Methodology
v1.0.0 ·
Model
Probabilities come from parametric sampling: each trade wins with the cell win rate (average win R, average loss 1 R), trades are independent, and trades per day are fixed. The interactive simulator historical mode instead bootstraps blocks of imported trades — it does not feed the precomputed tables.
Trading day and daily-loss anchor
The engine simulates day by day: trades are played, and the breach is checked against the floor before the next trade. The exact anchor (00:00 balance, equity, previous close) is the one written in the verified rule tables, taken from the firms official pages.
Grid and seed
Grid: win rates 35% to 60% (step 5), risk per trade 0.25% to 2%, average win 1 / 1.5 / 2 R. Each cell holds 10,000 Monte Carlo runs with the fixed seed 20261001: two executions produce byte-identical numbers (automated test).
Confidence intervals
Every probability carries a 95% Wilson confidence interval. Per-phase marginals derive from prefix chains: P(passing phase k) = P(k-phase chain) / P((k−1)-phase chain).
Limits
A run not completed after 730 simulated days counts as a time-limit failure — this mostly penalizes very low-risk profiles. No slippage, no commissions, no trade correlation, rules applied exactly as verified. Refundable fees are not deducted from the expected cost.
External cross-validation
Upcoming: comparison on identical scenarios against a third-party open-source engine. No cross-validation figures are published yet.