Types of indicators — leading, lagging and the four families

The answer

Every indicator is a calculation done on price and volume, and it answers 1 of only 4 questions: which way, how fast, how much movement, and how much conviction. Each one is also either leading, which means early and often wrong, or lagging, which means late and usually right — and no setting removes that trade-off.

Why this costs you money

Here is the most common chart in retail trading. Price at the top. Below it, RSI. Below that, Stochastic. Below that, MACD. Sometimes a fourth panel with CCI.

All 4 of those are momentum indicators. They are built from the same raw material, which is recent price change. On most days they say the same thing at almost the same moment. The trader looking at that chart believes 4 independent sources have agreed. In fact 1 source has spoken 4 times.

This is not a small error. It is the mechanism behind a specific and expensive behaviour: you take a trade you would have skipped, at a size you would not have used, because you thought you had confirmation. Then the trade goes against you and you hold it, because 4 indicators cannot all be wrong. They can. They were always 1 indicator.

The second cost is quieter. A trader adds indicators until the chart produces a signal in every market condition. That feels like progress. What has actually happened is that the chart now always says something, so it can never say nothing — and "nothing is happening here" is the single most profitable message a chart can send.

You do not need more indicators. You need to know which question each one answers, and to stop asking the same question 4 times.

How it works

An indicator takes the numbers you already have — open, high, low, close and volume — and rearranges them into a line. It adds no new information. That sentence is worth reading twice, because it destroys the idea that an indicator can see something the price does not contain.

What an indicator does is make 1 property of the price easier to see. That is a real service. Your eye is poor at judging whether the last 20 days moved faster than the previous 20. A calculation is good at exactly that.

The four families

Trend, momentum and volatility are all built from price. Only the volume family brings in a second, genuinely separate input. That is why a volume indicator adds more to a chart that already has RSI than a second momentum indicator ever will.

Leading and lagging

A leading indicator tries to signal a turn before it happens. Oscillators do this. They measure how far and how fast price has moved and they say "this has gone far enough". Sometimes it has. Often the move continues for months while the indicator screams.

A lagging indicator confirms a move that has already started. Moving averages do this. A 200-day average cannot turn up until a large amount of rising price has already happened, by construction.

The trade-off is not a flaw in any particular indicator. It is arithmetic. To be early you must act on less information. To use more information you must wait. There is no setting that gives you both, and every trader who has ever shortened a period to "get in earlier" has discovered they also got more false signals. That is the whole trade, and it never improves.

Leading indicators are early and frequently wrong. Lagging indicators are late and usually right. Choose which error you can live with, because you will get one of them.

What it tells you, and what it does not

An indicator tells you a property of the past price. It does not tell you the future price, and it does not know anything about the company, the sector or the news.

Three specific limits are worth stating.

An indicator cannot tell you which market condition you are in. RSI does not know whether it is in a trend or a range. It computes the same way in both. You have to supply that judgement, and if you supply it wrongly the indicator will confidently mislead you. This is why a trend measure such as ADX is often the first thing to put on a chart, not the last.

A default setting is somebody's choice, not a property of markets. RSI is 14 periods because Welles Wilder wrote 14 in 1978. There is no evidence that 14 is optimal for a stock listed in 2019 on a 15-minute chart. When you change the setting the signals change, which proves the setting was never a fact.

An indicator that fits the past can be an accident. Given enough rules and enough settings, something will look excellent on any historical chart. Sullivan, Timmermann and White showed in 1999 that a large set of technical trading rules, tested properly against the number of rules that were searched, produced far less evidence of profitability than the raw results suggested.

The decision rule

Choose 1 indicator from each family you actually need, and no second one from any family.

If you trade trends: 1 trend indicator, then 1 volume indicator to check conviction. A momentum indicator is optional and it must be used to measure strength, not to predict a top.

If you trade ranges: 1 momentum indicator, then 1 volatility indicator to tell you how wide the range is. Add a trend indicator for 1 purpose only — to tell you when the range has ended and to stop taking the momentum signals.

Unless 2 indicators from the same family disagree. That is not confirmation and it is not a signal. It means your settings differ, and it tells you about your settings, not about the stock.

Try this now

This takes about 5 minutes and it is the cure for the crowded chart.

