Moving averages

A moving average is the average closing price over the last N periods, recalculated each bar. That is the entire mathematics.

Everything else attributed to it — support, resistance, trend signals, golden crosses — is interpretation layered on top of a smoothed line, and some of that interpretation survives scrutiny while most does not.

What it actually does#

It removes noise to reveal direction. Nothing more.

Price is jagged, and the jaggedness is mostly meaningless. An average smooths it so the underlying drift becomes visible. That is genuinely useful, and it is the honest description of the tool.

The cost is lag. An average is by construction a summary of the past, so it always turns after price does. A 50-period average cannot tell you about a change that happened three bars ago — it is still averaging in the forty-seven bars before it.

Every property of a moving average is a trade-off between smoothness and lag. Shorter is more responsive and noisier. Longer is cleaner and slower. There is no setting that escapes this.

Simple versus exponential#

Simple (SMA) weights every period equally. The price from 50 days ago counts exactly as much as yesterday's.

Exponential (EMA) weights recent periods more heavily, so it turns sooner.

The practical difference is smaller than the debate around it. EMAs respond faster and whipsaw more; SMAs are steadier and later. Neither is superior — they sit at different points on the same smoothness-versus-lag line.

The one job they do well#

Defining the trend context.

Ask a single question: is price above or below the average, and is the average rising or falling?

• Price above a rising average → uptrend context
• Price below a falling average → downtrend context
• Price crossing back and forth around a flat average → range, and no trend approach will work

That third case is the valuable one. A flat, repeatedly-crossed moving average is the clearest available signal that you are in a range — and knowing which regime you are in matters more than any entry technique.

Why crossovers disappoint#

The golden cross (50 above 200) and death cross (50 below 200) are the most publicised signals in technical analysis, and they perform modestly at best.

The reason is structural. A crossover of two lagging indicators is doubly lagging. By the time a 50-period average crosses a 200-period one, the move has usually been underway for weeks.

They are not useless — they describe regime shifts reasonably well after the fact. But as entry triggers they arrive late and produce frequent false signals in ranging markets, which is where markets spend most of their time.

Reasonable uses#

Trend filter. Only take long setups while price is above a rising average. This one habit removes a large share of low-quality trades by keeping you out of counter-trend positions.

Slope as momentum. How steeply the average rises tells you something about the strength of the drift. Flattening is an early warning that a trend is tiring — often before structure breaks.

Distance from the average. Price far above its average is stretched and tends to revert. This is a real tendency and a terrible timing signal — stretched markets stay stretched.

Regime detection. Flat average, repeated crossings, no trend. Covered above and the most useful of the four.

For F&O specifically#

Two adjustments worth making.

Averages are for context, not entries. With a contract that expires, the lag matters far more than it does for someone holding equity. A signal that arrives a week late may arrive after your option is worthless.

Use them on the higher timeframe. A daily 50-period average as context for an intraday trade is reasonable. A 50-period average on a 5-minute chart is mostly noise dressed as a signal.

Check yourself

0 of 4 answered

1. 1.What is the unavoidable trade-off in choosing a moving average period?

   Accuracy versus costSmoothness versus lag — shorter is responsive and noisy, longer is clean and slowSimple versus exponentialDaily versus intraday

2. 2.Why do golden and death crosses perform modestly as entry signals?

   They are calculated incorrectly by most platformsA crossover of two lagging averages is doubly lagging, so the move is usually weeks old by the time it firesThey only work on individual stocksThey require volume confirmation to be valid

3. 3.Why is 'the 50-day average is acting as support' a weak claim?

   The 50-day is too short to matterAn average is a different price every day — no orders rest there and nobody remembers it, unlike a real levelSupport only exists on weekly chartsIt is only valid for exponential averages

4. 4.What is the most useful thing a flat, repeatedly-crossed moving average tells you?

   A breakout is imminentYou are in a range, so trend-following approaches will not workThe period setting is wrongVolatility is about to collapse

What to take away

• A moving average smooths noise to show direction. That is all it does.
• Every period is a smoothness-versus-lag trade-off. There is no optimal setting.
• SMA vs EMA matters less than the debate suggests.
• Best job: regime detection — flat and crossed means range.
• Crossovers are doubly lagging and whipsaw in ranges.
• An average is not support — no orders rest there, nobody remembers it.
• For F&O: context only, and on the higher timeframe.