Trading System Expectancy and Edge Math 2026: How to Know if Your CFD Strategy Actually Works

Every trader has at some point asked the question that separates professionals from gamblers: is my strategy actually working, or am I just lucky? Answering that question requires more than intuition. It requires measuring the strategy’s performance across enough trades to strip out noise, and it requires four distinct metrics — expectancy, R-multiples, profit factor, and the Sharpe ratio — each answering a different question.

This 2026 guide walks through each metric with worked examples, shows how the metrics fit together, and gives you the exact spreadsheet formulas to compute them on your UZFX trade history. By the end, you will be able to look at any strategy and answer: does it have edge, is the edge large enough to matter, and should I be risking this much on it?

Editor’s research note (2026-10-17): All worked examples use a $10,000 account with 1:500 leverage on EUR/USD. Contract specifications vary between brokers — verify on the UZFX product list. All formulas are broker-agnostic and work the same way on any CFD platform.

Why One Metric is Not Enough

A strategy with a 50% win rate and 1:1 reward-to-risk has positive expectancy. A strategy with a 50% win rate and 1:1 reward-to-risk has the same expectancy whether positions are 0.01 lots or 0.50 lots — but the second produces 50x the volatility of the first. Expectancy alone does not tell you whether the size is right.

A strategy with a 90% win rate and 0.05R winners against 5R losers has positive expectancy over 20 trades — but one loss wipes out 20 wins, and the required 21 trades to return to breakeven can take months. Win rate alone does not tell you whether the trade risk is survivable.

Each metric covers what the others miss:

Used together, they answer: does the strategy make money (expectancy), how much does it earn per trade (R-multiples), how robust is the edge (profit factor), how much volatility am I paying for that edge (Sharpe), and how much should I risk (Kelly).

Metric 1: Expectancy

Formula: Expectancy = (Win% × Average Win) − (Loss% × Average Loss)

Where Win% + Loss% = 1.0 (ignore breakeven trades or treat them as 0).

Worked example 1: A break-out strategy with 40% win rate

• Trades: 100
• Winners: 40 × average $180 profit = $7,200 gross profit
• Losers: 60 × average $60 loss = $3,600 gross loss
• Net profit: $3,600
• Expectancy: $36 per trade

Positive expectancy. The strategy wins money over the sample, but the volatility is high — you can be down $300 to $500 at any point during the run of losing trades.

Worked example 2: A mean-reversion strategy with 65% win rate

• Trades: 100
• Winners: 65 × average $40 profit = $2,600 gross profit
• Losers: 35 × average $55 loss = $1,925 gross loss
• Net profit: $675
• Expectancy: $6.75 per trade

Positive expectancy, but tiny. The strategy wins almost every trade, but each winning trade pays barely enough to cover the losses. Over 100 trades, you made $675 — barely above the noise. Over 500 trades, you might have made $3,375, but a single bad month of trend following could erase it.

Interpretation:

• Expectancy > $0.10R: strong strategy, likely profitable at scale
• Expectancy 0.05R - 0.10R: viable strategy, needs discipline
• Expectancy < 0.05R: borderline — costs and volatility will eat most of the edge
• Expectancy < 0: strategy is broken or not edge-positive

Metric 2: R-Multiples

Formula: R = Trade Profit ÷ Initial Risk

Initial risk is the distance from entry to stop in currency units. A trade on 0.10 lots EUR/USD with a 20-pip stop has an initial risk of $20. A winner closing at $40 profit has R = $40 / $20 = 2R. A loser closing at the stop has R = -1R by definition.

The point of R-multiples is that they let you compare strategies independent of position size. A 0.01 lot strategy with a 20-pip stop and a 0.50 lot strategy with a 1,000-pip stop — if both produce +1.5R average, they are functionally equivalent. R-multiples strip out the sizing decision so you can see the pure strategy performance.

Worked example 3: R-multiple distribution

Consider 100 trades:

• 30 trades closed at +2R (big winners)
• 30 trades closed at +1R (moderate winners)
• 10 trades closed at 0R (breakeven)
• 20 trades closed at -0.5R (small losers)
• 10 trades closed at -1R (stop-outs)

Average R: (30×2 + 30×1 + 10×0 + 20×(-0.5) + 10×(-1)) ÷ 100 = (60 + 30 + 0 - 10 - 10) ÷ 100 = 70 ÷ 100 = +0.7R per trade

Expectancy in R: +0.7R per trade. Excellent — this is what a professionally tuned system looks like.

