Moving Averages
What a moving average actually smooths, the tradeoff between responsiveness and lag, and how crossover signals work.
A moving average is introduced simply as the average price over a fixed trailing window — a 50-day moving average is the average closing price of the last 50 days, recalculated fresh each day as the window rolls forward — and its purpose is to smooth out short-term noise so the underlying trend direction is easier to read than from raw daily price alone. Murphy walks through the simple moving average and the exponential moving average as the two main variants, the key difference being that an exponential average weights recent prices more heavily, making it react faster to new information at the cost of also reacting faster to noise that turns out not to matter.
This produces the central tradeoff the whole chapter is organized around: a shorter average (fewer days in the window) is more responsive to genuine trend changes but also generates more false signals from short-term noise; a longer average is smoother and filters out more noise but lags further behind a genuine trend change, giving back more profit before it confirms the change has happened. There is no universally correct length — the right choice depends on the trader's own timeframe and how much lag versus false-signal risk they're willing to accept.
Recalculated each period as the trailing window rolls forward one day and drops the oldest day from the average.
The book's most common practical application is the crossover signal between two averages of different lengths — a shorter average crossing above a longer one is read as a bullish signal, crossing below as bearish. Murphy is candid that this class of signal is inherently a lagging one: because both lines are built from past prices, a crossover can only occur after the underlying price trend has already shifted, meaning a moving-average system will reliably miss the very start of a new trend and give back some profit at the end of an old one before confirming it has ended. The tradeoff is deliberate: what the system gives up in early entry and exit timing, it gains in filtering out the whipsaws that a trader reacting to every raw price wiggle would suffer instead.
A classic version cited in the book pairs a 50-day and 200-day moving average, where the shorter crossing above the longer is popularly known as a "golden cross" and the reverse as a "death cross" — widely followed in practice specifically because so many other market participants also watch it, which can itself become a partial, self-reinforcing reason it works.
- A moving average smooths price to make the underlying trend easier to read; shorter averages are more responsive but noisier, longer averages are smoother but lag further behind genuine trend changes.
- Exponential moving averages weight recent prices more heavily than simple moving averages, trading some smoothness for faster reaction to new information.
- Crossover signals between two averages are inherently lagging — they can only fire after price has already moved, which is a deliberate tradeoff against whipsaw noise, not a flaw to be eliminated.