Moving Averages Explained: What They Really Tell You
A moving average is the average price over the last N bars, recalculated on every new bar. That is the whole formula, and it explains both the value and the limit: it summarises where price has already been and says nothing about where it is going. Every honest use starts by accepting that it lags by design, and every expensive mistake starts by treating the line as a prediction.
SMA versus EMA, and the periods that matter
A simple moving average (SMA) weights every close in its window equally: a 50-day SMA is the last 50 closes divided by 50. An exponential moving average (EMA) weights recent closes more heavily, using a factor of 2/(N+1), so a 20-period EMA puts about 9.5% of its weight on the newest bar alone. The EMA therefore turns sooner, giving earlier signals and more false ones; the SMA turns later and filters more noise. Neither is better, and testing on your instrument and timeframe settles it.
Period length matters far more. Three settings dominate because so many participants watch them: 20 (about a month of trading days, the short swing rhythm), 50 (a quarter, the intermediate trend) and 200 (a year, the line funds and financial media treat as the bull-bear divide). Their popularity is part of why they react, and also the limit: reference points, not laws. An SMA also lags roughly half its window, so a 200-day average describes conditions from about 100 days ago.
The three honest uses
1. A trend filter. A moving average's most valuable job is cutting the number of trades you are allowed to take. A rule such as no longs while price sits under a falling 200-day removes a whole category: counter-trend buying in a downtrend, where most beginner accounts bleed. The filter never says when to enter, only which half of your ideas to throw away.
2. A dynamic level. In a sustained trend, pullbacks often stall near the 20 or the 50. Treat that as a zone a percent or two wide where you wait for a real entry trigger, rather than buying the touch. Set the stop from the structure of the move, not from the average, which drifts toward price every bar and will follow you into a loss.
3. A mechanical crossover system. Two averages, one fast and one slow, give complete testable rules: long when the 50 crosses above the 200, flat or short when it crosses below. Golden cross and death cross are the popular names for it. The payoff is classic trend following - a win rate near 35-45%, many small losses and a few very large winners that carry the year. Taking every signal is not optional: skipping the entry after three losses is the standard way to miss the trade that pays for the rest.
The trap: crossovers in a sideways market
A crossover system follows trends, and most markets spend most of their time not trending. In a range, price oscillates across its own average, so the fast line keeps cutting the slow one. A 20/50 pair on a daily chart can fire eight to twelve crossovers in one flat quarter, each entering late near the range high and exiting late near the low. That is whipsaw, and it drains an account quietly while every single loss looks small.
No setting fixes this, because it is a regime problem, not a parameter problem. Read the slope as well as the position - a flat 200-day means there is no trend to follow - and treat two averages that coil together and touch repeatedly as a market with no direction. Slower pairs whipsaw less but lag more: a trade-off, not a cure.
So discipline decides more than parameters. Tag each moving-average trade in your journal with the regime you thought you were in and the risk you took; after fifty trades the numbers say whether losses come from bad rules or from good rules used in the wrong market.
Indikora is built for that verification: a paper-trading simulator for running a crossover rule set across dozens of trades before it costs anything, a coach that holds you to the filter you wrote down, and a journal that separates trending results from range-bound ones.
Frequently asked questions
Which moving average is best for beginners? The 200-day, used as a direction filter rather than an entry signal. It is slow enough to be hard to misread, and keeping you out of counter-trend trades saves the most money early.
Is the EMA better than the SMA? No, it is faster: earlier entries in real trends, more false signals in noisy ones. Pick one, test it on your instrument and timeframe, and stop switching after every losing week.
Do moving averages work in crypto? The same way they work anywhere: as trend filters and reference zones, not forecasts. Crypto volatility produces more whipsaw, so slower settings or a higher timeframe are common.
Indikora is an AI-powered trading coach for crypto, forex, gold and indices - with a paper-trading simulator for testing moving-average rules, a coach that enforces the filters you set, and a journal showing which conditions your system actually survives. Try it free: https://indikora.com
This article is for educational purposes only and is not financial advice.