The R-multiple histogram tells you more than the average. A strategy with 90% +0.2R trades and 10% -2R trades has average R of +0R — breakeven. But it is a different strategy from one with 30% +1R trades and 70% -0.3R trades (also average 0R, but smoother). The first has fat tails to the downside; the second is smooth but has no winners big enough to make the risk worthwhile.

Metric 3: Profit Factor

Formula: Profit Factor = Total Gross Profit ÷ Total Gross Loss

Worked examples:

Benchmarks:

• Profit factor < 1.0: the strategy loses money. Do not run it.
• 1.0 < Profit factor < 1.3: barely profitable. Costs, slippage and spread widening will eat most of the edge.
• 1.3 - 1.5: viable retail strategy. Enough edge to survive normal volatility.
• 1.5 - 2.0: strong. Real money-making system.
• 2.0 - 3.0: very strong. Suspicious if from a small backtest — likely overfit.

• 3.0: almost certainly overfit or backtested on data leak. Do not trust without out-of-sample validation.

Sample size matters. A profit factor of 2.0 on 20 trades is meaningless — the confidence interval is wide. A profit factor of 2.0 on 300 trades is meaningful. As a rule of thumb, gather at least 100 trades before drawing conclusions from any single backtest.

Profit factor is more useful than win rate. Two strategies both 50% win rate:

• Strategy A: wins pay 3R, losses pay 1R → profit factor 3.0
• Strategy B: wins pay 0.5R, losses pay 3R → profit factor 0.17

Same win rate. Completely different strategies. Profit factor collapses both win rate and reward-to-risk into one number.

Metric 4: Sharpe Ratio

Formula: Sharpe Ratio = (Return − Risk-Free Rate) ÷ Volatility of Return

For a strategy with monthly returns:

• Annualised return: 24% (mean 2% per month × 12)
• Risk-free rate (10-year Treasury, 2026): 4.5%
• Annualised volatility: 30%

Sharpe = (24% - 4.5%) ÷ 30% = 0.65

Benchmarks:

• Sharpe < 0.5: poor. Not enough return per unit of risk to justify the trading time.
• 0.5 - 1.0: acceptable. Many retail strategies land here.
• 1.0 - 1.5: strong. Professional-level performance.
• 1.5 - 2.0: excellent. Institutional-level. Rare for retail.

• 2.0: extraordinary. Check for overfitting or survivorship bias.

Why Sharpe matters: A strategy with 30% annual return and 60% volatility has a Sharpe of 0.42 (after 4.5% risk-free). A strategy with 15% annual return and 20% volatility has a Sharpe of 0.53 — better despite half the raw return. Sharpe captures the trade-off between return and pain.

Sharpe has a subtle weakness. It only uses two moments of the return distribution (mean and standard deviation). A strategy with hidden negative skew — many small wins and occasional catastrophic losses — can have a respectable Sharpe ratio while being operationally dangerous. Always look at the return distribution alongside Sharpe.

Metric 5: Kelly Fraction

Formula: Kelly fraction = (Win% × Average Win R) − (Loss% × Average Loss R) ÷ Average Loss R

Or equivalently: Kelly = Edge ÷ Average Loss, where Edge is per-trade expectancy in R-multiples.

Worked example from above (R-multiple example 3):

• Expectancy in R: +0.7R per trade
• Average loss in R: 1.0R (all losers close at -1R by definition)
• Kelly fraction: 0.7 ÷ 1.0 = 0.70 = 70% of account

Yes, that is what the math says. And no, that is not what you should do.

The reality of Kelly sizing:

• Full Kelly (100% of the fraction) maximises long-term geometric growth but produces extreme volatility. Drawdowns of 40-60% are common before peak equity. Most traders psychologically break before the math realises.
• Half-Kelly (50% of the fraction) reduces volatility by half and loses only 25% of the theoretical long-term growth rate. This is what most professional traders use.
• Quarter-Kelly (25% of the fraction) is very conservative. Almost no loss of growth rate but a much smoother equity curve.