1. Open any chart you look at often, on your own broker app or on a free charting site. Set it to daily candles and 1 year.
2. Add RSI (14) in a lower panel. Add Stochastic (14, 3, 3) in another lower panel below it. Both are momentum indicators.
3. Look at the last 6 months. Count the number of times one of them went into its extreme zone (RSI above 70 or below 30, Stochastic above 80 or below 20) and the other did not do the same thing within 2 days.
4. Now remove the Stochastic. Add OBV instead, which is a volume indicator.
5. Look at the same 6 months. Find 1 date where RSI and OBV pointed in different directions.

What you should see. In step 3 the count will be low. On most charts the 2 momentum indicators enter and leave their extreme zones together, within a day or 2, almost every time. They are not confirming each other. They are the same measurement with different arithmetic.

In step 5 you will find disagreements easily, and they will be interesting ones — price rising while OBV falls, or the reverse. That is what a second family buys you. The disagreement is the information. Two indicators that never disagree cannot tell you anything you did not already know from 1 of them.

Three real cases

1. J. Welles Wilder Jr., New Concepts in Technical Trading Systems, 1978 (United States) — four families, one book A single self-published book introduced RSI (momentum), Average True Range (volatility), the Directional Movement system with ADX (trend) and Parabolic SAR (trend-following exit). Wilder was a mechanical engineer before he traded. He built these tools by hand, for hand calculation, on commodity futures with far higher volatility than a large-cap share. The default periods he printed — 14 for RSI, 14 for ATR and ADX — are still the defaults on almost every platform in the world, nearly 50 years later, on instruments he never saw. They are a 1978 convenience that became a global standard by repetition.

2. Brock, Lakonishok and LeBaron, Journal of Finance, 1992 (United States) — a lagging rule that worked, then stopped The 3 authors tested simple moving average rules on the Dow Jones Industrial Average over roughly 90 years of data ending in 1986, and reported that the rules had predictive power. The study is still cited by people selling moving average systems. What is cited less often is what followed. Sullivan, Timmermann and White re-examined the same universe of rules in 1999, correcting for the fact that many rules were searched before the best one was reported, and found the evidence far weaker — and specifically weak in the period after the original study ended. A rule that works on data you found it in is not yet a rule.

3. The COVID crash and recovery, February to July 2020 (United States and India) — leading and lagging, on one chart, 4 months apart In late February and March 2020 momentum indicators on the S&P 500 and the NIFTY 50 reached extreme oversold readings and stayed there for weeks while price kept falling. The leading indicator was early and wrong for the whole decline. Then the 50-day moving average crossed below the 200-day — the "death cross" — within days of the market low on 23 March 2020. The lagging indicator was late and its sell signal arrived at almost the worst possible price. Both families failed, in exactly the ways their construction predicts. Neither failure was a surprise to anybody who understood what the calculation does.

The question that resolves it

A novice looks at a chart and asks: what are my indicators saying?

An expert asks: how many independent things are on this chart?

Four momentum indicators is 1 thing. A trend indicator, a volume indicator and a price level is 3 things. The number of panels tells you nothing. The number of separate inputs tells you everything.

What would make this wrong

The claim here is that indicators from the same family carry nearly the same information, and that leading and lagging is a genuine trade-off rather than a solvable problem.

You would falsify the first part by finding 2 same-family indicators that routinely disagree on the same chart with default settings, in a way that is useful. Do the counting exercise above on 10 charts. If the momentum indicators disagree often, this article is wrong about your instrument.

You would falsify the second part by finding a setting that produces both early signals and few false ones. Nobody has, and the reason is structural rather than empirical. But test it: shorten any period and count the signals before and after.

The honest limits are 2.

First, "same family" is a rough grouping, not a law. MACD and RSI are both called momentum indicators, and they are built differently enough that they do sometimes diverge meaningfully. The grouping is a rule of thumb for reducing chart clutter, not a mathematical statement.

Second, none of this establishes that indicators work at all. Park and Irwin's 2007 survey of the academic literature found a majority of modern studies reported positive results for technical rules, but the survey itself flagged data-snooping and selection problems across the field.` This cluster teaches you what each tool measures and where it breaks. It does not promise that any of them make money.

Carry this

• An indicator adds no information. It rearranges price and volume so 1 property is easier to see.
• There are 4 questions: direction, speed, movement, conviction. Pick 1 tool for each question you actually need.
• Early and wrong, or late and right. Choose your error. There is no third option.

Knowledge check

Related

• Moving averages
• Oscillators explained
• Confluence and confirmation — combining signals without overload
• ← Indicators and confirmation