The problem with stacking
The instinct is sound: if one indicator is uncertain, three agreeing should be more certain. The execution usually is not, because most indicators you would reach for are measuring nearly the same thing.
An SMA(20), an EMA(20) and a Bollinger mid-band are, for practical purposes, the same line — one is a smoothed average, another is a differently-weighted smoothed average, the third is a smoothed average. If all three say "up", you have one piece of evidence presented three times. RSI(14) and Stochastic(14) are likewise both measuring recent directional strength on similar windows. Agreement between them is not independent confirmation; it is arithmetic inevitability.
This is collinearity, and it is why adding indicators makes people feel better calibrated while their hit rate does not improve. You are not adding evidence, you are adding confidence.
A rule that actually works: one per family
The fix is to pick indicators from different families, because then they genuinely measure different things:
- One trend indicator for direction and regime — a moving average, or MACD.
- One momentum indicator for thrust and exhaustion — RSI, or the Stochastic.
- One volatility measure for stops and size — ATR, or band width.
Three readings, three questions, no double counting. If you find yourself adding a fourth, you are almost certainly adding a restatement of one of the three.
Also: never stack without a time frame plan
Indicators on the same chart, same time frame, are the most correlated of all. If you want genuinely separate information, the cheapest source is a higher time frame: the daily trend sets the bias, the hourly gives you the entry. That is real, independent context — not another line computed from the same bars.
Write the rule down before you look
The last failure mode is not about indicators at all. It is adding a filter after seeing a setup you like, to justify taking it. If your rule is "long when price is above the 200-day and RSI is above 50", write it down and then go looking for setups. If you instead find a chart you want to buy and then go hunting for indicators that agree, you are not analysing — you are rationalising, and the indicators will oblige you every time.
How to check this yourself
Take your current setup, whatever it is, and list every indicator on your chart. For each one, write one sentence on what it measures. Now look for two sentences that say the same thing in different words — that is your redundancy, and deleting one of them will improve your decision-making without changing a single number on the screen.