Direct Answer

Correlated signals are readings from two or more technical indicators that move together because they are derived from the same underlying price or volume data rather than from genuinely independent information. When a trader stacks correlated indicators expecting extra confirmation, they are often just seeing the same underlying data point twice, dressed up in different formulas, not two separate votes in favor of a trade.

Key Takeaways

  • Correlated signals arise when multiple indicators share the same core input, most often closing price over a similar lookback window.
  • Agreement between correlated indicators is not independent confirmation. It is largely mathematically expected.
  • Momentum oscillators (RSI, Stochastic, Momentum, ROC) are commonly correlated with each other since all derive from recent price changes.
  • Moving averages of different lengths on the same price series are also correlated, especially when the lengths are close together.
  • Correlation can be checked informally, by comparing indicator formulas, or formally, with a correlation coefficient over historical data.
  • Diversifying indicator combinations across categories, trend, momentum, volume, volatility, reduces correlated-signal risk.
  • Correlated-signal risk is a common cause of overconfidence in multi-indicator trading systems.
  • Backtests that stack correlated indicators can look more robust than they actually are, since the "extra" filters rarely disagree.

What Are Correlated Signals?

A correlated signal problem occurs whenever the indicators in a combination are not measuring genuinely different aspects of market behavior. Every technical indicator is a transformation of some underlying data series, typically price, sometimes volume. If two indicators are both transformations of the same closing-price series over similar time windows, their outputs will tend to rise and fall together, even though each has its own formula, scale, and name. The apparent agreement between them adds little new information because it was largely predictable from the shared input alone.

This is distinct from indicators genuinely disagreeing or agreeing based on different information. A momentum oscillator and a volume indicator can both turn bullish for separate reasons, one reflects the speed of price change, the other reflects participation. That kind of agreement is closer to real confirmation. Two momentum oscillators both turning bullish because they are both built from the same 14-period closing-price window is a different, weaker kind of agreement.

How to Recognize Correlated Indicators

There is no single universal formula for "correlated signals" the way there is for a specific indicator, but the underlying statistical concept is the Pearson correlation coefficient applied to two indicator time series:

r = Σ[(Xᵢ − X̄)(Yᵢ − Ȳ)] / √[Σ(Xᵢ − X̄)² · Σ(Yᵢ − Ȳ)²]

where X and Y are the two indicators' historical values over the same set of periods, and X̄/Ȳ are their respective averages. A coefficient close to +1 means the two indicators move together almost in lockstep; a coefficient near 0 means their movements are largely unrelated; a coefficient near −1 means they move in opposite directions. Traders don't need to compute this formally for every pairing, inspecting whether two indicators share the same price input and a similar lookback period is often enough to flag likely correlation, but the coefficient is the precise way to quantify it when it matters, such as when validating a rule-based system.

A Hypothetical Example

Consider a hypothetical trader building a swing-trading system who combines the 14-period RSI with the 14-period Stochastic Oscillator, requiring both to cross above 50 before entering a long position. Suppose over a hypothetical 60-day sample, RSI and Stochastic %K produced a correlation coefficient of 0.91, both are built primarily from recent closing prices over the same lookback window, so they tend to rise and fall together. In that hypothetical case, "two indicators agreeing" happened on roughly 54 of the 60 days, close to what the correlation alone would predict, rather than reflecting two independent pieces of evidence lining up.

Now suppose the same hypothetical trader instead pairs RSI with a 20-day average-volume filter, requiring rising volume alongside the RSI cross. In this hypothetical, the correlation coefficient between the two series was only 0.14, volume and price momentum are measuring different things, so an entry signal that satisfies both conditions carries more genuinely independent support than the RSI/Stochastic pairing did.

Why It Matters

Traders combine indicators to reduce false signals, on the theory that requiring several independent conditions to align filters out weaker setups. That logic only holds if the indicators actually contribute independent information. When correlated indicators are stacked instead, a trader can end up with a system that looks more selective than it is, fewer trades pass the filter, but not because the surviving trades are meaningfully stronger, only because the redundant conditions rarely disagree with each other in the first place. This can produce a false sense of statistical rigor in backtesting, where "three indicators confirmed" reads as stronger evidence than "one indicator confirmed, restated three ways."

Recognizing correlated signals also shapes how traders design combinations going forward: pairing a momentum oscillator with a volume or volatility measure, or a trend indicator with a mean-reversion oscillator, tends to add more genuine information than pairing two indicators from the same family.

