I am no Python expert, but the begin – the Real and Imag calculation – does not look right to me. Also Python might have some subtle differences that affect the result.

def CCYState(Data, Length, Threshold, previous_angle):
Real = correl(Data, Length, cosFunc)
Imag = correl(Data, Length, sinFunc)

# Compute the angle as an arctangent function and resolve
ambiguity
if Imag == 0:
current_angle = 0
else:
current_angle = 90 + math.degrees(np.arctan(Real/Imag))
if Imag > 0:
current_angle -= 180

# Do not allow the rate change of angle to go negative
if previous_angle – current_angle < 270 and current_angle <
previous_angle:
current_angle = previous_angle

# Compute market state
if abs(current_angle – previous_angle) < Threshold:
if current_angle = 0:
state = 1
else:
state = 0

return state, current_angle

C.

I have just converted his code without giving it much consideration. But you are right, from math in school I remember that a phase angle is atan(sin/cos), not atan(cos/sin) as in the code. But swapping sin and cos is a 90 degrees rotation and I think that’s why Ehlers adds 90 to the result. The end result is anyway based on angle difference, not on absolute angle. Maybe Ehlers can explain why he has calculated it in this way.

In both your code above and the Ehlers paper the arctangent is described in terms of a the ratio of real to imaginary. Isn’t this, in fact, the reciprocal of the arctangent? If we think of the phasor diagram with the Imaginary axis being the ‘y-axis’ and the real axis being the ‘x-axis’, then tan of the phase angle would be y/x.
In addition, this definition is at odds with Ehler’s own ‘rocket science of traders’ book where he uses the same process to derive the phase angle via the Hilbert Transform.