I Am So So Frog

I have seen some of you claim for a pre-cup on FROG 15m! Indeed It hasn't been unnoticed.

I shall remind you that a high degree of variation around a "cup" lowers its validity, what might you say to this?

Assuming we have a model y representing our underlying trend, the degree of variation around our fitted model can be given by the sum of squared error (SSE):

se = (price[t] - y[t]) ^ 2
sse = SUM(se)

A cup alone makes the assumption of a quadratic underlying trend (that is if you remove cyclical and noisy variations from the price you end up with a curve looking like a parabola) given by:

price(t) = at^2 + bt + c + cycle(t) + noise(t)

where at^2 + bt + c is the quadratic trend, where a, b, and c are constants.

If the amplitude of the cycle(t) and noise(t) components are high then variability around the trend will be higher.



You might see cups everywhere, and that's cool, it's very open-minded of you, and you might even be right, however, trends buried in noise are not as pertinent if you truly believe they play a key role in your analysis.