Download MetaTrader 5

Indicators: Harmonic Pattern Finder V2 - page 6

To add comments, please log in or register
danizani95
142
danizani95  
Andre Enger:

Thanks for the feedback.

As far as I'm aware, Elliot waves are not harmonic by themselves, it is a more loosely defined theory of wave structures. Elliot waves have three successively higher tops before two lower bottoms (buy at second bottom), and various compositions and classifications on how this may occur. Some technical analysts seem to believe, the relation between harmonics and Elliot waves is that the X-A segment should correspond to the Elliot impulse phase, and the ABCD part correspond to the Elliot correction. Therefore different harmonic patterns, like the Gartley and the Bat, are different manifestations of the same elliot phenomena.

Adding Elliot wave markers to the indicator is of less use, as the some of the patterns themselves are full Elliot waves. One thing I've considered though in a new version is a filtering mechanism which makes it easy to add user defined filters to the pattern finder. It is then quick to add say an "Elliot wave filter" which removes those harmonic patterns where there is no finer-scaled impulse structure in the XA leg. This could be detected for instance by checking if a ZigZag on a lower timeframe has three successive higher tops.

Regards


Hi Andre thank you for your reply. Now i'm trying to understand what's the difference between the purist and the holist method to calculate prz and what should I use to have better results.. Could you explain to me in a very simple and practical way what change on the chart ? Also I set both the slack to 0.. Do I have a more precise pattern creation with 0 slack ? Thank you

Andre Enger
1840
Andre Enger  
danizani95:

Hi Andre thank you for your reply. Now i'm trying to understand what's the difference between the purist and the holist method to calculate prz and what should I use to have better results.. Could you explain to me in a very simple and practical way what change on the chart ? Also I set both the slack to 0.. Do I have a more precise pattern creation with 0 slack ? Thank you

Hello Danizani,

The discussion on purist vs holist patterns was about two different views on whether or not an occurring pattern really is valid. So for example on the Gartley, you know the D point should be at an 78.6% retracement of X-A and 127% - 161% of B-C. The thing is that different traders have different rules for validating patterns. Some might say that the 78.6% X-A retracement point must be within the 127% - 161% B-C retracement zone, others might claim that either the 127% or the 161% retracement of B-C should be really close to the 78.6% X-A and the price ideally reverse in between those points, yet others might think that the price could go below 161% B-C but the bar should not close there, etc. The “holist” view, which is what is implemented in the indicator, basically

  1. takes each ratio constraint in isolation
  2. applies a slack to it
  3. checks if the price did go that high/low but not higher/lower according to the slack
  4. and does so for each ratio constraint resulting in a unified PRZ.

The “purist” view code-snippet applies a filter to the patterns found using the holist method by additionally stipulating that the PRZ must have the precise ratios inside of it by disregarding slack. So whereas one might holistically say that a Gartley reversing at 75% X-A and 161% B-C is OK, puristically one would not because the point where the X-A is at an exact 78.6% is below the 161% B-C mark. So on the chart this pattern would be removed.

If you set the slack setting to zero all the ratio-constraints will need to be matched precisely, so yes, this increases precision and removes non-fitting patterns. Keep in mind that this might be unrealistic, and the slack setting applies symmetrically to all ratio-constraints on all patterns with respect to the leg they are on. If you are code savvy it is possible to modify this and implement more advanced rules. As is, the way to “calibrate” the slack is to consider your favorite pattern (e.g. the Gartley), pick a leg with a single numbered ratio constraint (e.g. the 78.6% X-A retracement), and find the limit to where you would consider the pattern as valid if it occurred (e.g. “if the price reverses at 72% its OK but not at 71%”). Then you have a slack of 78.6% - 72% = 6.6% and enter “0.066” in the settings. Do the same for interval ratio-constraints to find a suiting setting there.

danizani95
142
danizani95  
Andre Enger:

Hello Danizani,

The discussion on purist vs holist patterns was about two different views on whether or not an occurring pattern really is valid. So for example on the Gartley, you know the D point should be at an 78.6% retracement of X-A and 127% - 161% of B-C. The thing is that different traders have different rules for validating patterns. Some might say that the 78.6% X-A retracement point must be within the 127% - 161% B-C retracement zone, others might claim that either the 127% or the 161% retracement of B-C should be really close to the 78.6% X-A and the price ideally reverse in between those points, yet others might think that the price could go below 161% B-C but the bar should not close there, etc. The “holist” view, which is what is implemented in the indicator, basically

  1. takes each ratio constraint in isolation
  2. applies a slack to it
  3. checks if the price did go that high/low but not higher/lower according to the slack
  4. and does so for each ratio constraint resulting in a unified PRZ.

The “purist” view code-snippet applies a filter to the patterns found using the holist method by additionally stipulating that the PRZ must have the precise ratios inside of it by disregarding slack. So whereas one might holistically say that a Gartley reversing at 75% X-A and 161% B-C is OK, puristically one would not because the point where the X-A is at an exact 78.6% is below the 161% B-C mark. So on the chart this pattern would be removed.

