The key being able to define it in absolute terms...
Saying "a price turn that happens near the same level as another price term, when both turns are at an extreme and both terms are in a reasonable number of bars from each other", doesn't really make it.
Saying "A price turn that happens within 10% of a recent price turn, when both price turns happen after a 50% move opposite of the previous pivot and where the two turns are within 10 bars of each other" is a good start (but still leaves a few questions unanswered).
And no, I don't think the above is a good definition for a 2xtop or bottom, it is for example purposes only.
I could build something that identifies a resistance level or support level (highest high/lowest low after a defined up/down trend). Then look for a minimum retracement (say 4~10%), then a second attempt to reach the support/resistance level. But this would only be the pre-qualification for a double top or bottom.
The real trick would be to identify the "final move" as early as possible with some accuracy.
I think we could start with a "Pivot" app - like my TT Pivots - and build from there. As others have stated, the hardest part if to define the variations of the double top/bottom accurately so the EFS can identify them.
This is not an easy process to complete.. kinda subjective..
How far is the look back period - 2 weeks, 30 min?
How much of a retrace is required - 4-10% intraday is too large, should probably exresss in ticks?
I think the two paramters are related. If I found a 2xSomething formed in 10 5 minute bars, the allowable retrace would be smaller than if I found a similar pattern using 300 5 minute bars. I would guess you would want a dynamic % retrace which is based on the number of bars in the pattern.
Of course volaility of the symbol might effect the retrace amount also.
The classical 2x signal is when, on a closing basis, the current price breaks the retrace hi/lo that happened between the two points. So that is well definied, but I do agree that it would be nice to flag the 2x formation BEFORE it signals, but not until it has reached some reasonable level of maturity.
But other parameters that you would have to consider along with those you mentioned:
How small does the delta of the prices of the two points forming the 2x pattern have to be. If the two prices have to be exactly the same, well then you will get very few signals and also you would be ignoring some very valid patterns. Since this value will vary between symbols, it is best to use a % or ratio.
You have to consider the price movement before the formation...not just the retrace between. After all a 2x pattern needs to be reversing something.
And of course, it would be very nice to verify that volume is confirming the formation...
Did anyone work on any code for this double top/double bottom indicator? I have a pretty solid definition I could lend to the project but I could use the coding help. I know this thread is a decade old, but I do not see any newer threads looking for double bottoms.
Comment