That's the first time I've heard of anyone predicting signals.
It is a subtle point. Usually, it happens when the signal itself is ill defined. Things like: It is rising, or it is diverging. These type of "signals" are full of holes since they are not discrete but more on a continuum. Better is rising-diverging by XYZ relative to ABC, which can be discrete, which then can be put into Boolean logic, which therefore can be coded.
If you think that systems can be based on vague (ill defined, technically speaking) "continuums", then you are in the different school of thought, than what Alogs are based upon. Which can work, but a different "style" or "model" or to use the technical term, "ontology".
It is one of those clarifications that comes from studying logic, model theory and epistemology for years.
Bottom line is many things work, but mixing stuff can be dangerous when you do it to fit your perception, versus relying upon facts of the matter, i.e. reality.
But IMO, things happen so fast in Day trading, that just keeping up with the new data at regular intervals,
consistently, is enough of a chore to keep me busy. I can't deal with what is possible, only what has happened.