Gold sits at 2,400, a round number. It is also the prior week's high. It is also a level that produced a sharp $18 rejection three weeks ago. Three completely different reasons, all pointing at the exact same price. No single one of them would justify much confidence alone, but stacked together they explain why that particular dollar figure has repeatedly been where gold's momentum has changed. That stacking has a name, confluence, and it is one of the more genuinely useful ideas in level trading, provided it is not stretched into an excuse to see meaning everywhere.
What confluence actually means
Confluence is nothing more than multiple independent reasons landing at the same price. A round number is one reason. A prior swing high is a second, unrelated reason. A fib retracement level from a different move landing nearby is a third. A session extreme lining up is a fourth. None of these individually predicts anything with much reliability, which is exactly the honest starting point: a round number alone is weak, a random fib level alone is weak, a single old swing high alone is moderate at best. What confluence claims, and what tends to hold up, is that when several independently weak signals agree on one price, the odds of a real reaction there improve, because more separate groups of traders have an actual reason to act at that same price simultaneously.
Why stacked reasons upgrade a setup
The improvement is not magic, it is arithmetic on participation. A level with only a round number behind it attracts traders who watch round numbers. A level with a round number and a prior swing high attracts that same group plus everyone tracking swing structure. Add a session extreme and you add the intraday crowd marking yesterday's or last session's range. Each additional independent reason recruits a slightly different group of traders into defending or attacking the same price at the same moment, and it is that combined participation, not any mystical property of the number itself, that makes confluence zones react more reliably than single-reason levels.
Confluence you can actually find
The realistic version of this on a normal trading day: EURUSD resistance at 1.0900 (a round number) happens to also be the exact high from two sessions ago and sits within a few pips of a 61.8 percent retracement of the last big swing down. Three reasons, one price, and a trader marking all three independently before ever combining them has a genuinely stronger case than a trader who marked only the round number. The stacking has to be found, not manufactured: forcing a fib tool to land near a level by adjusting which swing you measure from is not confluence, it is confirmation bias wearing a ruler.
The honest limit of the idea
Confluence upgrades a setup's quality, it does not turn a bad trade plan into a good one, and it never guarantees anything. A three-reason level can still fail, and it will, regularly, because no combination of lines overrides genuine buying or selling pressure when it actually shows up. The discipline is using confluence to choose which of several candidate levels deserves your attention and a slightly tighter stop, not to convince yourself a trade cannot lose. Treat a confluence zone the way the earlier discussion of zones recommends: as an area, not a single price, and wait for an actual reaction there before trusting the stack of reasons over what price is doing in real time.