Quantifying divergence (A question) | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Divergence and hidden Divergence 122 replies

Divergence Question: What would you do? 5 replies

Question on Divergence Set-ups 4 replies

Quantifying our Strategies 18 replies

Hedging Question - Quick Question Looking for A Quick Answer 1 reply

Attachments: Quantifying divergence (A question)
Exit Attachments

Quantifying divergence (A question)

  • #1
  • First Post: Mar 29, 2012 12:32pm Mar 29, 2012 12:32pm
1) Oscillators can indicate divergence. 2) Deceleration is a big part of divergence. 3) But oscillators don't quantify deceleration. What if you could? and what if you can gather statistical knowledge about them? For example do you think there are optimum ranges in the market for such and such time frame for rate of deceleration that can improve turning point prediction?

PS: I am talking about using divergence as part of a setup to trade.
Riding the order flow
  • #2
  • Mar 29, 2012 3:00pm Mar 29, 2012 3:00pm
Quoting Binyamin
1) Oscillators can indicate divergence. 2) Deceleration is a big part of divergence. 3) But oscillators don't quantify deceleration. What if you could? and what if you can gather statistical knowledge about them? For example do you think there are optimum ranges in the market for such and such time frame for rate of deceleration that can improve turning point prediction?

PS: I am talking about using divergence as part of a setup to trade.
I am not sure what you are asking but.... you could begin by obtaining statistical data on (Ranges)(time frames)

high-low= range, then interpret the data

deceleration is pretty easy to see actually, the market does that really often
acceleration, loss of momentum, correction or can be seen as clustering, then acceleration can be seen as liquidity finding the equilibrium in supply and demand.

divergence i am assuming you are using MACD or something of the sort, well its basically when the difference in the exponential averages do not reflect the current formation...but my hypothesis is that because the averages are being calculated to normalize data in different time windows, when price moves up and re tests a level there is less differential from the past in terms that there is less distance from the point when the Exponential readings are correcting to the current points loss in momentum.

the thing is that deceleration although will turn to correction the distance is never really known nor the completion of the correction, what you CAN observe is price clustering and at that point you can infer on which binary decision to take according to the most likely outcome, and with good sound risks parameters.

likely is not = future
AVT INVENIAM VIAM AVT FACIAM
  • #3
  • Mar 29, 2012 7:34pm Mar 29, 2012 7:34pm
Quoting Binyamin
1) Oscillators can indicate divergence. 2) Deceleration is a big part of divergence. 3) But oscillators don't quantify deceleration.
Standard Oscillators don't quantify deceleration, but a Modified Oscillator can identify and quantify deceleration. These Modified Oscillators are very hard and very rare to find.
Your next hurdle is, if you get a quant/physicist/mathematician to formulate it, you still need an expert statistician to dissect the information and incorporate it in a system.
What you are searching for has been done, good luck with your endeavor.

Quoting Binyamin
What if you could? and what if you can gather statistical knowledge about them? For example do you think there are optimum ranges in the market for such and such time frame for rate of deceleration that can improve turning point prediction?
PS: I am talking about using divergence as part of a setup to trade.
Trust me on this one, there are NO optimums.
  • #4
  • Edited Mar 30, 2012 1:41am Mar 29, 2012 11:20pm | Edited Mar 30, 2012 1:41am
Quote
the thing is that deceleration although will turn to correction the distance is never really known nor the completion of the correction, what you CAN observe is price clustering and at that point you can infer on which binary decision to take according to the most likely outcome, and with good sound risks parameters.
Agree but that is the point of doing statistical studies once you can quantify rates of acceleration/deceleration. Cheers.

Quote
What you are searching for has been done, good luck with your endeavor.
Thanks. I figured I am not the 1st person to ask this question, that is why I asked in this forum. Well an idea just came, my 1st place to search will be scientific literatures or 'reports' found in scientific databases. I will have to get a uni student to do me a favour.

