This is an article Todd prepared a few months ago on identifying divergence in technical analysis. I’m glad to share it here and hope that it is helpful.
____________________________________
When looking for divergence between price and an indicator (e.g.,
MACD, Stochastic, CCI, RSI) what we are looking for first is for the price to match or exceed the previous pivot high or low. (Any time MACD is mentioned, other oscillators can be substituted.)
Rule #1: If we have not established through price action either a new price extreme (e.g., a peak that is higher than the previous high or a valley that is lower than the previous low) or a double top or bottom (matching the previous price but not exceeding it) then there is no divergence, period, end of story. Price MUST meet this requirement before indicators can even be examined for the possibility of divergence.
It is the easiest if we examine just successive pivot highs or lows and only two extremes at a time. The pair must include the peak or valley representing the current price action. In other words, when price sets a new high, proceeds to do its normal pullback from that high, and then moves again to meet or surpass the earlier high, this is the time to look for divergence. In downward motion, the price must set a new low, pull up from that low, and then move down to the earlier level or lower. Only in these conditions is the first prerequisite met. Failure to meet or exceed the previous extreme invalidates searching for divergence.
Rule #2: When price has fulfilled #1, then draw the line from the recent price extreme backward to the price level that had previously set the high or low — the one that current price action had to match or blow through to qualify for Rule #1. This only works on successive major peaks/valleys — any little bumps or hills that price action went through between setting the extreme points are irrelevant to the purpose of this discussion.
Rule #3: Unless price action actually creates successive peaks/valleys instead of simply consolidating along a ceiling or a floor, looking for divergence is ineffective.
At such times, any curves seen in MACD or similar indicators are simply an indication that momentum has slowed down. They are not an illustration of reliable divergence.
Until a new peak/valley is formed, price can only do one of two things: consolidate in a range or reverse and pull back away from that resistance/support, perhaps to try again later to establish a new extreme. Only in the second scenario does the search for divergence provide any credible indication of imminent reversal.
Rule #4: When qualified peaks are established, then we are connecting the TOPS of those two price peaks. If valleys are set, then we connect the BOTTOMS of those valleys. It’s an easy mistake to make to be drawing the “price-slope” line on the wrong side of the price action.
If we have established that price action has fulfilled these first requirements, then we can look at MACD and compare it to price.
Rule #5: Whether one’s MACD contains one or two lines, we are examining the extreme points of the CURVES of the MACD’s movement, not, in the case of MACD indicators with 2 lines, the relationship of one line to the other. This is another very common mistake.
Read the rest of this entry »
