Trading Focal Points -- "A Leap Of Faith"... | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Broker Challenge "NDD/STP" vs. True "ECN" and "Mkt Mkr" broker talk 46 replies

Leap of faith... 18 replies

The old "Control Points" vs. "Every Tick" testing question 2 replies

Leap of Faith 9 replies

Leap of Faith 6 replies

Attachments: Trading Focal Points -- "A Leap Of Faith"...
Exit Attachments

Trading Focal Points -- "A Leap Of Faith"...

  • #21
  • Mar 22, 2022 6:46pm Mar 22, 2022 6:46pm
Another little bird asked what's with the backshiftng? Why backshift the SMA lines? Again, I haven't read the original reseach but I do have a hunch why...

The SMA(16), SMA(32), SMA(64) and SMA(128) are very different. Their moving average periods differ by a lot. Without backshifting, these SMAs would rarely cross. But by backshifting 1/2 their period then they come into better alignment and Focal Points occur more frequently.

However, just as in real-life, everything comes with a trade-off. The trade-off with backshifting is that the SMA lines end prematurely and require extrapolation to extend to the end of the chart. Those extensions are not based on price data and are simply extrapolations from where the SMA had ended many bars ago. If market conditions have changed since the SMAs had ended then the SMAs will appear to have no relation to the current price bars. However, according to the theory, the TargetPrice may still be good. That's where the Fractal Channel and your discretion come into play.

On the chart below the SMA extensions include small dots. If you remove those portions of the SMA lines then it should become clear what I'm talking about.

Attached Image (click to enlarge)
Click to Enlarge Name: Focal-Points-Introduction.png Size: 64 KB
  • #22
  • Mar 22, 2022 6:51pm Mar 22, 2022 6:51pm
Quoting robots4me
A little bird asked me a question -- what's with all these power-of-2 SMA lines? Why power-of-2? Why not power-of-3 or power-of-1.5? What's "magical" about power-of-2? I haven't read the original research papers from 60 years ago, but I have a good idea why power-of-2 is used. Back in the 1960's computers didn't have the benefit of today's floating-point processors. Those guys at the Jet Propulsion Laboratory were working on missile guidance systems and the computations had to be fast, fast, fast. Computer processors think in binary, and computations...
now looking back to my student days writing machine code with a 4-bit Motorola processor ... moving bits from one register to another... it was then not easy.

then came the 8 bits, then the 16 bits, and now how many bits processors... millions millions of transistors do the job super fast.
Honesty is a very expensive gift. You wont find it in cheap people.WBuffett
  • #23
  • Mar 22, 2022 6:56pm Mar 22, 2022 6:56pm
about Double Bottom the EURAUD 1 mn right now
Attached Image(s) (click to enlarge)
Click to Enlarge Name: screenshot.png Size: 157 KB
Click to Enlarge Name: screenshot.png Size: 138 KB
  • #24
  • Mar 22, 2022 6:57pm Mar 22, 2022 6:57pm
Quoting parisboy
{quote} Well it is a French guy! I have some knowledge of his work and tons of papers on its use in cycle analysis. (frequencies in time series). Spectral Analysis , Periodogram.
Wow, wow -- I am truly impressed. The software we develop for troubleshooting RF uses the Fourier Transform alogorithm to extract the accumulated frequencies (i.e. sine waves) into individual frequencies that are then displayed by the software.

So now we have a convergence of missile guidance systems, financial trading and RF diagnostics. Perhaps at some level every system can be studied as a sine wave?
  • #25
  • Mar 22, 2022 7:02pm Mar 22, 2022 7:02pm
Observe how the Centered Moving Average 128 in in the Center of Price fluctuations

Compare with SMA 128 (not centered)
Attached Image(s) (click to enlarge)
Click to Enlarge Name: screenshot.png Size: 109 KB
Click to Enlarge Name: screenshot.png Size: 127 KB
  • #26
  • Mar 22, 2022 7:09pm Mar 22, 2022 7:09pm
Quoting robots4me
{quote} Wow, wow -- I am truly impressed. The software we develop for troubleshooting RF uses the Fourier Transform alogorithm to extract the accumulated frequencies (i.e. sine waves) into individual frequencies that are then displayed by the software. So now we have a convergence of missile guidance systems, financial trading and RF diagnostics. Perhaps at some level every system can be studied as a sine wave?

