Moving averages — SMA, EMA, WMA, HMA

Open any daily chart and the price line looks jagged — up a day, down a day, a spike here, a gap there. Somewhere inside that mess there is usually a direction: the stock is broadly climbing, broadly falling, or broadly going nowhere. But the day-to-day jumps hide it. Your eye keeps getting pulled to the last big candle instead of the shape of the whole move.

So here is the plain question this module answers: is there a simple way to smooth the jaggedness away and leave just the direction? The answer is the oldest and most-used tool in all of technical analysis — the moving average. It is nothing more than the average price over the last few days, redrawn every day. And its single most important property, the one everything else in this module hangs on, is that it always arrives a little late.

A single day's price is loud and unreliable. It reflects that day's news, that day's mood, one large order that happened to land in the afternoon. Take any one day on its own and you cannot tell a real turn from a random twitch.

Averaging fixes this. If you take the last ten closing prices and average them, one wild day gets diluted by nine calmer ones. Do that every day — dropping the oldest price and adding the newest — and you get a line that moves across the chart: a moving average. Because it blends many days together, it ignores the small stuff and shows the broad drift. When the line slopes up, recent prices are on average higher than older ones; the trend is up. When it slopes down, the trend is down. When it runs flat, the market is going sideways.

This is why the moving average sits at the head of the "lagging indicators" part of this book. It is a lagging indicator — a tool that describes what price has already done, rather than guessing what it will do next. That word "lagging" sounds like an insult, but it is an honest label. The whole point of averaging the past is to be steadier than the present, and steadiness costs you speed. A moving average will never call the exact top or bottom, and it is not meant to. It exists to answer a calmer question: which way has this been going, once I stop staring at the last candle?

The trend, once established, tends to keep going for a while — that is the empirical regularity every moving-average user is quietly leaning on. A trend in motion usually persists longer than it feels like it should, which is exactly why a lagging line that simply follows the trend can still be useful even though it is always a step behind.

Every moving average answers the same question — what is the typical price lately? — but they disagree on one thing: how much should recent days count versus older days? That single choice produces the four you will meet everywhere.

SMA — the Simple Moving Average. Add up the last N closing prices, divide by N. Every day counts exactly the same. A 20-day SMA is the average of the last 20 closes, full stop. It is the smoothest and the most sluggish: because it treats a price from 20 days ago as equal to yesterday's, it is slow to react to fresh moves — but for the same reason it is the calmest and the hardest to fool with a one-day spike.

EMA — the Exponential Moving Average. Same idea, but recent days are given more weight and older days fade out gradually. Yesterday matters more than the day before, which matters more than the day before that. So the EMA hugs price more closely and turns sooner than the SMA. It reacts faster — which is helpful when a real move begins, and a nuisance when the "move" is just noise.

WMA — the Weighted Moving Average. A blunter version of the same instinct: it multiplies each day by a plain number (the newest day by N, the next by N−1, and so on down to 1). Like the EMA it front-loads recent prices, so it also reacts faster than the SMA. In practice the EMA is far more common; the WMA is worth knowing mainly because it is the building block inside the next one.

HMA — the Hull Moving Average. A clever construction (by Alan Hull) that combines weighted averages to cut lag hard while staying fairly smooth. The HMA hugs price the closest of the four and turns the soonest. That makes it feel wonderfully responsive in a clean trend — and dangerously twitchy in a sideways market, where it will follow every wiggle.

Notice the one axis running through all four: weight the recent past more, and you trade lag for noise. There is no setting that gives you fast and smooth. Every moving average lives somewhere on that trade-off, and choosing one is really just choosing where on the line you want to sit.

Watch what the lag actually costs and buys you. illustrative

Below is a composite stock that trends up cleanly, then chops sideways in a tight range for a stretch, then trends up again. Imagine a 20-day SMA drawn through it (we cannot draw the curve on a candlestick chart, so hold Figure 1 in your head). In the two clean trends, the average sits calmly below the rising price and simply confirms the direction — late, but right. In the choppy middle, the price keeps crossing back and forth over where the average sits, and every crossing looks like a "signal" that immediately reverses.

That is the entire honest story of a moving average in one picture. It is a trend tool. In a trend it is quietly excellent — it keeps you on the right side and stops you exiting at the first wobble. In a range it is actively harmful, firing a signal every few days that costs you money. The tool did not change; the market did. Knowing which market you are in matters far more than which average you picked.

A moving average cannot tell you the future. It is arithmetic on prices that have already printed, so by construction it can only ever describe the past a little late. When people say a moving average "predicted" a move, what they mean is that the trend it was tracking happened to continue — which it does often enough to be useful, and fails often enough to bankrupt anyone who forgets it can fail.

It cannot tell you when the trend will end. The line slopes up right until the day it doesn't; the turn only becomes visible in the average well after price has already reversed. That delay is the flip side of its calmness — you cannot have a line that both ignores noise and reacts instantly. A moving-average signal shifts the odds a little; it never settles the outcome, and treating the smooth line as a guarantee is how a lagging tool turns into a trap.

It cannot tell you anything at all in a sideways market — worse, it will actively mislead you there, because a flat, choppy price keeps crossing the average and manufacturing signals that are pure noise. This is the single most important limit to hold onto: a trend tool used in a non-trend is not neutral, it is dangerous.

And it cannot tell you what to own. The line knows nothing about the company — its cash, its debt, whether the business is any good. It is a picture of price, and price is the map, not the territory. The average is a summary of the crowd's recent behaviour, not a statement about the business underneath. Use it to think about when, never about what.

The same handful of mistakes catch people over and over.

1. Treating a slope or a "price above the average" as a buy signal. It is a description of an existing trend, not a trigger. By the time price is comfortably above a rising 200-day average, most of the easy move is already behind you.

2. Hunting for the "best" moving average. There is no best. Every length and type is a point on the same speed-versus-smoothness trade-off, and the one that looked best on last year's chart is simply the one that happened to fit last year's chop. That is the classic the setting that looks perfect on old data is usually fitted to that data's noise — a faster or slower line will win next year, and you cannot know which in advance.

3. Using a trend tool in a range. Covered above, but it deserves repeating because it is so common: in a sideways market a moving average whipsaws you to death. Check whether there is a trend before you trust a trend tool.

4. Forgetting that popular lines are partly self-fulfilling. Price reacts around the 50 and 200-day partly because everyone watches them. That makes them useful, but it also means they can be gamed — a big move that "defends the 200-day" may be traders acting on the line, not new information about the business.