Detecting Fake Price Movements - Convergence Divergence Indicator | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Destroy Markets With Convergence and Divergence 39 replies

Detecting Fake Price Movements thread - Warning to Programmers 14 replies

Detecting entry point with Zig Zag indicator? 1 reply

Convergence/Divergence (wip5.51.ex4) 32 replies

Relative Strength Convergence Divergence? 4 replies

Attachments: Detecting Fake Price Movements - Convergence Divergence Indicator
Exit Attachments

Detecting Fake Price Movements - Convergence Divergence Indicator

  • #1
  • First Post: Edited May 19, 2017 11:17am May 7, 2017 8:28am | Edited May 19, 2017 11:17am
Thread Closed.
  • #2
  • May 7, 2017 9:59am May 7, 2017 9:59am
hi, i use this indicator
Attached File(s)
File Type: mq4 _Correlation IND HSTlite.mq4   7 KB | 1,231 downloads
  • #3
  • May 7, 2017 10:09am May 7, 2017 10:09am
Fake price movement? I don't understand this term. Anytime price moves - it's real - there's a reason behind every single price move that occurs in the market.
  • #4
  • May 7, 2017 11:40am May 7, 2017 11:40am
I like Bill Williams.
  • #5
  • May 7, 2017 11:45am May 7, 2017 11:45am
Nice eh?
Attached Image (click to enlarge)
Click to Enlarge Name: C__Data_Users_DefApps_AppData_INTERNETEXPLORER_Temp_Saved Images_usd_cadbil1_53.png Size: 34 KB
  • #6
  • May 7, 2017 12:00pm May 7, 2017 12:00pm
Quoting samsha786
THREAD RULES: If you discuss or attach anything unrelated to this thread you will be banned. Please read the provided pdf. Make sure you understand and have context. We look at developing a new pair trading method to detect fake price movements and arbitrage opportunities that is based on a convergence/divergence indicator (CDI) belonging to the oscillatory class. The proposed technique is applied to a cross-currency pair EURAUD and trading rules based on CDI signals are obtained. The CDI indicator is shown to outperform others of the oscillatory...
Sounds Interesting Samsha. Subscribed.

Good Luck!
  • #7
  • May 7, 2017 12:35pm May 7, 2017 12:35pm
Quoting aaven
{quote} Sounds Interesting Samsha. Subscribed. Good Luck!
Thanks!
  • #8
  • May 7, 2017 1:00pm May 7, 2017 1:00pm
I've tried some variations on these strategies but have not had much luck. I will have a look at your indicator requirements though and see if I can help. I did not know people actually wrote thesis on this stuff... being a scientist, I find it exciting to see a well researched strategy like this.
Better to run, than curse the road
  • #9
  • May 7, 2017 1:01pm May 7, 2017 1:01pm
You need to give us more information of changes to make for the indicator. I think maybe the indicator levels might be off by a digit?
  • #10
  • May 7, 2017 1:17pm May 7, 2017 1:17pm
Quoting steve2010
I've tried some variations on these strategies but have not had much luck. I will have a look at your indicator requirements though and see if I can help. I did not know people actually wrote thesis on this stuff... being a scientist, I find it exciting to see a well researched strategy like this.
Hi Steve

Welcome! Thanks for joining the thread. As the indicator on which we will base our strategy is well documented, backed by research and historical findings - there must be some merit to it. This is a play on statistical arbitrage, which implies that it should give us some kind of edge.

You being a scientist makes it all that more exciting
  • #11
  • May 7, 2017 1:19pm May 7, 2017 1:19pm
Quoting packgreen49
You need to give us more information of changes to make for the indicator. I think maybe the indicator levels might be off by a digit?
Hi Packgreen

