Moving averages, explained

What a moving average is

A moving average is the average of the last few closing prices, recalculated every day. A 5-day moving average adds up the last five closes and divides by five. Tomorrow it drops the oldest day and adds the newest, so the average "moves" along with the price.

Say a stock closed at 10, 11, 12, 11 and 13 over five days. The 5-day average is (10 + 11 + 12 + 11 + 13) ÷ 5 = 11.4. If it closes at 14 the next day, the 10 drops out and the 14 comes in: (11 + 12 + 11 + 13 + 14) ÷ 5 = 12.2.

The point is to smooth out the noise. Daily prices zig-zag; the average shows which way they're heading underneath. When the price is above a rising average, the trend is up. When it's below a falling one, the trend is down.

Simple versus exponential

A simple moving average (SMA) treats every day equally. In a 21-day SMA, each close counts for 1/21, about 4.8%.

An exponential moving average (EMA) gives the most recent days more weight. In a 21-day EMA, today's close counts for about 9%, yesterday's a little less, and so on back in time. That makes an EMA react faster when the price turns.

There are more exotic versions too. Alan Hull's average combines weighted averages to cut lag further; Perry Kaufman's adaptive average speeds up when the price moves cleanly and slows down in chop; Patrick Mulloy's triple EMA strips out most of the lag of an ordinary EMA. Each is a different answer to the same trade-off.

Neither is better. Faster averages give earlier signals and more false alarms. Slower ones give fewer, later signals. Many of the plays on Be The Puck are really a choice about where to sit on that trade-off.

Why they always lag

An average is built from past prices, so it can only confirm a turn after it has happened. A 200-day average might take weeks to turn down after a peak. That lag is the cost of filtering out noise, and it's why trend plays never sell at the top or buy at the bottom. They aim to catch the middle of big moves.

Longer averages lag more. Some common lengths, roughly translated: a 10-month average is about the same as a 200-day one; a 30-week average is about 150 days; a 20-week average is about 100 days.

Bands and clouds

Some plays use two averages and shade the space between them, called a band or a cloud. The PB EMA uses two EMAs; Ben Cowen's bull market support band pairs a 20-week SMA with a 21-week EMA; Ripster's cloud uses the 34- and 50-day EMAs.

The band works like a buffer. Price above the whole band reads as an uptrend, below it as a downtrend. Price inside the band is undecided, so band plays hold their last call until price closes clearly on one side. That one choice cuts out a lot of false signals.

Crossovers

A crossover is the moment one line crosses another: the price crossing above its average, or a faster average crossing a slower one. The best-known is the golden cross, when the 50-day average rises above the 200-day, and its opposite, the death cross. Ed Seykota's teaching system is a crossover of two exponential averages, 15 and 150 days.