Confluence of Signals: Why Three Weak Pieces of Evidence Together Equal One Strong One

It's Tuesday evening, you open the chart, and the moving average has just crossed the price from below. Your heart races — "this is it." Your finger hovers over the button. But two days later, the price drops by 6%, and you stare at the chart feeling like you've been deceived. You were misled by a single isolated signal that you believed in as if it were evidence in court. Yet it was more like an anonymous text: "something's happening, maybe."

One Signal is a Witness, Three Signals are Evidence

Imagine a court case based on a single witness who isn't entirely sure. The chance of error is high. Now add fingerprints, a video recording, and a bank statement — each imperfect on its own, but all pointing in the same direction. This is what is called confluence in analysis: a point where multiple methods independently converge and say the same thing.

The key word is independently. When you have three indicators that essentially calculate the same thing (like three different moving averages), you don't have three pieces of evidence — you have one piece of evidence said three times. It's like three brothers swearing to the same alibi. True confluence mixes different perspectives: trend, price level, volume, fundamentals, sentiment.

Why It Works: A Bit of Probability Math

Let's put it in numbers, because that's where it shines.

Let's say you have one signal with an approximate 55% success rate — slightly better than a coin toss. That sounds dull. Now imagine a second, independent signal with the same success rate, and a third as well. If they are truly independent and you require all three to agree, your trade profile changes significantly:

• You have fewer opportunities (logically — it's rarer for three things to agree at once).
• But the opportunities that pass the filter have a significantly higher success rate than a lone signal.

A small illustration of the power of independence: the probability that two independent events with 55% truthfulness both err is 0.45 × 0.45 = roughly 20%. For three, it's 0.45³ ≈ 9%. This doesn't apply literally to markets (signals are never perfectly independent, and the market isn't a fair coin), but the principle is solid: combining weak but independent hints reduces the risk that they all err at once.

And now the important part, what the data says — and what it does NOT. Confluence does not increase the return of an individual trade. It doesn't make anything certain. What it does: improves the signal-to-noise ratio. It filters out a lot of random fluctuations that would otherwise cost you fees and nerves.