Combining chart signals: why stacked indicators can overstate agreement

1. How to read it / Where it misleads

Traders who read several chart tools at once can mistake one piece of price information repeated three times for three separate confirmations, and the creator of one widely used tool warns against exactly that. Crypto price charts often carry more than one tool: a moving average, an RSI panel underneath, perhaps bands or a volume bar. The habit has a name in chart analysis: confluence, meaning several separate readings that describe the same thing about a price at the same time. This Field Guide entry explains what confluence is meant to measure, why one of the tools' creators warns about how it is done, and where the practice tends to fool the person doing it.

This explains how traders describe and combine chart readings. It is not a signal, forecast or recommendation, and past readings do not predict prices. This entry uses no market data.

2. The mechanism

Each indicator compresses price history into one number. A moving average smooths closing prices into a direction; RSI, the relative strength index created by J. Welles Wilder Jr., scores recent gains against recent losses on a scale of 0 to 100 (source 2). How each is built is covered in Le Coin's entry on moving averages and oscillators (source 3).

The argument for combining them is about independence. If two readings draw on different information, agreement between them says more than either one alone. John Bollinger, who created Bollinger Bands, puts the condition plainly in his published rules: when more than one indicator is used, they "should not be directly related to one another." His example is that a momentum indicator might complement a volume indicator, "but two momentum indicators aren't better than one" (source 1).

Confluence only adds information when the inputs come from different places: price direction, momentum, volatility, volume, or positioning data such as open interest. Bollinger's own list of indicator sources is momentum, volume, sentiment, open interest and inter-market data (source 1).

3. The evidence on the pitfalls

The most common failure is counting the same thing twice. RSI, MACD (moving average convergence divergence, a gap between two moving averages) and a short moving average are all arithmetic on the same closing prices. When all three turn up together, a chart can look as if three witnesses agree. Three momentum readings on one price series amount to one reading drawn three ways.

The second failure is treating a level as a verdict. Bollinger writes that a price touching his upper band "is NOT in-and-of-itself a sell signal," and a touch of the lower band is not a buy signal; in trending markets, prices can "walk up the upper Bollinger Band" for long stretches (source 1). The same caution applies to RSI. TradingView notes that Wilder's 70 and 30 thresholds are ones some traders widen to 80 and 20 "at the trader's discretion" (source 2). A reading of 72 does not mean the same thing in a quiet range as in a strong advance.

The third failure is tuning settings until the past looks tidy. Default inputs, such as 14 periods for RSI or 20 periods and two standard deviations for Bollinger Bands, are starting points. Bollinger calls his own defaults "just that, defaults" (source 1). Adding settings and indicators until a chart lines up with past turns risks fitting the history rather than describing anything that carries forward.

Crypto adds a fourth. Markets trade around the clock, so the daily close behind every indicator depends on the cutoff each data provider chooses. Two platforms can show different readings for the same coin on the same day, and a confluence that appears on one chart may not appear on another (source 3).

4. The objection

Experienced users of these tools do not agree on what their signals mean, which is a reason for caution about stacking them. TradingView's RSI guide records that Wilder read a bullish divergence, where price makes a new low but RSI makes a higher low, as a sign that a rise was coming. Andrew Cardwell, a later RSI analyst, read the same pattern as a brief correction inside a continuing downtrend (source 2). When two serious users interpret one pattern in opposite ways, adding a second indicator to confirm it can simply import a second person's assumptions.

There is a fair counterpoint. Bollinger's rules describe comparing price action with indicator action, measured against the bands, as a basis for more rigorous analysis (source 1), and confluence can be read as a prompt to look closer rather than as a trigger. In Le Coin's view, that is the most defensible way to read it: as a check on whether the information sources differ, not as a count of how many lines agree.

5. The tell

The useful question about any confluence is what each input actually measures. If the answer for every panel on the chart is "recent closing prices," the agreement is one observation repeated. If the inputs come from different families, such as price, volume and volatility, the agreement is at least drawn from separate evidence, though it still describes only what has already happened.

Not financial advice.