For the example above, half-Kelly = 35% of account per trade. That is still a lot — most practitioners would use quarter-Kelly (17.5%) or a fixed 1-2% risk rule (which is close to quarter-Kelly for a strategy with 0.7R expectancy).

The relationship between risk per trade and Kelly:

For most retail strategies with expectancy between +0.3R and +0.7R, a 1% risk rule is close to quarter-Kellly or less — which is a defensible, low-volatility choice.

Putting It All Together: Scoring Your Strategy

Use this rubric to score your strategy before running it with real money:

A strategy scoring “Strong” on 4 or more criteria is production-ready. A strategy scoring “Poor” on any criterion should not be traded with real capital until that criterion is fixed.

Tracking Metrics on UZFX

UZFX’s Web Terminal and mobile apps export closed trade history with the fields you need: entry, exit, entry time, exit time, position size, entry price, exit price, P&L in currency, swap charges, and commission. To compute the four metrics:

1. Export the trade history CSV from the Web Terminal.
2. Add a risk column — either from a trading journal (pre-trade stop distance in currency) or by computing Position size × (Entry − Stop) for historical stops.
3. Compute per-trade R: Trade P&L ÷ Risk.
4. Aggregate: expectancy = mean of R; profit factor = sum of positive R ÷ sum of absolute negative R; Sharpe (using monthly returns aggregated from daily close).

Sample spreadsheet formulas:

Expectancy (R) = AVERAGE(R_column)
Profit Factor = SUMIF(R_column,">0") / ABS(SUMIF(R_column,"<0"))
Win Rate = COUNTIF(R_column,">0") / COUNT(R_column)
Sharpe (annual, monthly) = (AVERAGE(monthly_returns) - RiskFree/12) / STDEV(monthly_returns) * SQRT(12)

For the full workflow with position sizing, read our position sizing and 1% rule guide and for recovering from a bad month, the drawdown recovery strategy.

What Kills Good Strategies

1. Small sample size. 20 trades is not enough. A +0.7R average on 20 trades has a 95% confidence interval that spans -0.3R to +1.7R. Wait for 100+.
2. In-sample overfitting. If you tune the strategy on the same data you use to test it, you are not measuring edge — you are measuring noise. Split the data: develop on the first 60%, test on the last 40%.
3. Not accounting for costs. Slippage, spread widening on entry, swap on overnight positions — a strategy with +0.1R expectancy before costs might be -0.05R after.
4. Survivorship bias. If you only count the trades you would have taken, you are biased. Include the trades you passed on.
5. Assuming one timeframe = one strategy. A strategy that works on daily charts might not work on 5-minute charts. Test on the timeframe you actually trade.
6. Ignoring regime shifts. EUR/USD trend-following strategies perform differently in a USD-driven macro regime than in an Asia-driven divergence regime. Check the regime of your backtest period against the regime of the next month.

Final Verdict

Metrics are not a substitute for discipline, but they are the only tool that tells you whether your discipline is applied to something that works. Expectancy tells you whether the strategy makes money. R-multiples tell you how much. Profit factor tells you how robust the edge is. Sharpe tells you how much volatility you are paying for the edge. Kelly tells you how much you should risk.

Run every strategy through all five metrics before risking real capital. A strategy that scores “Viable” or better on at least four of the five is worth running. A strategy that fails on more than one is not.

The best traders do not have a secret strategy. They have a measurable edge, and they measure it constantly. Open a UZFX demo account (60024310, $100,000 virtual), run your strategy for 100+ trades, compute the metrics, and only then decide whether it is worth running live.

Risk Disclaimer

Trading forex and CFDs carries a high level of risk and may not be suitable for all investors. You could lose more than your initial deposit. Leverage amplifies both gains and losses — UZFX offers leverage up to 1:500 on forex pairs, and every asset class carries its own margin requirements. The expectancy, R-multiple, profit factor, Sharpe ratio, and Kelly calculations in this article are mathematical tools, not guarantees. All worked examples use illustrative data and do not reflect live market conditions. Kelly sizing assumes known probabilities that most retail strategies do not have with statistical confidence. Backtests assume conditions do not change — real markets do change. Past performance is not indicative of future results. This article is for information and education only and does not constitute investment advice. Verify current trading conditions and regulatory details on the official UZFX website and the ASIC professional registers before opening an account, and trade responsibly with capital you can afford to lose.

Recommended Broker: Visit UZFX Official Website