Confluence in trading — how to count overlapping signals

Confluence in trading is the situation in which several independent analytical tools point to the same place on the chart as significant. A single signal taken in isolation — a support break, an RSI divergence, a pin bar candle — is rarely enough for any edge to show up across a long series of trades. Only when several independent arguments point at the same price does the chance rise that the market will genuinely react there. I show how to write the list of conditions before the session, how to avoid the illusion of confluence assembled in hindsight, and why every notch you raise the bar costs you in trade count.

What confluence actually means

Confluence is a situation in which several independent technical tools converge on the same price level as important and on the same future behaviour as likely. The idea is straightforward. If 1.0850 is at once former resistance turned support, the fifty percent Fibonacci retracement, the fifty-period EMA on the four-hour chart and the place where a pin bar forms, it is no accident that price reacts precisely there. A single signal standing alone gives you far less to work with, and whatever it leaves is easily eaten by spread and commission.

The clearest analogy is courtroom evidence. One witness can be mistaken. Two witnesses who agree are more credible. Five independent witnesses who converge on the same account of events leave very little room for doubt. Overlapping technical signals at one price work the same way. The assumptions of technical analysis on which that logic rests are covered in more depth in the technical analysis section on ForexMechanics.

Eight core signals to count for confluence

The list of available technical tools is effectively endless, but eight of them appear in every serious approach — from the classical canon of John J. Murphy to the price action school of Al Brooks. They form the core of any practical way of counting confluence and are the first thing a trader should master when learning to work with support and resistance levels and Fibonacci retracements.

Why higher confluence improves precision but lowers frequency

The higher the threshold a trader sets, the less often a position gets opened — and the more carefully chosen the ones that do. The mechanism is not complicated. Stricter requirements filter out part of the false signals, but they filter out part of the genuine ones too, because the market is under no obligation to line every argument up neatly in one place.

The trade-off shows up after a few weeks of watching. At a two-signal threshold, a dozen or more candidate trades appear on a single pair in a day, and a good share of them are impossible to tell apart from noise. At a four-signal threshold you are left with a handful a week, but each one has a reason you can write down in a single sentence. At a six-signal threshold you trade a few times a month and need the temperament not to click for the rest of the time. There is no single correct value — only the one you can actually hold to and the one your own records support. Four is a sensible starting point for someone fitting trading around a day job.

The deadliest trap — confluence assembled in hindsight

The most serious risk in this approach has nothing to do with statistics. It is something far more subtle and far more damaging to the account, namely the habit of layering justifications onto a trade only after it has worked out. A trader opens a position on a single signal, celebrates the win, then returns to the chart and conveniently discovers that the support break was accompanied by a Fibonacci level, a moving average, an RSI divergence and a volume spike. Six signals. Suddenly there is a story about a marvellous six-factor setup — completely useless as a system, because none of those signals existed in the trader's reasoning before entry.

This is textbook confirmation bias, well described in the literature on trading psychology. The remedy is singular. The list of conditions must be written down before the session, not after it. The trader opens the journal, writes out the chosen eight signals, declares the minimum threshold required to open a position, and only then begins to observe the market. Every signal added after entry is a piece of fiction, not a system. A formal setup checklist consulted before every click of the button is a robust antidote.

A clearly-hypothetical six-signal example

The best way to understand how stacked confirmations work in practice is to walk through one example from start to finish. The setup below is illustrative only — it demonstrates the structure of the reasoning, not an actual trade pulled from a historical archive. Imagine you are watching EUR/USD on the four-hour chart, with the daily trend already established as bullish.

The example shows how to document multiple conditions before entry. Because these tools can share the same underlying prices, calling them independent does not establish statistical independence. Whether additional conditions improve expectancy requires a test of the complete rule, costs and missed trades.

Practical takeaway: Several conditions derived from the same prices need not provide independent evidence. Test their combined effect.

Three rules of discipline before the session

Three rules allow the trader to harness the strength of confluence without falling into the after-the-fact justification trap. First, the list of signals must be written down the day before and must not change during the session. Second, the minimum confluence threshold is declared quantitatively — four signals, say — and every entry below that threshold is simply a breach of the system. Third, the trading journal contains a mandatory field labelled "number of signals at entry" and that field must be filled before the click, never after.

After a hundred recorded trades you have your own data, and only then is it worth discussing results in numbers. They will almost certainly come out worse than scrolling back through a chart suggested — a spread widening during a macro release, slippage and ordinary departures from the plan all take their share. What matters is not whether you land on some particular percentage, but whether trades with more confirmations do systematically better than trades with fewer. If they do not, the threshold is improving nothing for you and the cause lies elsewhere.

What to do tomorrow — implementing confluence step by step

1. Write down the eight signals tonight on a permanent reference card and place it next to your trading station. For each item, specify how you will confirm its presence — which indicator, which window, which timeframe. Without that grounding the next step cannot be done consistently, because each time you would be counting something slightly different.
2. Declare a minimum threshold of four signals and do not deviate from it for the next three months, not even by one setup. The first four weeks will be dull — most of the trading day will pass in observation rather than execution. That is the price of ending up, three months later, with a sample large enough to judge at all.
3. Record the number of signals in your trading journal at every entry and review your results monthly, sorted by confluence threshold. One question interests you: do trades with five confirmations come out better than trades with three? The answer comes from your own journal, not from a table found on the internet.
4. Begin to vary your position size gradually in response to the number of confirmations, but only after a hundred documented trades using your own version of the system. Earlier than that the data is too thin to justify increasing risk. The standard one percent of capital per trade remains the starting point until your own statistics justify any calibration.