I am dubious on the results. . It depends a lot on how you group the results together. I wil pst what I have oon the subject if you are interested. (the Hurst Model)
  • #27
  • Mar 22, 2022 7:19pm Mar 22, 2022 7:19pm
it is late here - see you tomorrow
  • #28
  • Mar 22, 2022 7:21pm Mar 22, 2022 7:21pm
Quoting parisboy
{quote} I am dubious on the results. . It depends a lot on how you group the results together. I wil pst what I have oon the subject if you are interested. (the Hurst Model)
Yes -- please, go for it. I'm not sure how much I'll understand or even know what question(s) to ask, but if you are up to it then let's keep on trucking and see where it leads.

BTW -- now that we've got the initial posts out of the way I'll relax the posting requirements so that others can participate if they are so inclined.
  • #29
  • Mar 22, 2022 7:22pm Mar 22, 2022 7:22pm
Quoting parisboy
it is late here - see you tomorrow
Chat with you later -- and thank you @parisboy...
  • #30
  • Edited 7:43pm Mar 22, 2022 7:32pm | Edited 7:43pm
Attached File(s)
File Type: pdf FOURIER TRANSFORM FOR TRADERS-Copier.pdf   233 KB | 954 downloads
Attached File(s)
File Type: pdf Fourier Spectral Analysis - W.Taylor (Cycles 1985)-Copier.pdf   247 KB | 434 downloads
File Type: pdf SHORT TERM FOURIER ANALYSIS-Copier.pdf   982 KB | 524 downloads
  • #31
  • Mar 22, 2022 8:06pm Mar 22, 2022 8:06pm
Quoting parisboy
{image} {image} {image}
Thanks -- I did browse the first paper "Fourier Transform for Traders" by John Ehlers and got the drift. In the end he concluded that the resolution of his Discrete Fourier Transform (DFT) did not provide sufficient accuracy in low signal-to-noise environments. So, though not directly useful for trading, it's still interesting to learn that someone as accomplished as John Ehlers was even exploring the possibility of applying DFT to trading.
  • #32
  • Mar 22, 2022 8:22pm Mar 22, 2022 8:22pm
Philisophical note: I suspect at some level everything boils down to waves / cycles / frequencies / pendulum behavior. Consider this:

* sub-atomic particles are associated with particular frequencies
* light travel can be described in terms of a wave
* sound travels as a wave
* the earth revolves around the sun in a cyclic manner
* the ocean currents behave as waves
* there is the "circle of life"
* and so on

I believe the reason is because all of these can be represented in terms of a system that hovers around a midpoint / equilibrium / fair market value. The cyclic behavior is caused by feedback mechanisms that repeatedly pull the system back to midpoint whenever it strays too far to one side. Pulling it back in one direction causes it to travel too far and so a different feedback mechanism pulls it back in the other direction. Competing feedback mechanisms pulling the system back and forth across a midline creates the cyclic behavior.

So, how does this apply to trading? Well -- certain indicators are better than others at revealing the pendulum behavior. When you trade in the direction of the pendulum then that provides an edge -- in my opinion...
  • #33
  • Mar 22, 2022 8:29pm Mar 22, 2022 8:29pm
Quoting robots4me
A little bird asked me a question -- what's with all these power-of-2 SMA lines? Why power-of-2? Why not power-of-3 or power-of-1.5? What's "magical" about power-of-2? I haven't read the original research papers from 60 years ago, but I have a good idea why power-of-2 is used..
Interesting idea but this is not the reason Hurst did it this way. The reason for using multiples of 2 for moving average lengths relates to the nature of sine waves and cycles and how they interact. I'll attempt to show rather than tell so you can reproduce this if you wish. If you want more details and examples, read Brian Millard's "Channels and Cycles" book which I think covers this in greater detail.

