N
Number of Trades
Total trades in a backtest. Higher N = more statistical confidence in metrics.
WR
Win Rate
% of trades that closed positive.
Exp / Expectancy
Expectancy
Average profit per trade in baht (or %). = (WR × avg_win) + (LR × avg_loss). Most important single number.
MaxDD
Maximum Drawdown
Largest peak-to-trough loss in equity curve. Measures worst-case experience.
P&L
Profit and Loss
Net result of a trade or period.
R / 2R / 3R
Risk Multiple
1R = initial risk on the trade. 2R win = sold at 2× initial risk profit. 3R = 3×.
R:R
Risk-to-Reward Ratio
Expected reward ÷ risk taken. 1:3 means risk $1 to make $3.
Walk-Forward
Walk-Forward Testing
Rolling out-of-sample validation: rules fixed on window 1, tested on window 2, then window shifts forward. Repeated across 6+ windows. Passes if ALL windows are positive — not just the average. Prevents overfitting to one historical period.
Bootstrap CI
Bootstrap Confidence Interval
Non-parametric statistical test: resample the trade population 10,000× with replacement, compute the mean each time, derive the 5th–95th percentile range. If the CI excludes 0, the edge is statistically real. If it includes 0, the result may be noise.
OOS
Out-of-Sample
Test data NOT used during system development. Critical for proving the edge isn't curve-fit.
Train / Test
Train / Test split
Train period (e.g. 1990-2010) is where rules are designed. Test period (2011-2026) validates them.
CAGR
Compound Annual Growth Rate
Annualized return assuming compounding. Buy-and-hold S&P 500 ≈ 10% CAGR.
IC
Information Coefficient
Spearman rank correlation between a signal and forward returns. Measures predictive power. IC=0.10 is meaningful.
Stop / SL
Stop Loss
Order to exit if price hits a predetermined level. Protects capital.
TP
Take Profit
Order to exit at a predetermined gain level.
Slippage
Slippage
Difference between expected fill price and actual fill price. Includes commissions in our backtests (0.3%).