Home › FX Trading School › Pro
Expectancy and Statistical Edge Explained
Trading expectancy is the number that tells you, honestly, whether a strategy is worth trading at all — the average amount you can expect to win or lose per trade once you strip away hope, hindsight and small-sample luck. This lesson is Module 14 of the PipTax FX Trading School, and it builds directly on risk-reward ratio and position sizing from earlier modules — if you haven't covered those yet, go back before continuing, because expectancy is where they combine into a single decision rule.
What Trading Expectancy Actually Measures
Expectancy answers one question: over a large number of trades, what's my average result per trade, expressed in the same unit I use for risk (pips, £, or R-multiples)?
It is not:
- Win rate alone — a 70% win rate can still lose money if losers are large.
- Risk-reward alone — a 1:3 risk-reward system can still lose money if the win rate is too low.
- A backtest headline number — profit factor and total return hide how the average trade behaves.
Expectancy forces win rate and risk-reward to be judged together, which is why professional traders quote it in preference to either number in isolation. It's also the number that determines, mathematically, whether adding more trades makes you richer or poorer over time — a positive expectancy strategy improves with volume; a negative one gets worse.
The Expectancy Formula
The standard formula is:
Expectancy = (Win rate × Average win) − (Loss rate × Average loss)
Worked example, using R-multiples (R = your risk per trade):
| Input | Value | |---|---| | Win rate | 40% | | Average win | 2.5R | | Loss rate | 60% | | Average loss | 1R |
Expectancy = (0.40 × 2.5) − (0.60 × 1) = 1.0 − 0.6 = +0.4R per trade
That means, on average and over a large sample, this system nets 0.4R per trade. Risk £100 per trade consistently and the long-run average is roughly £40 per trade — before costs, which is the crucial caveat covered next.
Why Real Trading Costs Change the Answer
Backtests and hypothetical examples rarely include full transaction costs. Live trading does. Every one of these needs to be subtracted from your average win and added to your average loss before expectancy means anything:
- Spread — paid on every entry, varies by instrument and broker
- Commission — charged per lot on some account types, e.g. raw/ECN-style accounts
- Swap/rollover — accrues on positions held overnight, can be positive or negative
- Slippage — the gap between requested and filled price, worse in fast markets
A strategy showing +0.4R on paper can easily fall to +0.1R or below once real costs are applied, especially for shorter-term or high-frequency approaches where the average win is small relative to the spread. This is precisely why PipTax builds the cost audit tool — run your actual instrument and trade frequency through /audit.html to see the real cost drag, and compare it against /cost-impact.html to understand how much of your edge costs alone can consume.
Sample Size: Why Small Runs Lie
Expectancy is a long-run average, and small samples are dominated by variance, not edge.
- Under 30 trades: essentially meaningless — a lucky or unlucky streak can flip the sign entirely.
- 30-100 trades: directional signal, but treat with caution.
- 100-200 trades: reasonable confidence, still verify across different conditions.
- 200-300+ trades: solid basis for a live decision, ideally spanning trending and ranging markets.
A common mistake is calculating expectancy from a single strong month and assuming it holds. Markets change regime, and a strategy tuned to trending conditions can show a completely different — sometimes negative — expectancy in a ranging market. Log every trade with entry, exit, cost, and market condition so you can segment the data later.
Comparing Win Rate and Risk-Reward Trade-Offs
Two strategies can produce identical expectancy through very different paths:
| Strategy | Win rate | Avg win | Avg loss | Expectancy | |---|---|---|---|---| | A (high win rate) | 65% | 1R | 1R | +0.30R | | B (trend-following) | 35% | 2.7R | 1R | +0.295R |
Strategy A will feel psychologically easier — frequent small wins — but is more exposed to a run of costs eating the thin margin per trade. Strategy B will feel harder — long losing streaks are normal — but each win does more work. Neither is objectively better; the right choice depends on your temperament, drawdown tolerance and how consistently you can execute the plan under pressure. This is a judgement call for position sizing and psychology modules, not something expectancy alone decides.
