Combining indicators
The natural instinct after learning five indicators is to load all five and wait for unanimous agreement. This produces either paralysis or an illusion of certainty. Here is the more useful frame.
One indicator per job
Indicators come in families, and members of the same family are the same maths in different clothes. RSI, Stochastic and Williams %R agreeing is not confirmation — it is the same measurement three times. A functional combination takes one tool per question:
| Question | Family | Pick one |
|---|---|---|
| Is there a trend, and which way? | Trend | 50/200 MA, ADX |
| Is the move gaining or losing force? | Momentum | RSI or MACD |
| How much should I expect price to move? | Volatility | ATR (sizing), Bollinger (regime) |
| Is participation backing the move? | Volume | Volume bars, OBV, Volume Profile |
| Where exactly am I wrong? | Levels | Support/resistance |
A worked confluence read
Five tools, five different questions, one coherent plan — including the exit, before entry.
What indicators cannot do
- They lag. Every indicator is arithmetic on past prices. None contains information about the future; they organise the past so you can bet on continuation or exhaustion.
- They are self-referential. Widely-watched levels and crossovers "work" partly because many traders act on them — and fail precisely when the crowd is positioned the same way (see Derivatives data).
- They know nothing about events. A Fed decision or an ETF headline (see Macro data) will slice through the cleanest setup.
A note on what this guide deliberately leaves out: Elliott Wave, Gann tools, and Point & Figure charting all have devoted followings and long histories. We omit them because their frameworks are flexible enough to fit any outcome in hindsight — the opposite of testable. If they intrigue you, read critically and paper-test first (see Practice & review).
If some indicator combination reliably predicted price, it would be arbitraged into silence. The realistic goal is a modest, repeatable edge applied with strict risk control — which is why the next category matters more than this one.
Build a small stack: one trend filter, one momentum read, ATR for sizing, levels for stops, volume for confirmation. Then let risk management — not indicator count — carry the season.