Quote
These Modified Oscillators are very hard and very rare to find.
All the more I want to do it now.

Quote
Your next hurdle is, if you get a quant/physicist/mathematician to formulate it, you still need an expert statistician to dissect the information and incorporate it in a system.

Well if this analysis or research have been done before but the knowledge of it is still rare, it could mean that as you said there aren't really any optimum or knowledge in there that can give an edge. I don't know, but I want to find out.
Riding the order flow
  • #5
  • Mar 30, 2012 10:55pm Mar 30, 2012 10:55pm
G'day Bin.

What exactly do you mean by "decelleration/acceleration" in that thread? If it is merely the slowdown of a move, then the slope of a MA shows that. MACD is a way of showing that as one MA changes faster than the other, - a sort of rough derivative (velocity) and basically the OSMA is an acceleration/deceleration indi, as it is a MACD of a MACD (second deivative = acceleration/deceleration).

You might also want to search http://www.mql4.com/ for a start -lots of indis, etc over there with explanation, discussion.

Here are a couple ...
http://ta.mql4.com/indicators/bills/...n_deceleration
http://codebase.mql4.com/247
http://codebase.mql4.com/2308
  • #6
  • Mar 31, 2012 4:41am Mar 31, 2012 4:41am
Quoting Binyamin
1) Oscillators can indicate divergence. 2) Deceleration is a big part of divergence. 3) But oscillators don't quantify deceleration. What if you could? and what if you can gather statistical knowledge about them? For example do you think there are optimum ranges in the market for such and such time frame for rate of deceleration that can improve turning point prediction?

PS: I am talking about using divergence as part of a setup to trade.
The divergence that an oscillator shows is difficult to measure because essentially it is a signal that the indicator has failed. You can see this more clearly if you change the setting from the standard MACD to longer periods and observe the increasing divergence in a trend. The oscillator keeps correcting itself and the signals are meaningless.

You are on the right lines but my advice would be to forget trying to create more indicators (it has already been done) and concentrate more on the behaviour of existing oscillators with chart time and study. Spotting bearish divergence on pullbacks in a downtrend, for example, is well worth while, as opposed to looking for bearish divergence to pinpoint a top in an uptrend.
Gone to a better place
  • #7
  • Apr 2, 2012 5:00am Apr 2, 2012 5:00am
Quote
What exactly do you mean by "decelleration/acceleration" in that thread?
Hi BT! the image bellow illustrates what I mean. In the left of the highlighted region(phase 1) the market drop in 5 bars but after the pull back the market took many more bars to achieve about half the distant as the first(phase 2). That means the velocity have decreased. In the 1st phase the velocity was more or less constant, but in the second phase the velocity changed. Changes in velocity over time is called acceleration. If velocity were the same for phase 1 and 2 the chart will display a straight line not a curve shape like the image bellow.

Quote
a sort of rough derivative (velocity) and basically the OSMA is an acceleration/deceleration indi, as it is a MACD of a MACD (second deivative = acceleration/deceleration).
Ok cheers will check them out and see if I can find anything.
Attached Image (click to enlarge)
Click to Enlarge Name: 7123923 Interbank FX Trader 4 - Demo Account - GBPUSD,M15_2012-03-29_16-59-26.jpg Size: 135 KB
Riding the order flow
  • #8
  • Last Post: Apr 2, 2012 5:04am Apr 2, 2012 5:04am
Quoting smikester
You can see this more clearly if you change the setting from the standard MACD to longer periods and observe the increasing divergence in a trend..
Interesting approach.

Quoting smikester
Spotting bearish divergence on pullbacks in a downtrend, for example, is well worth while, as opposed to looking for bearish divergence to pinpoint a top in an uptrend.
I have been doing that by going into the lower time frame to look for divergence. Good advice.

Quoting smikester
You are on the right lines but my advice would be to forget trying to create more indicators (it has already been done)
Yeah will try and search first.
Riding the order flow
0 traders viewing now