic ideas before I can ever take you to the next level. So -I am just giving signals right now so that learning and making money can go hand in hand!
  • #3,713
  • Jul 21, 2013 4:48pm Jul 21, 2013 4:48pm
Quoting billbss
Very interesting methodology, but the thread is getting impossible to read. It's turning into a garbage dump of BS comments and miscellaneous charts. Eurusdd's posts only account for 20% of the posts in this thread. He should be doing the great majority of the posting and the rest of should be reading.
Err.. no. First and foremost, this is not a platform for a classroom with a syllabus. FF is a platform for open discussion. Thread starters are given certain powers over their thread but it is primarily an open platform to discuss trading strategies.

Eurusdd has chosen to reveal a principle in the same way great sages of old have done - cryptically, so that students who are ready can appear. Those who are not yet ready can have a place to discourse... this place.
A pip is only worth it if you know how much you risked to earn it
  • #3,714
  • Jul 21, 2013 4:56pm Jul 21, 2013 4:56pm
I'm a little tired and may have some inconherencia. small steps rather go and so I developed a work plan. I have considered the following:

1 - Perform the test formulas several technical indicators for sampling similarities (dis-similarity) over a period of time. Look about three equations for two TF are a total of 6 equations. (leaving aside time since we are only drafts)

2 - An analysis of the data to see if any treatment necessary data, although unlike artificial intelligence systems, I think in this case it would not be necessary because it would distort the series.

3 - calculation of Hurst coefficient (H> 0.5)

4 - Calculation of fractal dimension.

5 - Representation logarithmic and observe the "regression of the data"

This would be my working algorithm
  • #3,715
  • Jul 21, 2013 4:58pm Jul 21, 2013 4:58pm
Quoting shiva
{quote} MM, isn't MTF an obscene word?
A pip is only worth it if you know how much you risked to earn it
  • #3,716
  • Jul 21, 2013 5:11pm Jul 21, 2013 5:11pm
eurusd has to take out 1.31583 quick!!!

She will go for it sooner rather than later!!!
  • #3,717
  • Jul 21, 2013 5:12pm Jul 21, 2013 5:12pm
Well all the big boys use complex algorithms. I remember a BBC financial journalist was invited to a $billion hedge fund in Zurich and he watched such a computer system make $1 million in 20 minutes.
  • #3,718
  • Jul 21, 2013 5:30pm Jul 21, 2013 5:30pm
but doing so would justify the accuracy of the model, after the new pivot points as would be calculated? by straight regression?
  • #3,719
  • Jul 21, 2013 5:59pm Jul 21, 2013 5:59pm
Quoting Eurusdd
eurusd has to take out 1.31583 quick!!! She will go for it sooner rather than later!!!
2 longs placed as soon as 5M printed a ZZBOT.
A pip is only worth it if you know how much you risked to earn it
  • #3,720
  • Jul 21, 2013 6:05pm Jul 21, 2013 6:05pm
Eurusdd, I found this article for you. They speak your language... talks about a Market Instrument Function...
(way over my head)
Attached File(s)
File Type: pdf Market Instrument Function.pdf   5.8 MB | 1,471 downloads
1 trader viewing now