Welcome! What specifically are you looking for? After you calculate correlation for the 90 day and 24 hour periods - you should end up with values between 0.0 and 1.0. The correlation calculation ensures this. Here's a simple explanation of correlation:

https://www.mathsisfun.com/data/correlation.html
  • #12
  • May 7, 2017 1:37pm May 7, 2017 1:37pm
In order to program this indicator, you need a formula and something to go on. The pdf does not say too much.
D.
  • #13
  • May 7, 2017 2:06pm May 7, 2017 2:06pm
Quoting DirkH143
In order to program this indicator, you need a formula and something to go on. The pdf does not say too much. D.
Well, it seems to just use correlation determined over two separate periods. It should basically be this stuff: https://en.wikipedia.org/wiki/Pearso...on_coefficient
I'll try code it tomorrow after work and see
Better to run, than curse the road
  • #14
  • May 7, 2017 2:57pm May 7, 2017 2:57pm
Quoting samsha786
... We look at developing a new pair trading method to detect fake price movements and arbitrage opportunities that is based on a convergence/divergence indicator (CDI) belonging to the oscillatory class. ....
Hi samsha,
Your approach looks interesting and promissing. Exploring correlation I've also noticed that it gives some beautiful signals. You may read my findings here and have a look at some screenshots. There is also my Pearson correlation indicator attached to this post which you may use.
Here are my comparing conclusions Pearson vs Spearman correlation.
I'll study deeper your CDI pdf and try to contribute a bit later.

Quoting samsha786
{quote} ... This is a play on statistical arbitrage, which implies that it should give us some kind of edge.
Statistical arbitrage is based on cointegration which is different from correlation.
BR, alex
  • #15
  • May 7, 2017 3:01pm May 7, 2017 3:01pm
Looks very interesting.
In the detail appendix of the PDF it shows a return of approx 5% total in 3 years, which is only about 1.6% a year. Am I reading this correctly?
Still I think it is worth following up, maybe with shorter time frames.
Have you tried asking the authors if it is possible to get a copy of their indicator?
  • #16
  • May 7, 2017 3:20pm May 7, 2017 3:20pm
Nice idea, I am watching closely.
  • #17
  • Edited 12:48am May 8, 2017 12:22am | Edited 12:48am
Quoting OrdrFlowEdge
Fake price movement? I don't understand this term. Anytime price moves - it's real - there's a reason behind every single price move that occurs in the market.
Hi

What do we mean by fake price movement?

If for the past 90 days 2 securities have a high positive correlation (move in same direction) of 70% or more then if within a 24 hour period they diverge (move in opposite directions), that is an indication of a temporary move - they are most likely to become positively correlated again.

If for the past 90 days 2 securities have a high negative correlation (move in opposite directions) of 70% or more then if within a 24 hour period they converge (move in same directions), that is an indication of a temporary move - they are most likely to become negatively correlated again.
  • #18
  • May 8, 2017 12:24am May 8, 2017 12:24am
Quoting sevan62
{quote} Hi samsha, Your approach looks interesting and promissing. Exploring correlation I've also noticed that it gives some beautiful signals. You may read my findings here and have a look at some screenshots. There is also my Pearson correlation indicator attached to this post which you may use. Here are my comparing conclusions Pearson vs Spearman correlation. I'll study deeper your CDI pdf and try...
Hi Sevan

Welcome! The attached paper refers to this method specifically as correlation analysis - unless I've misunderstood. Looking forward to your contributions.
  • #19
  • May 8, 2017 12:30am May 8, 2017 12:30am
Quoting profitkeeper
Looks very interesting. In the detail appendix of the PDF it shows a return of approx 5% total in 3 years, which is only about 1.6% a year. Am I reading this correctly? Still I think it is worth following up, maybe with shorter time frames. Have you tried asking the authors if it is possible to get a copy of their indicator?
Hi

Yes, I have attempted to contact these authors - unfortunately, no response. Remember, the article basically describes the CDI indicator to provide more accurate signals then the popular RSI. 5% in this context is irrelevant. Why? CDI will form the basis of our strategy, we will obviously be adding additional filters that make sense and we will employ some form of money management.

Baby steps! First let's create the CDI indicator. Remember, there is no "Magic Wand" or "Holy Grail", there are however tools which are better suited for particular situations.

We know from the attached paper that they ran a test with CDI by itself and it was profitable. So there is opportunity here to exploit and develop the idea further.
  • #20
  • May 8, 2017 12:32am May 8, 2017 12:32am
Quoting DirkH143
In order to program this indicator, you need a formula and something to go on. The pdf does not say too much. D.
Hi DirkH

Just to add to what Steve2010 has already posted, here's a simple explanation and example of the correlation formula:
https://www.mathsisfun.com/data/correlation.html

Excel also has a function called CORREL, which can be used to check if the indicator results are correct or not.
0 traders viewing now