Most traders know their win rate. Far fewer know whether their system actually has an edge worth scaling. Win rate alone is meaningless — a 70% win rate with an average loss three times larger than an average win produces a losing system. This guide covers four metrics — expectancy, profit factor, SQN, and R standard deviation — that together give you an objective, quantified verdict on your trading system.
This guide is for intermediate to advanced traders who have at least 30 completed trades in a single system and want to go beyond basic statistics.
Step 1: Calculate Your System’s Expectancy
Expectancy is the average R-multiple earned per trade. It answers: “On average, how much do I make for every unit of risk I take?”
Formula: Expectancy = (Win Rate × Average Win in R) − (Loss Rate × Average Loss in R)
Example with 100 trades:
- Win rate: 45%
- Average winner: +2.1R
- Average loser: −1.0R
Expectancy = (0.45 × 2.1) − (0.55 × 1.0) = 0.945 − 0.55 = +0.395R per trade
A positive expectancy confirms the system has edge. A system producing +0.20R or more per trade with 50+ trades is worth taking seriously. Anything below +0.10R is marginal and highly sensitive to execution slippage and spread costs. See how to calculate expectancy for a detailed worked example.
Step 2: Calculate Profit Factor
Profit factor compares the total gross profit of all winning trades to the total gross loss of all losing trades.
Formula: Profit Factor = Total Gross Profit / Total Gross Loss
Example:
- Total gross profit across 45 winning trades: $4,720
- Total gross loss across 55 losing trades: $2,890
Profit Factor = 4720 / 2890 = 1.63
Benchmarks:
- Below 1.0 — losing system
- 1.0–1.25 — break-even range, costs will likely make this a loser
- 1.25–1.50 — marginal, needs more data
- 1.50–2.0 — viable system
- Above 2.0 — strong system
Profit factor is sensitive to outliers. A single 20R winner can inflate the number significantly. Cross-reference it with expectancy to confirm the edge is distributed across trades, not concentrated in one or two outliers.
Step 3: Calculate the System Quality Number (SQN)
SQN, developed by Van Tharp, combines edge size and consistency into a single score. It accounts for how reliably the system produces its edge, not just whether the edge exists.
Formula: SQN = (Mean R / Standard Deviation of R) × √(Number of Trades)
Using the same 100-trade sample:
- Mean R: +0.395
- Standard deviation of R: 1.42
- Trade count: 100
SQN = (0.395 / 1.42) × √100 = 0.278 × 10 = 2.78
That scores as “good” on Van Tharp’s scale. To calculate in a spreadsheet: list all R-multiples in one column, use =AVERAGE() for mean R and =STDEV() for standard deviation, then apply the formula.
A critical constraint: SQN inflates with sample size. A score calculated on 30 trades is not comparable to one calculated on 300. Always report the trade count alongside the SQN.
Step 4: Assess Consistency with Standard Deviation of R
The standard deviation of your R-multiples reveals whether your system’s results cluster tightly or scatter widely. A system with mean R of +0.40 and a standard deviation of 0.8 is far more consistent than one with the same mean R and a standard deviation of 3.5.
Interpreting standard deviation of R:
- Below 1.0 — very consistent, tight risk control
- 1.0–2.0 — normal range for trend-following and swing systems
- 2.0–3.0 — high variance, likely includes runners with wide exits
- Above 3.0 — results are dominated by outliers; system may not be reliably repeatable
If your standard deviation of R is above 2.5, examine your largest 5 winners. If removing them drops your expectancy below +0.10R, your system depends on catching occasional outliers rather than a repeatable edge. That is a different system than it appears to be on paper. Review your equity curve alongside this data to visualise the variance in real time.
