Confluence in Technical Analysis: When Agreeing Signals Add Information

Confluence in technical analysis means several separate signals pointing to the same trade at the same place and time: a support level that is also a Fibonacci retracement and a moving average, or an uptrend confirmed by a momentum indicator and by volume. The idea is that each extra reason makes the trade more likely to work. That only holds if the reasons are independent, and many common indicators are not: on five years of the S&P 500 ETF, the RSI and a 20-day moving average gave the same reading on 95.7% of days.

Types of confluence

• Level confluence. Several tools put a level at the same price: a previous support or resistance, a Fibonacci retracement, a long moving average, a round number. The overlap is often drawn as a zone.
• Indicator confluence. A trend tool and a momentum tool agree, for example price above its moving average and the RSI above 50.
• Timeframe confluence. The daily and weekly charts point the same way.
• Pattern and level. A candlestick or chart pattern forms at a level, such as a double bottom on an old support.
• Volume. A move on volume above its average, read as more participants behind it.

Two signals, one reason: how often indicators agree

Three popular readings of the same chart: the 14-day RSI above or below 50, the MACD above or below its signal line, and the close above or below its 20-day exponential moving average. Each is a bullish or bearish signal on every day. If they were three independent opinions, they would disagree often. Here are the moving average and the RSI on the same 120 days of the S&P 500 ETF:

On these 120 days, from May 8 to October 28, 2026, the RSI and the moving average gave the same reading on 109 days, the RSI and the MACD on 72, and all three on 70. Over the full sample, the counts are just as lopsided.

S&P 500 ETF: 1,357 days, June 4, 2021 to October 28, 2026, dividend-adjusted prices. Bitcoin: 495 days, June 21, 2025 to October 28, 2026. Both start once all three indicators have enough history.

The RSI above 50 and the close above a short average are almost the same statement: both say that recent closes have been rising more than falling. Counting them as two reasons is counting one reason twice. The MACD is closer to a separate opinion, and it still agreed with the others on about seven days in ten.

Did agreement improve the next 20 sessions?

If confluence adds information, days when all three signals were bullish should have been followed by better returns than days when none were. We grouped every day by its number of bullish signals and measured the return over the next 20 sessions.

Returns are measured over the 20 sessions after each day, from close to close, so the last 20 days of each sample have no result and are left out: 1,337 ETF days and 475 Bitcoin days are grouped.

On the ETF, the days with full bullish agreement were followed by the lowest mean return of the four groups. On Bitcoin, all-bullish days did better than all-bearish ones, but both were followed by losses on average. Neither market shows returns that rise with the number of agreeing signals. One limit: consecutive days share most of their next 20 sessions, so 579 days are not 579 independent trades, and the gaps between groups are smaller than they look.

Level confluence: what to check

The other kind of confluence is a level confirmed by several tools, as in the schematic below. It is harder to judge than indicator agreement, because each tool comes with settings (which swing, which retracement, which average), and once enough tools are drawn almost any price sits near one of them. Two habits keep it honest: choose the tools and their settings before looking at the chart, and define "near" in advance, for example within half an average true range of the level. A zone found after the move, by trying tools until two of them line up, is a description of the past, not a reason for the next trade.

This page does not measure how often overlapping levels held. Our support and resistance page tests single levels, swing lows and highs on the ETF, with rules written in advance.

A confluence checklist that avoids double counting

1. One trend tool. A moving average, or the direction of swing highs and lows. Not both a moving average and the RSI.
2. One level. Where is the nearest support or resistance, and is it a swing point, a retracement or an average? Two tools on the same price count as one level, with more than one name.
3. One momentum reading, if it is built differently. The MACD adds more than a second short average does.
4. Volume. Is the move on volume above its recent average?
5. The calendar. Is an earnings report, a central bank decision or a data release due before the trade can work? A chart cannot see it.
6. The stop. Where is the trade wrong? Confluence does not shrink the loss when it is.

Chart patterns are one way to combine a level and a shape into a single rule, and our bench measures them that way, as described in our methodology. For trends, see trend trading; more guides are in technical analysis.