Overview
A moving average (MA) is the average price over the last N periods, recalculated each bar. It smooths the noise of raw price into a single line that reveals the underlying trend — the most widely used indicator in all of technical analysis, and the foundation of trend-following.
Its job is simple: turn a jagged price series into something you can read at a glance, give you an objective definition of "the trend," and act as a moving line of support or resistance.
Origins & history
The moving average is an old statistical tool, but its application to markets has a clear pioneer: Richard Donchian.
- 1948Donchian founded Futures, Inc., one of the first publicly managed futures funds, built on moving-average rules.1
- 1960As Director of Commodity Research at Hayden Stone, he wrote the weekly "Commodity Trend Timing" letter around his famous 5- and 20-day moving-average crossover system.1
- LegacyHe became known as "the Father of Trend Following"; his mechanical-system ideas directly inspired traders like Ed Seykota and the entire trend-following industry that followed.1
How it works
There are two main kinds:
The SMA weights every period equally. It is the smoothest and the slowest.
The EMA weights recent prices more heavily, so it turns faster and lags less — at the cost of more false moves. Common lengths: the 20 (short-term), the 50 (intermediate), and the 200-day (long-term regime). Crossovers have names: a fast MA crossing above a slow one is a golden cross; crossing below is a death cross.
Market psychology & mechanics
A moving average has no magic of its own — but it becomes meaningful because so many participants watch the same ones. The 200-day in particular is a line that funds, algorithms, and the financial media all treat as the dividing line between "healthy" and "broken," which makes reactions around it partly self-fulfilling. More fundamentally, a rising average simply encodes a real fact: that the average buyer over the period is now in profit, which supports further buying on dips toward that average.
Honest assessment
Strengths
It is objective, simple, and universal: it converts "is this a trend?" from an opinion into a rule, works on any market and timeframe, and provides dynamic support/resistance and a clean trend filter. As a risk filter it has real historical value (below).
The academic evidence
The landmark test is Brock, Lakonishok & LeBaron (1992), which ran simple MA rules on the Dow Jones from 1897–1986 and found genuine support — a basic MA rule added roughly 3–4% per year over buy-and-hold, mostly by sidestepping large declines.2 Mebane Faber's later work showed a ~200-day (10-month) average used purely as a risk switch sharply reduced drawdowns while keeping most of the upside.3
Evidence rating: historically supportive — especially as a drawdown-reducing risk filter — but weaker and more regime-dependent in recent decades, and sensitive to costs and parameter choice. Not a standalone edge.
Weaknesses & failure modes
- LAGIt always lags. An average of past prices can only confirm a trend after it has begun — it never predicts. Crossovers are late by design.
- WHIPSAWIt fails in ranges. In sideways markets price oscillates across the average, generating a stream of false crossovers and losses. MAs only earn their keep in trends.
- CURVE-FITThe length is arbitrary. The "best" period is usually hindsight; an MA optimized on the past often disappoints live.
- DESCRIPTIVEIt describes, it doesn't forecast. The average has no information price doesn't already contain.
Professional uses vs. retail misuses
How professionals use it
- As a trend filter / regime switch (trade longs only above a rising 200-day).
- As dynamic support to time pullback entries (e.g., the 20-EMA pullback).
- Combined with structure and volume — never alone.
Common retail misuses
- Trading every crossover mechanically in a choppy market.
- Hunting for the "magic" length by optimizing on history.
- Treating a touch of the MA as a guaranteed bounce.
How two modern traders use them
Two traders profiled in the next Market Wizards — Kristjan Qullamaggie and prop trader Lance Breitstein — learned the markets independently yet both lean heavily on moving averages, treating them as a read of equilibrium and trend. Both stress the condition that matters: MAs only work in trending tickers, not chop — echoing the whipsaw warning above. Qullamaggie uses a simple regime filter (10-day above a rising 20-day = good environment) and "surfs" trends in a stair-step pattern, trailing the 10- or 20-day and selling only on a close below it. Breitstein, trading shorter intraday legs, often exits the moment price breaks the average, expecting follow-through. A telling detail on not over-optimizing: Qullamaggie once used the wrong MA setting by accident and it made almost no difference — the discipline matters more than the exact number.
WATCH Lance Breitstein — "Reacting to Kristjan Qullamaggie's Moving Average Trading Strategy"
Going deeper
Variations: weighted (WMA), Hull (HMA, reduced lag), volume-weighted (VWMA), and the VWAP — itself a volume-weighted average. Systems built on MAs: Donchian's 5-20 crossover, Elder's Triple Screen, and Grimes' point that the average isn't magic — the edge is a disciplined, consistent entry near an average, not a specific number.
Multi-timeframe: the same average means different things on different timeframes; the 200-day is a position-trader's line, the 20-period a day-trader's. Alignment across timeframes (price above the average on both) is stronger than either alone.
Practice
Observation exercise: put a 20-EMA and a 200-SMA on a chart and scroll history. Mark where price trended (the averages helped) versus where it ranged (the averages whipsawed). Then count how many crossover signals would have been profitable in each regime.
Quiz 1 — Why does a moving average always "lag" price?
Because it is an average of past prices — it can only reflect what has already happened. It confirms a trend after it starts; it never predicts.
Quiz 2 — What's the practical difference between an SMA and an EMA?
An EMA weights recent prices more, so it reacts faster and lags less — but gives more false signals. An SMA weights all periods equally, so it's smoother and slower. Speed vs. stability.
Quiz 3 — In what market does a moving-average crossover system perform worst?
A sideways / ranging market. Price keeps crossing back and forth over the average, producing repeated whipsaw losses. MAs only pay in trends.
This concept in the knowledge graph
Resources
- TRADERRichard Donchian — the father of MA-based trend following.
- PLAYBOOKHoly-Grail pullback & Donchian breakout.
- GLOSSARYMoving average, golden cross, death cross.
References (primary where possible)
- Richard Donchian — biography & the 5-20 system — Wikipedia.
- Brock, Lakonishok & LeBaron, "Simple Technical Trading Rules and the Stochastic Properties of Stock Returns," The Journal of Finance, 1992 — Wiley.
- Mebane Faber, "A Quantitative Approach to Tactical Asset Allocation" (the 200-day / 10-month risk filter) — SSRN.
Note: a 1999 follow-up by LeBaron found the historical MA edge had weakened with more recent data — included above in the honest evidence rating.