If you set the slack setting to zero all the ratio-constraints will need to be matched precisely, so yes, this increases precision and removes non-fitting patterns. Keep in mind that this might be unrealistic, and the slack setting applies symmetrically to all ratio-constraints on all patterns with respect to the leg they are on. If you are code savvy it is possible to modify this and implement more advanced rules. As is, the way to “calibrate” the slack is to consider your favorite pattern (e.g. the Gartley), pick a leg with a single numbered ratio constraint (e.g. the 78.6% X-A retracement), and find the limit to where you would consider the pattern as valid if it occurred (e.g. “if the price reverses at 72% its OK but not at 71%”). Then you have a slack of 78.6% - 72% = 6.6% and enter “0.066” in the settings. Do the same for interval ratio-constraints to find a suiting setting there.


Ok thanks now I've understood what are the differences and can say that I prefer purist view. I have another doubt, e.g. the crab. The point B is a retraicement of XA between 0.382 - 0.618. This means that all of the values between this two ratios are valid or only harmonic ratios, considering the slack, between the min and max (0.382 - 0.618) can be accepted, e.g. 0.382 - 0.5 - 0.618 +- slack ?

Andre Enger
1840
Andre Enger  
danizani95:

Ok thanks now I've understood what are the differences and can say that I prefer purist view. I have another doubt, e.g. the crab. The point B is a retraicement of XA between 0.382 - 0.618. This means that all of the values between this two ratios are valid or only harmonic ratios, considering the slack, between the min and max (0.382 - 0.618) can be accepted, e.g. 0.382 - 0.5 - 0.618 +- slack ?

First one, all values between the two ratios are accepted on these "in between" retracements. So 0.382 'minus' slack to 0.618 'plus' slack.
danizani95
142
danizani95  
Andre Enger:
First one, all values between the two ratios are accepted on these "in between" retracements. So 0.382 'minus' slack to 0.618 'plus' slack.

I think that second option could be more profitable.. I have read Scott Carney s books and he considers valid patterns only those who have precise ratios to find each point. What do you think about this ?

Andre Enger
1840
Andre Enger  
danizani95:

I think that second option could be more profitable.. I read Scott Carney s books and he considers valid patterns only those who have precise ratios to find each point. What do you think about this ?

I actually thought so too so ran some experiments trying out the different harmonic numbers for the Bat and Gartley on several timeframes and currency pairs. I could not see any notable change in the statistical success rate, i.e., the chance of the ZigZag reversing, so do not think so anymore. Note: the success rate does not take into account the "reward" on the successes which would be dependent on some trailing/take-profit strategy in any case.

What I did find however, was that the reversal chance increased the higher the C-point retracement was, so patterns with a C-point at 0.88 (and even slightly above) tended to reverse more than the others, even those at 0.618 which is a primary harmonic number. This seems reasonable of course since it shows some pressure in the trade-direction.

I've also looked at Scott Carneys work and in many of the examples he talks about the patterns do not have precise ratios.

Additonally: "text book" harmonic ratios are impossible since the math does not add up, there must be some slack in the PRZ.
danizani95
142
danizani95  
Andre Enger:

I actually thought so too so ran some experiments trying out the different harmonic numbers for the Bat and Gartley on several timeframes and currency pairs. I could not see any notable change in the statistical success rate, i.e., the chance of the ZigZag reversing, so do not think so anymore. Note: the success rate does not take into account the "reward" on the successes which would be dependent on some trailing/take-profit strategy in any case.

What I did find however, was that the reversal chance increased the higher the C-point retracement was, so patterns with a C-point at 0.88 (and even slightly above) tended to reverse more than the others, even those at 0.618 which is a primary harmonic number. This seems reasonable of course since it shows some pressure in the trade-direction.

I've also looked at Scott Carneys work and in many of the examples he talks about the patterns do not have precise ratios.

Additonally: "text book" harmonic ratios are impossible since the math does not add up, there must be some slack in the PRZ.

thank you now it's all more clear. For curiosity, what success rate did you find with you experiments ?

Andre Enger
1840
Andre Enger  
danizani95:

thank you now it's all more clear. For curiosity, what success rate did you find with you experiments ?

It depends on the market and pattern, some pairs seem to like a pattern very well while other pairs doesn't. It can be anything from 40 up to 90 %, with about 70% success rate being rather normal. The increase with higher C-point was like the difference between 60-ish% to 70-80-ish%.

Andre Enger
1840
Andre Enger  
danizani95:

thank you now it's all more clear. For curiosity, what success rate did you find with you experiments ?

So just ran the pattern finder again on GBPCAD daily. In my source I've changed the Gartley 113 definition into finding 0.88 C-point Gartleys exclusively. The statistics which are shown in the screenshot tells that 3 out of 4 times the Gartley 113 reversed (75%), while both the standard and Max Gartley have 50% success rate.


This was how my "experiments" were done, did not take notes or anything but remember this conclusion and seeing it also happen with the Bat.

danizani95
142
danizani95  
Andre Enger:

So just ran the pattern finder again on GBPCAD daily. In my source I've changed the Gartley 113 definition into finding 0.88 C-point Gartleys exclusively. The statistics which are shown in the screenshot tells that 3 out of 4 times the Gartley 113 reversed (75%), while both the standard and Max Gartley have 50% success rate.


This was how my "experiments" were done, did not take notes or anything but remember this conclusion and seeing it also happen with the Bat.

What movement from point D do you consider a reversal ? Do you consider a specific ratio to reach ?

To add comments, please log in or register