This might also be a documentation issue -- either way, it appears to me that the rSign variable generated by ZigZag() has the wrong sign. From the online manual for ZigZag():



Unless I misinterpret this, rSlope and rSign should always have the same sign; however, they have opposite signs (in Zorro S 2.25), and rSign appears to contain -1 for a prior top, and 1 for a prior peak. rSlope appears correct. I assume this could be fixed by changing either the documentation (for backwards compatibility) or the implementation...

I would interpret it so that rSlope and rSign have a different sign. If the last peak was a top, then rSign is 1, and rSlope is negative because the slope goes downwards from a top.

Well, I understand one could interpret it that way -- only, that's not what rSign actually does :-)



Yields the attached result. It appears that as soon as rSign recognizes that a new leg is being formed, it assumes -1 for a prior top and 1 for a prior bottom (as seen by the length of the legs in the image, and by the last value for rSign)...

In any case. Not a big deal, one just has to be aware of it either way.

Ok, we'll check what rSign really does and will make the documentation more clear.

Thank you!