If you fit a complex enought model to a data set, the algorithm will eventually claim to find a fit. Unless you can explain why the resulting model makes sense, expect that the addition of new data will not fit and the model will be proved useless. Any model of price movement that does not consider underlying value, would be also be considered by me to be highly suspect.
Anchored VWAP can be useful. eMini today - VWAP's drawn from 830am (PPI release) and 930am RTH open lining up @ 110pm swing low:-
The simple model that I like is that price is analogous to pink noise, until you get near the Nyquist frequency. Because financial data is sampled data, near the Nyquist frequency introduces a lot of aliasing noise. These articles by John Ehlers explains, and shows what to do about it.
Actually … You might want to see people front running the level. This means price will slow down ahead of the level and reverses some tick / point before it. If the price run through it and the speed accelerate for a little while then stops might have been triggered. I can’t predict where the price is going to reverse and I don’t need to. I look for a reversal. What’s sometimes helpful (outside of events) are the measured moves. If the previous bull leg was 5 points then we can expect the current or next one to be about 5 points. Maybe confluence between different elements could give a nice probability of a reversal. The trend … the time … the measured moves … previous levels … etc Personally I always try to trade “The right side of the V” - Lance Breitstein
The paper gets even the basic math for sampling wrong. (Sampling is not multiplying two sine waves. Here is a link if you want to learn the real math: https://ocw.mit.edu/courses/6-341-d...05/b3c3a089731eef0250a6eefe1d8c78a0_lec04.pdf)
%% Good good signals + not noise; those can confuse price noise with price trend /maybe well confused,SuntraderLOL Coal miners daughter\ lived in a cabin on a hill\ in Butcher holler\ liked mail order catalogs money made from killing hawgs-Loretta Lynn =fossil fuel song >14,777,888 video views Hawg = hog+ holler means valley in hillbilly speak
If you can identify an algo being employed to slice a big buy or sell—like VWAP—the price level should generally hold. Example: JPM wants to buy 10000000 shares of AMD. They’ll not put the whole order out there because they’ll move the price against themselves. Instead they’ll run an algo to slice up the buy into smaller chunks. Not easy to perceive. Once in a while I spot one and make a dollar or two on it.
Price predicts it's own future, (when you know what to look for) but so does cyclic volume analysis, especially with crude oil. This gentleman's work is also still predictive.