Perturbation theory with Damped Harmonic Oscillation
Hello World.
I'm a bit of a math geek, there's pieces of paper on the wall that say I'm supposed to be smart (PhD, Chemical Engineering, 2003 and BS, Magna Cum Laude, Chemistry, 1999) but stupid is as stupid does, I guess. Anyway, I've been trying to get this code done-- hopefully by end of Q4 here- and then ported to Pine by sometime either end of Q1/beginning of Q2 next year.
The idea is simple. Not easy, but simple. Frequently, when looking at charts- any chart, really- I see a perturbation, meaning a large PA change. Direction is irrelevant for now, but these large PA changes can be modeled with a delta function. I'm partial to Dirac delta functions, only because I know how to handle them in my favorite languages, but any perturbation method will suffice. Stopping right there: The MAGNITUDE and DIRECTION of the perturbation actually DO matter, because they will suggest the correct dampening factors and correct initial amplitude of the damped harmonic oscillation. I've steadily been accumulating a list of "this happens, then this much perturbation occurs, and in this direction" --- things like the queen dies, JPOW talks and raises rates, Ukraine gets raided, LUNA collapses, etcetera etcetera. These are correlary only, but they usually give a good starting point for the models.
Once the perturbation itself is modelled, it will lead to the aforementioned initial amplitude and dampening factors for the harmonics. The only thing remaining then is the timescale- which is, currently, only correlary to the initial perturbation event. That correlation is somewhat weak and I'm still analyzing how to model it correctly.
The other part that is also difficult to predict is whether there is a scalar that "tilts" the harmonic up or down. So far, the best I've figured is to use PA leading up to the perturbation to "guide" whether price will fall or rise or simply average to a new.... average.
SInce I don't have this in pine yet, it is hard to publish here, but I thought I'd at least get a head start (and a kick in the ass for myself to finish it) by putting it out into the tradingverse.
Happy Trading, and if you have any comments, suggestions, or other insight please comment!!!
I'm a bit of a math geek, there's pieces of paper on the wall that say I'm supposed to be smart (PhD, Chemical Engineering, 2003 and BS, Magna Cum Laude, Chemistry, 1999) but stupid is as stupid does, I guess. Anyway, I've been trying to get this code done-- hopefully by end of Q4 here- and then ported to Pine by sometime either end of Q1/beginning of Q2 next year.
The idea is simple. Not easy, but simple. Frequently, when looking at charts- any chart, really- I see a perturbation, meaning a large PA change. Direction is irrelevant for now, but these large PA changes can be modeled with a delta function. I'm partial to Dirac delta functions, only because I know how to handle them in my favorite languages, but any perturbation method will suffice. Stopping right there: The MAGNITUDE and DIRECTION of the perturbation actually DO matter, because they will suggest the correct dampening factors and correct initial amplitude of the damped harmonic oscillation. I've steadily been accumulating a list of "this happens, then this much perturbation occurs, and in this direction" --- things like the queen dies, JPOW talks and raises rates, Ukraine gets raided, LUNA collapses, etcetera etcetera. These are correlary only, but they usually give a good starting point for the models.
Once the perturbation itself is modelled, it will lead to the aforementioned initial amplitude and dampening factors for the harmonics. The only thing remaining then is the timescale- which is, currently, only correlary to the initial perturbation event. That correlation is somewhat weak and I'm still analyzing how to model it correctly.
The other part that is also difficult to predict is whether there is a scalar that "tilts" the harmonic up or down. So far, the best I've figured is to use PA leading up to the perturbation to "guide" whether price will fall or rise or simply average to a new.... average.
SInce I don't have this in pine yet, it is hard to publish here, but I thought I'd at least get a head start (and a kick in the ass for myself to finish it) by putting it out into the tradingverse.
Happy Trading, and if you have any comments, suggestions, or other insight please comment!!!
Comment:
TO take this one step further: Once the damped harmonic oscillation is established by the model, it predicts entry and exit points in BOTH DIRECTIONS, meaning it's good for Long and Short. Take long at the bottom, exit at the top of the oscillation, and enter short, and so on. Easy money. Happy trading!!!