To Light a Candle
By Gordon Philips, Director of Trading and Research
Institute of Higher Earning
Let us insert our tongues firmly into our cheeks and take a little trip out
to the wood shed where we shall administer some necessary and long overdue
medicine to a most revered technical indicator.
No doubt you've heard the expression 'tough love.' Here we go a step farther
and deliver a technical trading reality slap!
Due to our fond affection for the ephemeral pip we will be directing our
remarks to Forex traders, however these comments would apply equally well to
traders of other markets, be it baseball cards, beanie babies or stocks.
We're here to skewer a sacred cow; that holiest of holy technical
indicators, brought down from Mystery Mountain on the same tablet as the
slow stochastic, but that's an article for another time.
We're speaking of the venerable Moving Average Convergence Divergence or
simply MACD.
Enough articles have been written about the MACD to repopulate half the
world's rain forests. Perhaps this will be the last. Although given human
nature we'd have to forecast (without using the MACD) that more trees will
needlessly die.
Here it is with both barrels: the MACD doesn't work. It's a glorified (if
more complex) moving average, and moving averages are pathetic in
forecasting price direction. They're great for telling you where, how and
when price has already been -- but that's in the past, and there's a reason
why they call this futures and not histories.
When applied over a long enough time frame a given moving average will be
right half the time and wrong the other: the very definition of randomness.
Jimmy the Greek would take a pass.
It's true that when paired with a second moving average of differing and
reasonable periodicity (3:1, 4:1, 5:1, etc.) it is possible to maintain a
modest positive trading expectancy over a sufficient number of trades, but a
pair of endlessly crisscrossing moving averages will 'get you in' too late
and just as often 'get you out' even later the vast majority of the time.
The MACD is a seductive indicator: by eschewing the simple moving average in
favor of the exponential it is just advanced enough to entice the delight of
the technically-minded, but not so arcane as to perplex the most
propeller-driven beanied amongst us.
As for sex appeal, it's got two colorfully wavy lines crossing back and
forth over each other (that's to keep you hypnotized), plus this really cool
histogram in the middle for an added air of sophistication. Of course, you
could program charting software to reprint the physical separation between
any two moving lines, but so far only the MACD holds that claim, and it's an
attribute that fascinates.
But the true allure of the MACD is that it requires -- ta, da! --
interpretation. This fact alone has guaranteed the MACD perpetual placement
in the pantheon of trading studies. For once a thing can be interpreted, it
can be elevated to mystical status. From there it's just a short trip to
worship which is where the MACD finds itself today.
MACD apologists would argue that we're being too harsh. That looking back
there are countless times when the venerable work horse presaged a
profitable move. No doubt. Of course that would also be true of tea leaves,
drifting clouds or a spinning Coke bottle. Even a stopped clock is correct
730 times a year.
Alas, when used as an entry signal to the lush pip-filled world lurking just
beyond the hard right edge of our computer screen, the MACD fails time and
again as one would expect from any mathematically AVERAGED smoothing of PAST
price action.
If only one could trade in reverse!
Those inclined to count how many pips can fit on the head of a pin might
wish to tweak their MACD in an attempt to make it even more 'predictive.' To
our way of thinking this is like feeding a variety of performance enhancing
drugs to a snail and watching to see if its meandering trail bends just a
little more to our liking. But would it help us to know where the snail is
going next? No, we won't have a clue, and the snail probably won't either.
The MACD uses as its farthest backward looking value 26 periods. On a 15
minute chart, 26 periods total 6 hours and 30 minutes. Fifteen minutes goes
by pretty fast. There are 96 such periods in a 24 hour day, 480 in a 5 day
trading week. A single period of 15 minutes is really just noise in the
larger scheme of things.
Let's say it's 8:00AM New York time and I'm sitting here watching the 15
minute chart. What is the MACD doing? Still paying attention to what
happened back at 1:30AM when New York traders were sawing wood.
I'm sitting here trying to get pips based on where I think the market is
going in the next 15 or 30 minutes. I'm trying to get pips RIGHT NOW. What
do I care where price was trading SIX AND A HALF HOURS AGO?
My favorite pair may have traded above the daily pivot during the Asian
session, taken a bounce off strong long term resistance on the London open
and dropped like a rock right into the New York session, leaving the MACD
looking like a drunken hula dancer trying to outrun a volcano.
But I'm trying to divine what the Big Boyz are likely to do in the next 15
to 30 minutes, which means I need to understand what they were thinking 15
or 30 minutes ago, not in the middle of last night!
What I need is a predictive, laser-guided aiming system. What I get instead
is a meandering lava flow with a bunch of crazy people trying to stay one
step ahead. (Fortunately, most of those sprinters are serial pip donors, so
I encourage their efforts and offer them a glass of water as they run by!)
In the end, the performance of moving averages and indicators based on
moving averages will always be, well, average. So let's move on. What can we
use to tell where price is going next?
Hah, I caught you. The answer is: nothing. There is nothing in existence
that can tell us where price is going next. But there are a few simple tools
that can tell us where price is likely to go next, perhaps 80% of the time
or better.
They are:
1. The trusty trend line.
2. That freshest of all price footprints in the sand (or snow, depending on
where you live): the candlestick.
These tools are not sophisticated and require no divinity degree to
'interpret.' In fact, they are not subject to interpretation. They are that
stupid. They just sit there and look at you. And you look back, and you know
what to do.
Don't believe me? [Using a demo account] try removing all indicators except
for support and resistance (trend) lines and candlesticks. Here is a short
list of what you do NOT need:
1. Moving averages
2. Stochastics
3. Parabolic SARS
4. Bollinger Bands
5. Chaos fractal alligator
6. Ichimoku ginko biloba
7. Fibo-tunnel price-ray vectors
We have learned to recognize that price is a fickle thing that can change
its mind faster than Liz dumped Dick; that price moves like the people who
move it. That like Sherlock Holmes who could walk into a room and know
instantly how the killer had left.
(Answer: Through the window. The candle on the table next to the window had
been burning. A breeze had been blowing into the room. All of the dried wax
had dripped down the candle opposite the window. The window had been open
and that's how the killer had escaped).
First and foremost, learn to recognize candlestick action. We can look at a
candlestick and instantly see telltale marks of reversal and rejection
(fear!), or momentum and continuation (greed!) that drive markets.
When candlestick reversals occur at places where fear and greed occurred
before (tilted support or resistance = trend lines), or at anticipated price
levels (horizontal support or resistance = pivot points), we have an
extraordinarily high probability trade.
Price zigs ... a hammer at key support!
Price zags ... an evening star formation following a tweezer top at key
resistance!
And the MACD?
Still meandering through the charts like a picturesque mountain stream,
'weaving through the canyons of your mind.'