Below is a sine wave (white) that has two cycles of 200 periods each. The vertical magenta line shows where the first cycle ends and the second cycle starts. Each cycle is a complete 360 degrees or in other words it returns back to the starting point.
Attached Image (click to enlarge)
Click to Enlarge Name: Focal_1_of_4.png Size: 2 KB

To isolate the cycle, a moving average of half the cycle length is used. The cycle is 200 periods so a 100 period moving average is shown in red below. This moving average has interesting properties because it is related to the underlying cycle length of the white sine wave (divided by 2). The yellow vertical lines show where the moving average slope changes from up to down, which also matches where the sine wave (price) is half way between up and down. The middle magenta line shows where the first cycle ends but also where the red MA changes slope from down to up and matches the 50% mark between the cycle low and cycle high.
Attached Image (click to enlarge)
Click to Enlarge Name: Focal_2_of_4.png Size: 3 KB

The next picture below shows cyan vertical lines that indicate the high and low of the cycle. This is of interest to traders who wish to sell/buy at a good price. The cyan lines match up with the 50% mark of the red moving average. In essence these cyan vertical lines are showing the focal points. But how can one identify the 50% mark of a moving average?
Attached Image (click to enlarge)
Click to Enlarge Name: Focal_3_of_4.png Size: 3 KB
The final picture below shows a green horizontal line. This is also a moving average of the sine wave but it is 2x the length of the red line or in other words it is equal to the full cycle length (200). So we see that where the red line and the green line cross at the zero horizonal line, it also matches up with the cyan vertical lines. Said another way, the 1x length moving average crosses the 2x length moving average at the focal point and this moving average crossing can be used to identify highs and lows of a cycle.
Attached Image (click to enlarge)
Click to Enlarge Name: Focal_4_of_4.png Size: 3 KB


This is the theory anyway. In practice price is not a sine wave but as can be seen from parisboy's many examples, the theory can be applied and is useful quite often. The theory shows that you really just need a MA with half the cycle length and a second with the full cycle length. Tillman used a convergence of moving averages of length, length*2 and length*4 (where length should be half the cycle length) because in practice it can get a bit messy with just two averages and the cycles aren't perfectly repetitive or bounded (trends exist).
  • #34
  • Mar 22, 2022 10:32pm Mar 22, 2022 10:32pm
Quoting FXEZ
{quote} Interesting idea but this is not the reason Hurst did it this way. The reason for using multiples of 2 for moving average lengths relates to the nature of sine waves and cycles and how they interact. I'll attempt to show rather than tell so you can reproduce this if you wish. If you want more details and examples, read Brian Millard's "Channels and Cycles" book which I think covers this in greater detail. Below is a sine wave (white) that has two cycles of 200 periods each. The vertical magenta line shows where the first cycle ends and the...
Thank you very much -- I need a bit more time to digest all this. One question though: You begin by referring to "multiples of 2". I understand that powers-of-2 are also multiples of two, but there is also is a big difference, right?

Back in the day when memory was limited to 64K running on 8-bit CPUs then then your code really had to match with the CPU's capability. It didin't matter how elegant the theory -- the real-time performance was critical and using power-of-2 integers saved instructions. Saving on the number of instructions was very, very important. In the 1990's I worked on real-time brain-imaging systems, coding in C. My co-engineer would keep track of the number of lines of assembly code. Back then performance outweighed the elegance of an algorithm. And on a guided missile they didn't have access to a main frame computer for computational purposes -- this all had to be run on miniature hardware in a missile head.
  • #35
  • Mar 22, 2022 10:39pm Mar 22, 2022 10:39pm
Quoting robots4me
Philisophical note: I suspect at some level everything boils down to waves / cycles / frequencies / pendulum behavior. Consider this: * sub-atomic particles are associated with particular frequencies * light travel can be described in terms of a wave * sound travels as a wave * the earth revolves around the sun in a cyclic manner * the ocean currents behave as waves * there is the "circle of life" * and so on I believe the reason is because all of these can be represented in terms of a system that hovers around a midpoint / equilibrium / fair market...
Continuing with the pendulum analogy...

