This source says that wave A in a zig zag must be impulsive? Ive heard a few EW analysts say this before but my experience tells me its not true. correctives literally have no motive they can produce some real curve balls.
Zig-zags are 5-3-5 (impulse-corrective-impulse) a corrective structure built from impulsive intent and lack the follow through to get that 5th wave.
I understand that Bitcoin throws out some real shit at times, but this is truth. ANY EW site will say the same about Zig-Zags being 5-3-5. When you have a 3-3-5 ABC where C is far longer than a flats' C allows for, It isn't a zig-zag just because it's not a flat. Bitcoin is heavily reactionary. Mainly because it's full of novice traders that don't know better. You can get a weak A (due to reluctance to sell) and the C will be far deeper than it should simply because of panic and herd mentality. This is mainly on intraday time frames. Larger picture counts don't have this phenomena because the moves are much larger so take much more volume to do the same.
What I mean about reactionary is that many traders don't know whats coming next. So their idea of a "trade plan" is to take the next big move, and follow. This is usually futile since it may or may not continue. They tend to be the C-wave panic crew or the power behind the 3 of 3.
I really hope this all makes sense from the psychology side of EW.
Actually I think EW works in any large liquid system on the assumption that traders haven't got a damn clue. buying at the top is irrational, and there's only one kind of irrational imo
certainly traders are not actively forming EW patterns as if that was the goal.
I think the analyst that sees all zig-zags must be 5-3-5 is overcomplicating things where there is a better explanation imo. If its difficult to identify 5-3-5 they might see instead (5-3-5)-x-(5-3-5) for the sake of that rule where reducing to that scale of fractal is not ideal. There are too many solutions to counting complex correctives in xyz formations. we could simply say (5-3-5)-x-(5-3-5) = 3-3-3.
Imo I think we just have to remember the roots of EW analysis. the fib sequence. We only use the two ratios from the sequence, the first two, 1/2 and 2/3. these are the strongest harmonic ratios that occur everywhere just as the golden ratio does, for example in music every note is derived from a single pitch and the ratios 1/2 and 2/3. All other ratios we use are actually just terms of these. We observe that movement in the direction of the trend are of 2/3 (3 up, 2 down) and corrective 1/2 (2 down, 1 up)
EDIT: of coarse my point being that 1/2 ratio leads to a 3 wave species and a 2/3 ratio leads to 5 wave species.
EDIT AGAIN: and I see that 5-3-5 woud actually correctly describe the direction of the trend within a correction but where a correction may be summarised to xyz-X-abc etc... practically we can consider the whole nature of a corrective as non impusive when we are not gifted impulsives.
That's why zig-zags are absolute in their definition and what should be expected through the duration. 
