Moving averages explained: SMA, EMA and how prop traders use them

A moving average is one of the simplest tools on a futures chart, and one of the easiest to misuse. Traders running several funded NQ or ES accounts often treat a favourite period as a signal, then discover in a choppy New York open that the same line that filtered trend last week is stopping them out of every scratch today.

This guide covers what a simple moving average (SMA) and an exponential moving average (EMA) actually are, how they behave, the ways prop traders usually apply them, and the pitfalls that matter when a daily loss limit can end the account. Numbers below are worked examples for illustration, not a strategy to copy.

What a moving average is#

A moving average smooths recent prices into a single line. Each new bar drops the oldest price in the window and adds the newest one, so the line "moves" with the market.

Traders use it for three jobs:

1. Describe the recent average price, so you can see whether price is above or below that average.
2. Filter direction, for example only take longs when price is above a longer average.
3. Mark a dynamic level that can act like support or resistance while a trend is intact.

It does not predict the next bar. It summarises the ones you already have.

Simple moving average (SMA)#

An SMA is the arithmetic mean of the last N closing prices.

SMA = (P1 + P2 + … + PN) ÷ N

Example (illustrative): a 5-period SMA on five closes of 100, 102, 101, 103 and 104:

(100 + 102 + 101 + 103 + 104) ÷ 5 = 102

If the next close is 106, the oldest close (100) drops out:

(102 + 101 + 103 + 104 + 106) ÷ 5 = 103.2

Every price in the window has the same weight. A spike ten bars ago pulls the SMA just as hard as yesterday's close, until that spike ages out of the window.

Exponential moving average (EMA)#

An EMA gives more weight to recent closes, so it reacts faster when price turns.

The usual recursive form uses a smoothing factor α based on the period N:

α = 2 ÷ (N + 1)

EMA_today = α × Price_today + (1 − α) × EMA_yesterday

Example (illustrative): a 5-period EMA, so α = 2 ÷ 6 ≈ 0.333.

Suppose yesterday's EMA was 102, and today's close is 106:

EMA_today = 0.333 × 106 + (1 − 0.333) × 102
= 35.3 + 68.0
= 103.3

Compared with the SMA example above, the same new close of 106 lifts the EMA a touch more in one step, because more of the weight sits on today. Older prices still matter, but their influence fades instead of falling off a cliff when they leave the window.

Seed the first EMA with an SMA of the first N closes, then apply the recursive formula from there. Charting platforms do this for you; the important part is that two platforms can disagree slightly if they seed differently.

How SMA and EMA behave#

Neither is "better". The trade-off is always the same: responsiveness versus noise. A short EMA hugs price and flips often. A long SMA ignores noise and lags a real turn by more bars.

Lag, in plain numbers#

Example (illustrative): price has been flat at 100, then prints five closes of 110.

• A 10-period SMA still has five old 100s in the window, so it only rises halfway toward 110 on that fifth bar.
• A 10-period EMA, with more weight on the new closes, sits closer to 110 by then.

That lag is useful when you want to stay with a trend and ignore scrapes against the average. It hurts when you use a slow average as an exit and give back a large part of an open profit before the line turns.

Common uses on a prop desk#

1. Trend filter#

Trade only in the direction of a longer average. For example, on a 5-minute NQ chart during the New York session, some traders only look for longs when price and a faster average are both above a 200-period SMA on that same timeframe, or when a higher timeframe (15-minute or 1-hour) average is sloping up.

The filter does not pick the entry. It vetoes trades against the larger drift.

2. Dynamic support and resistance#

In a clean trend, pullbacks often stall near a widely watched average (for example a 21 EMA or 50 SMA on the trading timeframe). That is observation, not a law. In a range, price will cross the same average repeatedly, and treating every touch as support is how a trend tool becomes a whipsaw machine.

3. Crossovers#

A classic setup is a faster average crossing a slower one (for example 9 EMA over 21 EMA). A cross after a long trend can mark a regime change. A cross inside a tight range usually means noise.