Does anyone else see how well DssBressert -- the "pendulum" indicator -- aligns with Focal Points?

Attached Image (click to enlarge)
Click to Enlarge Name: Focal-Points-plus-DssBressert.png Size: 87 KB

Attached File(s)
File Type: ex4 DSS Bressert.ex4   12 KB | 539 downloads
  • #36
  • Mar 23, 2022 3:55am Mar 23, 2022 3:55am
Using the DSS Bressert.ex4 as a possible over sold / brought indicator to show your swinging back to the target point?
Pip Student
  • #37
  • Mar 23, 2022 3:58am Mar 23, 2022 3:58am
Interesting thread.
I was confused at first, because you went on the first few posts, without actually defining or explaining what constitutes a "focal point". That is, until PB came along to clear things out. So apparently they are a bit discretionary, but on 2 screenshots of the same pair, you draw 2 different sets of focal points.
Can you explain the difference?

Attached Image (click to enlarge)
Click to Enlarge Name: FP1-High-FP2.png Size: 61 KB

Attached Image (click to enlarge)
Click to Enlarge Name: FP-Trading-Strategy.png Size: 71 KB

I can see that on one, it's when the majority of MA's cross the 128. But on the other screenshot of the same price action structure, it seems to be more arbitrary, like when red & blue are somewhat close together, but it makes less sense.

Thanks for explaining,
  • #38
  • Mar 23, 2022 4:05am Mar 23, 2022 4:05am
I would think from just observing that there might be a need to further seperate focal points into major and minor focal points.

The big yellow circled are major focal points
The squares are minor focal points

Would that make sense?

@Pairsboy @Robots4me

I am closely following this Thread but I likely wont write much, I know that I dont know anything about this. I might ask some stupid questions here and there though if it doesnt disturb the Thread.
  • #39
  • Mar 23, 2022 4:15am Mar 23, 2022 4:15am
Quoting FXEZ
{quote} Interesting idea but this is not the reason Hurst did it this way. The reason for using multiples of 2 for moving average lengths relates to the nature of sine waves and cycles and how they interact. I'll attempt to show rather than tell so you can reproduce this if you wish. If you want more details and examples, read Brian Millard's "Channels and Cycles" book which I think covers this in greater detail. Below is a sine wave (white) that has two cycles of 200 periods each. The vertical magenta line shows where the first cycle ends and the...

Thank you for this great post FXEZ
  • #40
  • Mar 23, 2022 4:17am Mar 23, 2022 4:17am
Quoting Oliver.Twist
Interesting thread. I was confused at first, because you went on the first few posts, without actually defining or explaining what constitutes a "focal point". That is, until PB came along to clear things out. So apparently they are a bit discretionary, but on 2 screenshots of the same pair, you draw 2 different sets of focal points. Can you explain the difference? {image} {image} I can see that on one, it's when the majority of MA's cross the 128. But on the other screenshot of the same price action structure, it seems to be more arbitrary, like...
Keep in mind this is a continuation of the Focal Points discussion that @parisboy and I began in the PlanetX thread. Wasn't that apparent from what I wrote in post #1?

The definition of a Focal Point is not discretionary -- it is well-defined. The crossing of backshifted SMA lines. However, they don't necessarily look as perfect as the cherry-picked ones that have been chosen as examples. Now, as far as how you trade them then, yes, I would say that part is highly discretionary and doesn't lend itself to full automation. However, since you can walk away after opening a trade then one could say they were "semi" automated.

I think you'll need to re-read the first few posts until the light bulb goes on. It sounds like you are attempting to fit Focal Points into your current trading methodology. Forget about market structure. Forget about price action. Re-read my first few posts without any bias. That doesn't mean you have to agree. Your initial goal should be to just get your head around it. It's not my or @parisboy's algorithm -- as I wrote above, Focal Points have been around for a long time and have been used in a variety of systems and later adapted to financial markets.

@parisboy is much more familiar with Focal Points, their origin and history. I'm here only as the mechanic who builds an indicator that may hopefully facilitate other members to look into them more deeply and consider adding them as a tool to their toolbox.
0 traders viewing now