System Quality Number (SQN)

System Quality Number (SQN), developed by Van Tharp and published in Trade Your Way to Financial Freedom (2nd ed., 2006), answers a question no single metric can: is this trading system genuinely good, or was the outcome luck? By combining average expectancy, the consistency of results, and the size of the trade sample into one score, SQN gives systematic traders a statistically grounded benchmark for evaluating their edge — and knowing when the evidence is too thin to trust.

Formula & Calculation

SQN = (Expectancy / StdDev of R) × √N

Where:

• Expectancy = Mean R-multiple per trade (average P&L divided by initial risk per trade)
• StdDev of R = Standard deviation of all R-multiples in the sample
• N = Number of trades
• √N = Square root of trade count

The ratio (Expectancy / StdDev of R) is essentially a signal-to-noise ratio for the system — how large the average gain is relative to the volatility of outcomes. Multiplying by √N rewards systems that have produced that signal consistently across many trades. A high expectancy means little if results are chaotic; SQN penalizes that chaos through the StdDev denominator.

To calculate R-multiples: divide each trade’s net P&L by the initial dollar risk on that trade. If a trade risked $400 and returned $600, the R-multiple is +1.5R. If it lost $200, it is -0.5R. Collect these across all trades, compute the mean and standard deviation, then apply the formula.

Benchmarks

Van Tharp’s published rating scale (from Trade Your Way to Financial Freedom, 2006):

Level	Range	What It Means
Holy Grail	7.0+	Theoretical upper bound; practically unachievable at scale
Superb	5.0 – 7.0	Exceptional consistency and edge; rare in live trading
Excellent	3.0 – 5.0	Strong, reliable system worth scaling
Good	2.5 – 2.9	Solid edge with acceptable variance
Average	2.0 – 2.4	Marginal edge; little room for execution error
Below Average	1.6 – 1.9	Insufficient edge; likely fails under real drawdown
Poor	Below 1.6	No reliable edge detected

These ratings assume a statistically meaningful sample (100+ trades). Apply them to small samples at your peril.

Practical Example

A futures trader runs 100 ES mini trades over 6 months, risking exactly 1R = $500 per trade (2 ES points × $250/point). After logging all trades:

• Mean R-multiple per trade: +0.35R
• Standard deviation of R-multiples: 1.05R

SQN = (0.35 / 1.05) × √100 = 0.333 × 10 = 3.33

This places the system in the Excellent tier — a meaningful result after 100 trades.

Now consider the same system evaluated after only 25 trades with identical stats:

SQN = (0.35 / 1.05) × √25 = 0.333 × 5 = 1.67

That reads as Below Average — a score that would lead most traders to abandon a viable strategy. The edge is identical; only the sample size changed. This is the sample size trap: at N=30, √N ≈ 5.5; at N=100, √N = 10; at N=400, √N = 20. A system needs enough trades to let the multiplier reflect genuine evidence, not just inflate a small lucky streak.

How to Track SQN

1. Record every trade as an R-multiple — divide net P&L by the initial dollar risk on that trade before position sizing is applied.
2. Accumulate at least 30 trades before calculating; target 100+ for any system evaluation that will influence live capital decisions.
3. Calculate mean and standard deviation of the full R-multiple distribution using a spreadsheet or journal software.
4. Apply the formula — (mean R / StdDev R) × √N — and note the N alongside the score every time you report it.
5. Track SQN over time as new trades are added; a stabilizing score is itself a signal that the sample is approaching statistical validity.

How to Improve SQN

1. Reduce outlier losses — trades that close at -3R or worse inflate StdDev disproportionately. Moving stops to breakeven after 1R of favorable movement caps the left tail without sacrificing upside.
2. Filter low-conviction setups — removing trades that don’t meet all entry criteria raises mean expectancy; even one fewer bad trade per 20 meaningfully shifts the ratio.
3. Increase frequency on proven setups — if the expectancy and StdDev are already favorable, more trades directly improve SQN through the √N factor. Scan more instruments or lower the timeframe if the setup quality holds.
4. Normalize position sizing to 1R before calculating — if your risk per trade varies widely, R-multiples become distorted and StdDev rises artificially.
5. Audit recurring -2R+ losses — these outliers are usually execution failures (late entries, missed stops) rather than system failures. Fixing execution improves SQN faster than changing entry logic.

Common Mistakes

1. Evaluating SQN on fewer than 30 trades — the √N multiplier makes a 20-trade SQN of 4.0 nearly meaningless. Always show N alongside any SQN figure you report or act on.
2. Using dollar P&L instead of R-multiples — dollar-based SQN is distorted by inconsistent position sizing and cannot be compared across instruments or time periods when risk varied.
3. Abandoning a system after a low reading on small N — the example above (3.33 at N=100, 1.67 at N=25) is not hypothetical. Many traders quit systems that would have validated at 100 trades.
4. Ignoring the StdDev component — a system with 0.8R mean expectancy but StdDev of 3.0R scores lower than one with 0.4R mean and StdDev of 0.8R. Consistency matters as much as magnitude.
5. Comparing SQN scores from different sample sizes — an SQN of 3.5 at N=50 and 3.5 at N=200 are not equivalent signals of edge quality. The larger sample is substantially more meaningful.

How JournalPlus Calculates SQN

JournalPlus calculates SQN automatically from your trade log, converting each closed trade into an R-multiple using the initial stop distance you logged at entry. The analytics dashboard displays your current SQN score alongside the trade count used in the calculation, so you always see N in context. As you add trades, the score updates in real time, and the performance charts let you track SQN over rolling windows (30, 50, 100 trades) to see how it stabilizes. You can filter by setup type, instrument, or date range to calculate SQN for a specific strategy subset — useful for comparing two setups running simultaneously without mixing their distributions.

For deeper analysis, export your R-multiple distribution to CSV and inspect the full histogram of outcomes, which reveals whether a mediocre SQN is driven by a few outlier losses or a persistently low expectancy — two problems that require very different fixes.

Internal links: Expectancy · Profit Factor · Win Rate · Risk-Reward Ratio · Sharpe Ratio