Step 5: Interpret and Act on the Composite Score
Combine all four metrics into a structured scorecard before making any decision about live trading, position sizing, or system retirement.
| Metric | Your Value | Benchmark |
|---|---|---|
| Expectancy | +0.395R | Minimum +0.20R |
| Profit Factor | 1.63 | Minimum 1.50 |
| SQN | 2.78 | Minimum 2.0 |
| Std Dev of R | 1.42 | Prefer below 2.0 |
Decision rules:
- All four metrics meet benchmarks on 100+ trades — trade live, scale gradually
- Two or more metrics below benchmark — extend testing, do not increase size
- SQN below 1.6 after 100 trades — treat as a losing system until evidence changes
- High profit factor but low SQN — results concentrated in outliers, reduce size or filter setups
Break down this scorecard by setup tag to identify which setups drive your edge. A composite SQN of 2.5 can easily contain one setup at 3.8 and another at 1.2. Trade more of the former, less of the latter.
Pro Tips
- Calculate SQN on a rolling 50-trade window, not just total history. A deteriorating rolling SQN warns you that edge is eroding before your account balance reflects it.
- Never compare SQN scores calculated on different sample sizes. Normalise by reporting SQN per 100 trades: divide by √N then multiply by 10.
- Profit factor above 3.0 on fewer than 50 trades is almost always noise. Require at least 100 trades before trusting any metric above its upper benchmark.
- Segment metrics by session (London, New York, Asian) — many systems are genuinely strong in one session and nearly flat in another. Your composite score is hiding that split.
- After a losing streak, recalculate SQN on your last 30 trades only. If it has dropped below 1.5, reduce size immediately until it recovers — this is a mechanical rule, not discretion.
Common Mistakes to Avoid
-
Using win rate as the primary quality metric. Win rate without average win-to-loss ratio is meaningless. A system at 35% win rate and 3.0R average winners outperforms a 65% win rate system with 0.7R average winners — always evaluate the full distribution.
-
Calculating metrics on demo data only. Execution quality, slippage, and spread widening during news events are absent from demo results. Require at least 30 live trades in the metric calculation before trusting any score.
-
Ignoring the standard deviation of R. Traders optimise for expectancy and SQN but ignore how volatile their R-distribution is. High variance systems require larger drawdown tolerance and more capital — sizing them as if they were low-variance systems causes account blow-ups.
-
Pooling data from multiple systems or setups. Mixing an ICT order block setup with a news fade strategy into a single dataset produces a composite score that represents neither. Score each system independently.
-
Treating SQN as a fixed property of the system. Edge degrades. Market conditions change. Recalculate every 30 trades, and set a hard rule: if rolling SQN drops below 1.5 for two consecutive 30-trade windows, reduce size by 50% until it recovers.
How PipJournal Helps
PipJournal automatically calculates expectancy, profit factor, and R-multiple distribution for every system and setup tag in your journal — no spreadsheet required. The analytics dashboard lets you filter by setup, session, or date range and immediately see how each slice scores across all four quality metrics. When your rolling performance shifts, PipJournal’s AI co-pilot surfaces the change and identifies whether the degradation is isolated to a specific setup or affecting your system as a whole. With every trade logged and tagged, you always have a statistically current scorecard — not a snapshot from three months ago.
People Also Ask
What is a good SQN score for a forex trading system?
Van Tharp's scale rates 1.6–1.9 as below average, 2.0–2.4 as average, 2.5–2.9 as good, 3.0–5.0 as excellent, and above 5.0 as superb. Most retail systems with genuine edge score between 1.8 and 2.8. Be skeptical of any system scoring above 4.0 on fewer than 100 trades.
How many trades do I need to calculate a reliable SQN?
Van Tharp recommended a minimum of 30 trades, but 100 or more significantly reduces sampling error. For short-term systems trading 5 or more times per week, aim for at least 3 months of live or demo data.
What is the difference between expectancy and profit factor?
Expectancy measures average profit per trade in R-multiples (e.g., +0.25R per trade), while profit factor is the ratio of gross wins to gross losses (e.g., 1.8). Both measure edge but profit factor is more sensitive to outlier trades, while expectancy normalises by trade count.
Can I calculate SQN in a spreadsheet?
Yes. List each trade's R-multiple, calculate the average (mean R) and standard deviation of R, then apply the formula — SQN = (mean R / standard deviation of R) multiplied by the square root of the trade count.
Should I calculate these metrics separately for different setups?
Absolutely. A composite score across all setups can mask a weak setup dragging down a strong one. Filter by setup tag and score each independently to identify which setups deserve more size and which should be cut.