Example (illustrative): on a quiet overnight ES range, a 9/21 EMA cross can fire three times in two hours with nothing but small scratches. The same cross after a multi-day trend on the 15-minute chart is a different event. Context first, cross second.

4. Multi-timeframe alignment#

A common prop-trading habit: read direction on a higher timeframe average, execute on a lower one. For example, only take New York continuation longs on the 1-minute or 3-minute chart when the 15-minute 50 EMA is rising and price is above it. That keeps the lower timeframe from talking you into fades against the session drift.

Pitfalls that matter on funded accounts#

Lag cuts both ways#

By the time a slow average confirms a turn, a large part of the move can be gone, and a tight daily loss limit does not care that your filter was "right" on a higher timeframe. If you need speed, shorten the period or use an EMA knowing you will accept more false flips.

Chop kills crossover systems#

Moving-average crossovers are trend tools. In balance (lunch hour, post-FOMC coil, overnight thin liquidity) they will overtrade. A simple session filter, or standing aside when average true range compresses, usually matters more than optimising 8 vs 9 vs 10.

Curve-fitting the period#

Any market has some period that looked perfect last month. Fitting NQ's last three weeks to a 17 EMA and then hard-coding that into every funded account is how a backtest turns into three daily-limit hits in a week. Prefer a small set of standard periods, test them across regimes, and keep the rule boring.

An average is not a stop#

Parking a stop "just under the 21 EMA" without a structure-based invalidation is common and fragile. One impulsive bar through the average can tag every copy of that stop across your accounts at once. Place the stop where the trade idea is wrong; use the average as context, not as the idea itself.

A trend filter does not replace firm rules#

Staying on the "right side" of a 200 SMA does not add room under a trailing drawdown or a daily loss limit. Size from room left, not from how neat the chart looks. Our guides on daily loss limit maths and trailing drawdown cover that arithmetic.

A practical way to test before you risk a funded account#

Before you rely on an SMA or EMA rule live:

1. Write the rule in one sentence. Example: "On 5-minute NQ, only long if price is above the 50 SMA and the 9 EMA has crossed above the 21 EMA after 9:30 New York."
2. Define the exit and the stop without reference to hope: where the idea is wrong, and how you take profit.
3. Backtest across different weeks, including a trend week and a chop week, not only the week that inspired the idea.
4. Check the result in R, after costs, against a daily loss cap like the one your firm uses.
5. If you run several accounts, apply the same rule and the same per-account risk, rather than improvising size on the second and third copies.

Tradesurge can backtest a rule like that against your own idea and help you execute it across prop accounts with each account's risk limits in view. It will not turn a noisy crossover into an edge; it will show you what the rule actually did, and stop one account's mistake from becoming every account's breach when you go live. See Breach Guard for how per-account limits are enforced.

FAQ#

Should I use SMA or EMA?#

Use SMA when you want a stable read of the average price over a window. Use EMA when you want the line to react faster to recent closes. Many traders use a longer SMA as a filter and a shorter EMA for timing. Test both; do not switch mid-week because one looked better yesterday.

What periods should I use on NQ or ES?#

There is no universal answer. Common starting points people discuss are 9, 21, 50 and 200 on the timeframe they trade. The useful question is whether your period matches the timescale of your setup, not whether a forum prefers 20 or 21.

Do moving averages work in the New York session?#

They describe price in every session. They work as trend filters when New York is trending, and they chop when New York is balancing. Session context matters more than the indicator.

Can I use a moving average as my only strategy?#

You can, but most funded traders who last treat it as one filter among others (level, session, risk), not as a standalone signal generator. A single cross with no risk plan is how accounts die quietly.

Do prop firms care which moving average I use?#

Firms care about loss limits, consistency rules and banned strategies, not about your indicator. Your chart can be perfect and the account can still fail on risk. Read your rulebook.

---

This article is for education only and is not financial advice. Trading futures and trading with prop firms involve risk of loss. Examples are illustrative arithmetic, not results. Always check your firm's current rulebook.

Cover photo by Vitalii Skachko on Unsplash.