Logarithmic Spirals appear in nature and may be helpful to predict if a price bar is a potential turning point. I represent them similar to the way in this book with a Center Point -- x,y coordinate First Point -- x,y coordinate Growth Rate (greater than one) -- how much the distance to the center is multiplied for each revolution added Direction -- counterclockwise or clockwise. In each the following charts, The center point is a + Five Wave points defined similar to those in this post are squares Each bar is for a calendar day, and when the market is closed, the price is linearly interpolated from the previous market close to the next market open Prices are adjusted for splits and dividends, and the prices and times are scaled from 0 through 100 from the actual price and bar index values of the wave points The bar the most recent wave point was detected on has an X at the end The triangle at the end of a bar is the current bar to be checked. It will always be at or after the X. Bars after that show what happened next The dates on the X axis label are for the first wave point and the current bar The small diamonds are points on the spiral separated by one degree of arc. To keep the charts readable, the spirals are cut off if they go too much out of the chart time and price range I wrote software to find a spiral that goes near the wave points. The code uses the scaled x,y coordinates of the wave points and no other market data to create the spirals. When the spiral also goes near the current bar, the current bar seems to often be at or near the next wave point. Has anyone used logarithmic spirals to help with trading?
No, but but I love the intellectual stimulation. I've always wondered how to apply fib spirals. This uses a Fourier Series to model complex wave patterns; The ability to re-create a line drawing simply using circles is impressive, although the main criticism would be literally be curve-fitting.
When I was a child there was a mathematician in my home town that lost his marbles, he used to draw spirals all over the city connecting things with it. He sprayed the city for months with all sort of spirals and comments underneath them. So every time I see these spirals I think, "ok... another one. First sign of madness."
What I did is literally curve fitting, since it's trying to fit a curve to the wave points. I would also consider putting just about any markings on a price chart as curve fitting. Speaking of spirals and madness here is a product with a feature that has its users fit artifacts to price charts like this. That chart has four golden spirals (logarithmic spiral whose growth rate is =~ 1.618 for each 90 degrees of arc) and four lines that users would manually fit to the candlesticks!
Looks like if green axis line was lined up with that low it would then line up with next swing high/DB at next green axis line perfectly rather than "in the ballpark". But like any line(s) it all depends of using the correct anchor points that PA is responding to.
Yes, that tool and Investor/RT require users to put the spirals in the right position. Another feature of these tools is they use a fixed growth rate for the spirals. The "Fibonacci Vortex" from wave59.com appears to only use golden spirals (growth rate of Φ ≅ 1.618 per quarter revolution), and logarithmic spiral tool from Investor/RT appears to use only Φ as its growth rate per whole revolution. In the natural world, logarithmic spirals have different growth rates. https://sites.millersville.edu/rumble/Math.457/nautilus.pdf I thought fixing the growth rate was very limiting, so I made my software pick the growth rate and also center point, and first point of found spirals. And my software uses clockwise spirals for predicting swing low points and counterclockwise spires for predicting swing high points. Originally, the code picked the spiral direction, but I found clockwise for potential long trades and counterclockwise for potential short trades seemed more accurate. Someone else here agrees with the direction part. One difference between what I did and what I've seen others do for logarithmic spirals is my software does not put the center point on a wave point (like it's done in the Investor/RT image above). Keeping the center point separate also seemed more accurate for the predictions, and my method finds a center point on or not too far off of the chart.
Some theory about spirals by Walter Russell; https://www.flickr.com/photos/esaruoho/albums/72157604234096846