Limitations and Common Mistakes

  • Assuming more indicators means more confirmation. Stacking indicators from the same category (multiple momentum oscillators, multiple moving averages) often adds redundancy, not independent evidence.
  • Skipping a correlation check on new systems. Traders building rule-based or backtested strategies rarely test whether their filters are actually correlated before treating agreement as meaningful.
  • Overfitting to a coincidentally low-correlation sample. Correlation between two indicators can shift across market regimes; a low coefficient in one sample period doesn't guarantee independence going forward.
  • Ignoring correlation between input data, not just indicators. Two different indicators can still be correlated indirectly if both weight recent price action heavily, even with different formulas.
  • Treating a correlation coefficient as causation. A high coefficient shows co-movement, not that one indicator is redundant to the other in every market condition.

Measure the Overlap Instead of Assuming It

Most discussions of correlated indicators stop at intuition: momentum oscillators go together, moving averages of similar length go together. That intuition is broadly right and it is not a substitute for checking. Two indicators can be more or less related than their formulas suggest, and on a given instrument over a given period the actual overlap is measurable rather than assumed.

The check is worth building into any rule-based or backtested system, because that is where the cost compounds. A strategy requiring three filters to agree is claiming three pieces of evidence, and if two of those filters are near-duplicates, the requirement is weaker than the design implies while producing fewer trades. Nothing in the backtest output flags this.

The result also has a shelf life. Correlation between two indicators shifts with the regime, so a comfortably low coefficient measured across one calm stretch can rise sharply when volatility changes. Treating a single sample as evidence of permanent independence is its own version of overfitting, and re-checking across different market conditions is the honest version of the test.

What stays true regardless of the number: agreement between tools sharing an input is close to arithmetic. It is not a second vote, and reading it as one is how a modest signal comes to feel like a confirmed one.

Frequently Asked Questions

What are correlated signals in technical analysis?

Correlated signals are readings from two or more technical indicators that move together because the indicators are derived from the same underlying price or volume data, rather than from independent information. When correlated indicators agree, it can look like strong confirmation even though the agreement was mathematically likely from the start.

Why is combining correlated indicators a problem?

Combining correlated indicators inflates a trader's confidence in a signal without adding new information. If two oscillators are both built from closing price over similar lookback periods, they will tend to turn up and down together regardless of whether the underlying setup is actually strong, so stacking them creates an illusion of multi-factor confirmation.

How can a trader check whether two indicators are correlated?

A trader can plot the two indicators' historical values against each other and compute a correlation coefficient over a sample period, or simply inspect the indicator formulas for shared inputs, such as the same price series, similar lookback windows, or one indicator being a derivative of another.

What is an example of two commonly correlated indicators?

RSI and the Stochastic Oscillator are frequently cited as correlated because both are momentum oscillators built primarily from recent closing prices over similar lookback periods, so they often rise and fall together and rarely provide fully independent confirmation.

How can traders reduce correlated-signal risk when combining indicators?

Traders commonly pair indicators from different categories, such as a momentum oscillator with a volume-based indicator or a trend-following moving average, so each one is drawing on a distinct dimension of market data rather than restating the same price information in a different form.

Is this the same thing as "double counting" a signal?

Yes. "Double counting" describes the same underlying problem from the trader's side: treating two indicator readings as separate pieces of evidence when they are really two transformations of the same input, most often the same closing-price series over a similar lookback window. Correlated signals is the more precise technical-analysis term for why the double counting happens.

What correlation between two indicators counts as too high?

There is no threshold that settles it, because the relevant question is what the second indicator adds rather than how similar it looks. Two signals correlated at a moderate level can still be effectively one input if they agree on the specific occasions that matter, and two with a low overall correlation can be redundant within the subset of conditions the rule actually trades. The correlation is a starting point, not a verdict.

Does the correlation between two indicators change with market conditions?

Frequently. Two oscillators can move almost identically through a sustained trend, where both sit pinned toward one end of their range, and decouple in a choppy period where their differing sensitivity matters. That means a correlation computed over a long history averages across regimes and can understate how redundant the pair is exactly when the decision is being made.

Can two indicators be decorrelated by feeding them different price fields?

Only marginally. Highs, lows, closes and typical prices from the same bars are themselves very highly correlated with each other, so swapping the input field changes the numbers slightly and leaves the shared information intact. Genuine separation requires a different data series altogether, such as volume, order book depth, or data from a related instrument.

References

Disclaimer

This page is for educational purposes only and does not constitute investment, financial, or trading advice. Technical indicators reflect historical price behavior and do not guarantee future results. Any prices, data, or correlation figures shown as examples on this page are illustrative and hypothetical, not live or historical market data. Swoopr Investment is not a licensed investment advisor; consult a qualified professional before making investment decisions.