Why indicators mislead – Trading with a plan

In this lesson

understand why indicators come late
see what false signals in a range cost
find out how not to be fooled by a “perfect” strategy from the past

Indicators look like magic: one line crosses another and you know what to do. If it worked like that, everyone who can open a chart would be rich. This lesson is about why they are not.

Indicators are calculated from the past. The crystal ball only shows what has already happened.

1. Delay#

Every indicator is calculated from prices that have already happened. A 200-day average needs a lot of new prices before it turns. That is why the signal arrives after a large part of the move has happened – you already know this from the lesson on moving averages.

2. False signals in a range#

In a trend, trend-following indicators make sense. In a range, however, the price jumps back and forth and the indicators with it. The result: one false signal after another, known as a whipsaw.

Bitcoin in 2024: from April to October the price moved in a range roughly between 49,000 and 73,000.

10 August – a death cross at a price of 60,924, five days after the panic day.
28 October – a golden cross at 69,962.

Anyone who sold and bought back on the crosses would have bought back 15% higher – plus fees.

SMA 50
SMA 200

In the range, the averages crossed at the moments when it was least useful.BTC/USDT · 1D · Data: Binance Vision · CC BY-NC-SA 4.0

3. Ten indicators, one piece of information#

RSI, stochastic, MACD, CCI… Most indicators are calculated from the same thing: closing prices. When they all agree, you do not have ten confirmations – you have one piece of information shown ten times. More lines on a chart mostly mean more chances to find the one that agrees with what you already think.

4. Overfitting to the past#

On historical data you can find a setting that would have "made" a great profit. Just keep trying average lengths and RSI thresholds until it works. This is called overfitting: the strategy learns random details of the past that will not repeat.

How to protect yourself:

Simple rules with few settings survive better than complex ones.
Test on data you did not use to invent the rules. If a strategy only works on the period it was built on, it does not work.
Count fees and slippage. Many strategies are profitable only on paper, without costs.

There is no indicator that always works

If there were, everyone would use it – and it would stop working. Indicators are aids for orientation, not a money machine. Anyone selling you a "90% accurate indicator" is mostly selling a dream.

What to use, then#

A sensible base is small: market structure, support and resistance zones, one trend filter (for example SMA 200) and volume. Add anything else only when you can say what it gives you on top.

Common myths#

"When five indicators agree, it is certain." It is one piece of information five times.
"This strategy would have made 300% over the last year." On the past, you can make anything.
"I just have not found the right setting yet." Maybe because there is none.

Summary

Indicators are calculated from past prices, so they come late.
In a range, trend indicators give false signals that cost money.
Most indicators carry the same information – more lines do not mean more certainty.
A strategy tuned to the past (overfitting) often fails in the future.