The System Quality Number, coined by Van Tharp, rolls three things every trader cares about into one score: how big your average edge is, how steady it is, and how many trades prove it. A high SQN means a strategy wins clearly, wins consistently, and has done it often enough to believe. A low one means the edge is either small, erratic, or barely tested — and the score won't tell you which without a look at the trades underneath.
What it is
SQN measures the quality of a trading system in units of R — the result of each trade expressed as a multiple of the amount you risked on it. A trade that made twice what you risked is +2R; one that hit its stop is −1R. Feed in the R-multiples of every trade and SQN rewards three things at once: a high average R, low variation in that R (consistency), and a large number of trades (evidence). It is a measure of edge quality, not of how much money the system made.
How it works
SQN = (average R ÷ standard deviation of R) × √(number of trades)
The first part — average R over the standard deviation of R — is the strength of the edge relative to how much it wobbles, essentially the same shape as expectancy divided by its own noise. Multiplying by the square root of the trade count is what stops a great-looking result from three trades scoring highly: the evidence has to scale with the claim. Van Tharp's rough bands, judged on roughly 100 trades:
- Below 1.6 — poor.
- 1.6 to 2.0 — average.
- 2.0 to 3.0 — good.
- Above 3.0 — excellent.
A worked example
An EA takes 100 trades. Across them the average result is +0.25R, and the standard deviation of the R-multiples is 1.25R.
SQN = (0.25 ÷ 1.25) × √100 = 0.20 × 10 = 2.0
That is an average system: a real, positive edge of +0.25R per trade, fairly ordinary consistency, backed by a decent sample. Now take a second EA with the same +0.25R average and the same 1.25R spread, but only 25 trades: SQN = (0.25 ÷ 1.25) × √25 = 0.20 × 5 = 1.0. Identical edge, identical consistency, half the score — purely because there is a quarter of the evidence. The square-root term did that, and it is the point of the whole formula.
Why it matters
- It combines edge and consistency. A system with a smaller average R but very steady results can outscore a bigger, wilder edge.
- It is honest about sample size. Two strategies with the same average can score very differently if one has ten times the trades behind it.
- It maps onto position sizing. Van Tharp built SQN partly to judge how hard a system could be pushed — a higher, steadier edge tolerates more.
Common misconceptions
- It needs a decent number of trades. Below about 30 the square-root term and the standard deviation are both unstable, and the score bounces around. Treat SQN on a tiny sample as a guess.
- It isn't a return figure. A high SQN says the edge is real and consistent, not that it made a fortune. A small, steady edge can score well.
- Outliers cut both ways. One enormous winner lifts the average R but also inflates the standard deviation, so it can lower the score rather than raise it.
- The bands assume ~100 trades. On far fewer, the same thresholds read too generously.
The Karnek note
Karnek computes SQN from your live terminal, using the same trades and window as the expectancy and R figures shown beside it, so they reconcile and the sample size is labelled — which matters most for a number that leans on trade count. It reads read-only, and it can never trade the account.
Written and reviewed by the Karnek Research team. Last updated August 2026.
Educational content only - not financial advice. Past performance does not predict future results. Trading carries significant risk.