Tuesday, May 17, 2011

The Difference ADX Makes When Examining Reversals of Breakdowns

In the past I’ve discussed how breaks from consolidations are often much less reliable when looking for a reversal than if you hit a new high or low during a trending market. This is because the breakout will often create new excitement in the direction it occurs. Some of that is stops being blown and some of that is from breakout players looking to take advantage of a newly emerging trend.

One way to measure the strength of a trend is with the ADX indicator. It was pointed out to me today by a subscriber that the SPX 14-day ADX was dropping under 12 – which is an extremely low level. (The 14-period reading is the default reading for ADX with most charting packages.)

So using ADX as a measurement of trend strength I created a study that demonstrated low-ADX vs. high-ADX breakdowns, and how the implications were much different. First let’s look at situations like the present where the SPX breaks down and ADX is under 20.

As you can see, there is no substantial edge suggested when you are looking at a sharp move out of a consolidation with a low ADX.

Next I ran a 2nd test. This one used all the same parameters except I required the ADX to close above 20 instead of below it. This suggests the market has been moving instead of meandering. Those results are below.

What we see here are vastly more compelling results.
Unfortunately, with the ADX now just under 12, there does not appear to be a substantial upside edge offered by the break lower.


Anonymous said...

From today’s post by Dave Steckler, ETFRoundup.com.. a perfect complement to your post, Rob. I’ve summarized a bit.. the quotes are just to indicate that I'm recapping someone else’s thesis here..

“ADX measures the strength – not the direction – of a trend. When the ADX is below 20, and especially when it’s below 15 (which was the case with our three ETFs--DIA, SPY, and QQQ--both in April and now), the security is not trending… and moving averages are less effective in holding as support or resistance.

“When the Qs fell below their 50-day SMA, on April 12, the ADX was 14.87. When the SPY closed below the 50-MA on April 13th, the ADX was 14.95. The ETF’s were trading in a range, not trending. And the lows established the bottom of the range.

“Back in March, when the ETF’s fell through the 50-day SMA and kept falling, note that their ADX’s were above 15-- and had started rising. As the downtrend became more pronounced, the ADX rose. Right now, with the low ADX, and the past history of these ETF’s bouncing higher within a few days of crossing below the MA (when that’s the case), there seems a reasonably good probability it will happen again.”

Miles Hoffman said...

If you (and I) understand the ADX correctly, the level of the ADX is generally NOT important, per se. It is the direction that is important, at least if the guy in the following link is correct:


You'll see this guy (LeBeau) believes Wells Wilder got us off to a bad start by saying the ADX level is meaningful. I would urge you to study Lebeau's use of the "delta" of the ADX as a signal, that is, what is important is the size of the CHANGE in ADX that is important. The CHANGES in the ADX have recently been "small", so the ADX is NOT telling us anything right now.

The level of the ADX is of MORE importance IF the ADX is above 40, which means the trend (be it up or down) is "long in tooth" and is at risk of ending (and this typically occurs when the CHANGE in the ADX has been consistently big).

However, the end of one trend (direction) does NOT mean the start of a trend in the opposite direction (and this is seemingly where we are right now). Quite the contrary, the end of a trend is usually a consolidation and eventual breakdown (OR resumption of trend), but NOT a sudden reversal.

This means when the ADX is rising, and the general usage is it breaks ABOVE 20 from below it, then a trend is starting and you should use trend following techniques to "place your bets". When the ADX begins to "break down", you should move away from trend-following techniques and concentrate on oscillators.

I combine the ADX with lessor-known trend indicators of the "Aroon Indicator" and the "Choppiness Indicator" for break out or break down and when all three agree, then it's a pretty good signal.

Currently, the Aroon Indicator is still bullish but weakening; however, the Aroon Indicator also has a bearish indicator, and it "is not speaking". The Choppiness Indicator is saying choppy.

So as I interpret your study, it really is only confirming the "definition" of ADX. And by definition, the level being around 12 means ABSOLUTELY NOTHING.

I'm sure I'm NOT conveying my interpretation succintly, but it's like the ADX is a red light. Your study is pulling out a color chart, studying the color wavelengths, and then telling us the red light is, well, red.

I am trying to be constructive here so please don't take me wrong. You do very interesting studies and I would LOVE for you to look at LeBeau and "quantify his methodology", especially since I've moved away from but once was a "quant".

I'm probably stupid for NOT quantifying LeBeau's methodology myself, for his logic appears very valid, at least from the visual standpoint. I've created technical indicators based on them, but like a quant, I'm always seeking "the best", so thanks for pointing me back to the CHANGE in ADX, which I have been overlooking.

Unfortunately, like the ADX, the CHG in ADX as yet is not telling us anything, except it's TOO early to be looking for a trend.

Miles Hoffman, CFA

Ron C said...

Great post. Personally, the most important component is price action as the indicators are lagging and are all dependant on the movement of price.