Deconstructing Expectancy in RSI Systems
A trader's equity curve is the direct result of their system's mathematical expectancy. A positive expectancy system makes money over a large series of trades. A negative expectancy system loses money. It is that simple. For RSI-based strategies, calculating expectancy is not an academic exercise. It is the core determinant of viability. The formula is:
Expectancy = (Win Rate * Average Win Size) – (Loss Rate * Average Loss Size)
Let's break this down. Win rate is the percentage of trades that are profitable. Average win is the average profit on winning trades. Loss rate is the percentage of trades that are losers. Average loss is the average loss on losing trades. A system with a 40% win rate can be highly profitable if the average win is a multiple of the average loss. Conversely, a system with a 70% win rate can be a consistent loser if the average win is a fraction of the average loss.
Consider an RSI mean reversion strategy on the 5-minute chart of NQ futures. A trader might short NQ when the 14-period RSI crosses above 75, targeting a reversion to the 20-period EMA. Over 500 trades, the system might have a win rate of 65%. The average winning trade nets 8 points. The average losing trade costs 12 points. The loss rate is 35% (100% - 65%).
- Expectancy = (0.65 * 8) – (0.35 * 12)
- Expectancy = 5.2 – 4.2
- Expectancy = 1.0
This system has a positive expectancy of 1.0 point per trade. Assuming a commission of $5 round turn, the net expectancy is still positive. This is a viable system. The key is the relationship between win rate and risk/reward (R:R). The R:R here is 8/12 or 0.67R. The system's profitability depends on the high win rate compensating for the poor R:R.
Institutional Approach to RSI Backtesting
Proprietary trading firms and quantitative hedge funds approach RSI backtesting with statistical rigor. They do not simply look at a handful of trades. They run simulations on years of historical data, covering various market regimes (bull, bear, range-bound). A typical institutional backtest of an RSI strategy on SPY might involve 10 years of 1-minute data, resulting in hundreds of thousands of trades.
The analysis goes far beyond simple expectancy. Quants look at:
- Profit Factor: Gross Profit / Gross Loss. A value above 2.0 is considered robust.
- Sharpe Ratio: (Average Return - Risk-Free Rate) / Standard Deviation of Returns. Measures risk-adjusted return.
- Maximum Drawdown: The largest peak-to-trough drop in the equity curve. This is a critical measure of risk.
- Monte Carlo Simulation: Running thousands of variations of the trade sequence to assess the probability of different outcomes and the risk of ruin.
For example, a firm testing an RSI divergence strategy on TSLA's daily chart would run a backtest from 2015 to 2025. They would find the optimal RSI period (e.g., 9, 14, 21) and overbought/oversold levels (e.g., 70/30, 80/20). The backtest would account for commissions, slippage (the difference between the expected fill price and the actual fill price), and dividend adjustments. A strategy might look great before costs, but fail once realistic transaction costs are factored in. An automated backtest on CL (crude oil futures) might show a 55% win rate using a 5-period RSI on a 15-minute chart, but with an average slippage of 2 ticks per trade, the strategy's edge evaporates.
Worked Trade Example: RSI Divergence Failure
Not all signals work. It is important to analyze failures to understand a strategy's limitations. Here is an example of a bullish RSI divergence that failed on the 15-minute chart of AAPL.
- Date: January 15, 2026
- Context: AAPL is in a short-term downtrend, making lower lows and lower highs.
- Signal: At 10:30 AM, AAPL makes a new low at $180.50. The 14-period RSI makes a higher low (32.5) compared to the previous price low (RSI was 28.0). This creates bullish divergence, suggesting waning downward momentum.
- Trade: A trader buys 100 shares of AAPL at $180.75, anticipating a reversal.
- Entry: $180.75
- Stop Loss: $179.95 (below the recent low)
- Target: $182.35 (targeting a previous resistance level)
- Position Size: 100 shares
- Risk: $0.80 per share ($180.75 - $179.95), total risk of $80.
- Potential Reward: $1.60 per share ($182.35 - $180.75), total potential reward of $160.
- R:R Ratio: 2:1
Outcome: The divergence fails. After a brief bounce to $181.10, the selling pressure resumes. The underlying downtrend was too strong. An institutional sell program may have been active, absorbing all bids. The price drops through the entry and stops the trader out at $179.95 for an $80 loss. This example shows that even classic signals like RSI divergence are not infallible. They are probabilistic, not deterministic. The broader market context (in this case, a strong downtrend) ultimately dictated the outcome.
Key Takeaways
- Expectancy is the single most important metric for any trading system.
- A high win rate does not guarantee profitability; risk/reward is the other half of the equation.
- Institutional backtesting is exhaustive and accounts for transaction costs, slippage, and various market conditions.
- RSI signals, like all technical indicators, can and do fail. Understanding the context of the failure is critical for system improvement.
- Profit factor and Sharpe ratio are more robust measures of a system's performance than win rate alone.
