Most traders lose money. The standard explanation is that they lack discipline, emotion interferes, or they chase hot tips. These explanations are not wrong, but they miss the underlying cause: most traders do not have an edge. They have a strategy — a set of entry and exit rules — but a strategy is not the same thing as an edge.
This lesson defines what an edge actually is, how to measure it, and why the distinction matters before you write a single line of backtesting code.
The Expectancy Formula
Expectancy is the expected return per dollar risked. Positive expectancy means the system makes money over time; zero expectancy means breakeven before costs; negative expectancy means losses compound until the account is gone.
Where:
- $W$ = win rate (fraction of trades that are winners)
- $A_w$ = average win size as a multiple of risk
- $L$ = loss rate (1 − W)
- $A_l$ = average loss size as a multiple of risk
The last row is instructive: a 60% win rate with an average win smaller than the average loss produces negative expectancy. This is the trap of most discretionary traders — they cut winners short and let losers run, producing a win rate that looks acceptable but an expectancy that is negative.
What Makes Something an Edge vs. a Strategy
A strategy is a set of rules. An edge is evidence that those rules produce positive expectancy in a defined market condition over a large enough sample.
The definition implies three things:
1. Statistical, not anecdotal. Ten winning trades is not evidence of an edge. The minimum sample to begin trusting results depends on win rate variance, but 100–200 trades is a reasonable starting point. Fewer trades and you cannot separate luck from skill.
2. Condition-specific. An edge in trending markets may be negative in ranging markets. An edge on large-cap tech may not exist on small-cap biotech. Defining the conditions where your system applies is as important as the rules themselves.
3. Measurable. If you cannot calculate the expectancy of your system from historical data, you do not know whether you have an edge. Gut feel and pattern recognition are inputs into forming a hypothesis — not evidence of an edge.
Why Markets Have Edges at All
Markets would have no exploitable edges if every participant were rational, equally informed, and trading without friction or behavioral bias. None of these conditions hold.
Edges exist because:
- Behavioral biases are systematic. Loss aversion, recency bias, and herding are predictable and recurring. A strategy that exploits these biases can extract a consistent premium.
- Market structure creates friction. Institutions must execute large orders and cannot exit immediately. This creates predictable order flow around certain price levels and times.
- Liquidity cycles. Markets cycle between trending and mean-reverting conditions based on the balance of informed vs. uninformed flow. Systems calibrated to one regime can extract edge during that regime.
None of these sources are infinite. Edges erode as more capital exploits them. A systematic trader tracks edge decay as carefully as edge discovery.
The Two Types of Edge
Statistical edge is quantifiable from price and volume data alone. Examples: mean reversion after a defined percentage gap, momentum continuation after a defined breakout structure, volatility expansion after a defined compression. You can measure it with a backtest, calculate the expectancy, and monitor for decay with ongoing live results. Statistical edge is the primary focus of this course.
Informational edge comes from knowing something the market does not yet reflect in price — earlier access to fundamental data, proprietary survey data, supply chain intelligence. For retail systematic traders, informational edge is largely inaccessible and often legally constrained. The systematic approach focuses on statistical edge, which is available to everyone with price data and a backtesting engine.
The Lesson in One Line
A strategy without measured expectancy is a guess. An edge is a tested, quantified reason why a specific set of conditions produces a favorable risk-adjusted return over a defined sample of market history — and continues to do so out-of-sample.
The remaining lessons in this course build the machinery to find, test, and operate systems with genuine edge.