How to Combine Multiple Indicators Without Overfitting Your Strategy

A single RSI strategy returned 30% on Bitcoin over 2.4 years. A five-indicator overfitted version made nothing. Zero trades. The difference is not skill. It is how you combine multiple indicators without overfitting.

Most traders assume more indicators equal better decisions. They stack RSI on top of MACD on top of Bollinger Bands on top of volume. The chart looks like a Christmas tree. The strategy feels sophisticated.

But backtesting tells a different story. Across 2.4 years of BTC/USDT daily data on CoinQuant's AI trading platform, a single RSI strategy returned 30%. A five-condition overfitted strategy never fired a single trade. A properly combined two-indicator strategy returned 25% with a 100% win rate. No coding required to build or test any of them.

The Three Strategies We Tested

We ran three approaches on BTC/USDT daily data from January 2024 through May 2026 using Binance price data from Kaiko. All three used long-only market entries with no stop loss. Initial capital: $10,000. Each strategy took under sixty seconds to build in CoinQuant: describe it in plain English and the AI generates the conditions.

V1: Single Indicator (RSI): RSI(14) crosses below 30 to enter long, crosses above 70 to exit. One indicator, two conditions. Simple. The baseline every trader starts with.

V2: Overfitted (Five Conditions): RSI below 30 AND MACD line crosses above signal AND SMA(50) above SMA(200) AND ADX above 25 AND volume above its 20-bar average. Five conditions, all measuring momentum or trend direction. Same signal, five different ways of asking for permission.

V3: Complementary Pair: RSI(14) crosses below 30 AND close price is above SMA(200). Two indicators answering fundamentally different questions: momentum (is price at an extreme?) and trend (which direction is the market moving?).

Backtest Results: January 2024 to May 2026

Data: Kaiko via CoinQuant API. Binance BTC/USDT daily. Backtests run May 21, 2026. $10K starting capital, long-only, no stop loss. Backtest IDs: V1 = 6ace52ef, V2 = 652da1fb, V3 = 46d981a2. Open at app.coinquant.ai/backtests/{id} for screenshots.

Three things stand out. First, the single RSI delivered strong returns but with significant drawdown: a 21% peak-to-trough decline means a $10,000 account hit $7,906 at its lowest point. Second, the overfitted V2 never traded: five conditions layered on top of each other left zero viable entry signals across 2.4 years of daily data. Third, the complementary pair won every trade: two for two with a 25% return, comparable to the single RSI but with cleaner, trend-confirmed entries.

The Overfitting Trap

The overfitted V2 is what happens when traders keep adding conditions until the backtest looks right. Five conditions that each filter out most trades leave you with no trades at all.

This is not an academic problem. It is the most common strategy-building mistake on any platform, including CoinQuant. A trader starts with RSI. Adds MACD for confirmation. Adds SMA crossover for trend. Adds ADX for strength. Adds volume for participation. Each condition feels responsible. Together, they produce a strategy that never deploys capital.

Why V3 Worked: Complementary Indicators

V3 combined two indicators that answer fundamentally different questions. SMA(200) asks: is the market trending up or down? RSI(14) asks: is price at an extreme that creates an entry opportunity?

The trend filter keeps you on the right side of the market. The momentum indicator tells you when to act within that trend. Neither duplicates the other. Each earns its place.

The complementary strategy won both of its trades. The single RSI won three out of four but carried significant drawdown risk between entries. Different approaches, different risk profiles, and the two-indicator version delivered comparable returns with cleaner signals.

How to Combine Multiple Indicators Without Overfitting

Redundancy Test: Do these two indicators measure the same thing? If yes, pick one. RSI and Stochastic both measure overbought/oversold conditions. MACD and moving average crossovers both measure trend direction. Stacking indicators from the same category adds noise, not information. V2 proved this: five momentum and trend indicators produced zero signals.

Complementarity Test: A strong combination answers different questions. Trend (SMA, HMA, ADX) asks which direction. Momentum (RSI, MACD) asks if the move is extended. Volume (OBV) asks if there is real participation. Each answers a question the others cannot.

Adding the SMA(200) trend filter to RSI produced a 100% win rate over 2.4 years. Both trades were profitable. The trend filter earned its place.

The Practical Rule

Combine indicators that answer different questions. Never combine multiple indicators that answer the same question. Two complementary indicators outperform five redundant ones. The best strategy is not the one with the most indicators. It is the one with the fewest indicators that still answers all the right questions.

Start simple. Backtest one indicator first. Add a second only if it measures something the first one does not. If you cannot explain what unique question the second indicator answers, do not add it.

Run this backtest free on CoinQuant