Building an Expectancy Tracker You Can Actually Use
You don't need software beyond a spreadsheet. Track, per trade:
1. Instrument, direction, entry/exit price 2. Risk in R (or £), and the actual result in R 3. Spread and commission paid (pull real figures from your broker statement — for example, check how Pepperstone's MetaTrader account statements itemise commission separately from spread, or how IG's platform reports costs on closed positions) 4. Swap charged, if held overnight 5. Market condition tag (trending/ranging/high-impact news)
From this you can compute rolling expectancy over the last 50, 100 and 200 trades, and see immediately if it's drifting. Recalculate whenever you switch broker or account type, since spread and commission structures differ and directly change the average loss input — the /brokers/index.html comparison pages and /audit.html tool exist specifically so you're using real, current figures rather than assumptions.
Conclusion: Making Expectancy Part of Your Process
Trading expectancy is not a vanity metric — it's the pass/fail test a strategy has to clear before you risk real money on it, and it only means anything once real spread, commission and slippage are included. Calculate it from a large enough sample, compare it honestly against the cost drag your broker actually charges, and revisit it whenever conditions or accounts change. Remember that even a strategy with a genuinely positive statistical edge can still lose money in the short run — that's the nature of variance, not a sign the edge is fake. Treat expectancy as one input alongside position sizing and risk management, run your numbers through PipTax's /audit.html and /methodology.html pages for cost clarity, and keep building the discipline that turns a paper edge into a survivable, tradeable one.
Key takeaways
- Trading expectancy tells you the average amount you win or lose per trade over the long run — it's the single number that decides whether a strategy is worth trading
- The formula is: (Win rate x Average win) minus (Loss rate x Average loss); anything above zero is a positive edge before costs
- Spread, commission and swap must be subtracted from every trade in the sample, not added as an afterthought — a strategy with positive edge on paper can be flat or negative after real costs
- You need a large enough sample (100+ trades minimum, ideally 200-300) before you can trust an expectancy figure — small samples are noise, not edge
- A high win rate with poor risk-reward can have the same or worse expectancy than a low win rate with strong risk-reward — the two numbers must be judged together
- Recalculate expectancy whenever you change broker, instrument, or timeframe, because execution costs and volatility shift the inputs
Frequently asked questions
- What is a 'good' expectancy value?
- There's no universal number because it depends on your risk unit and trade frequency. What matters is that expectancy is reliably positive after real trading costs, and that the strategy's drawdowns are survivable given your position sizing. A small positive expectancy traded consistently over hundreds of trades can be perfectly viable; a large expectancy from a 20-trade sample usually isn't real.
- Can a strategy with a low win rate still have positive expectancy?
- Yes. Many trend-following systems win less than half the time but hold winners far longer than losers, so the average win dwarfs the average loss. Expectancy depends on the balance between win rate and risk-reward, not on win rate alone.
- How many trades do I need before I trust my expectancy number?
- As a rough guide, treat anything under 100 trades as provisional. Random variance dominates small samples, so a strong result over 30 trades tells you very little. Aim for 200-300 trades, ideally across different market conditions, before drawing firm conclusions.
- Do spreads and commissions really change expectancy that much?
- Often yes, especially for higher-frequency or scalping strategies where the average win or loss is small relative to the spread. A strategy that looks profitable on a backtest with zero costs can turn negative once realistic spread, commission and swap are applied per trade. Always test with live cost assumptions from the cost tool, not backtester defaults.
- How does expectancy relate to position sizing and risk of ruin?
- Expectancy tells you if a strategy is worth trading at all; position sizing tells you how much to risk per trade given that edge. A positive-expectancy system can still be ruined by oversized bets during a losing streak, so the two concepts work together, not as substitutes for each other.