Evidence dashboard
Namibia payment journey: Public lists cannot confirm what is enabled for one profile. Use the authenticated Personal Area (PA) as the live source for methods, limits, fees and processing estimates.
| Metric | Question |
|---|---|
| Expectancy | Positive after all costs? |
| Maximum drawdown | Affordable? |
| Trade count | Enough independent observations? |
| Regime split | Works beyond one period? |
| Cost share | Edge survives spread/commission/conversion? |
| Process error | Missed/duplicate/rule breaches? |
Validation ladder
- Written rules.
- Historical out-of-sample.
- Cost/parameter stress.
- Demo forward test.
- Small live execution.
- Independent withdrawal test.
- Scale only after all gates.
Failure signals
| Signal | Response |
|---|---|
| One setting drives result | Reject/redo robustness |
| Few outlier wins | Reduce confidence |
| Live slippage erases edge | Pause |
| Drawdown exceeds limit | Stop |
| Extra deposits hide losses | Recalculate true return |
| Unknown order state | Reconcile before trading |
Capital accounting
Track deposits, withdrawals, rebates and trading P/L separately. Calculate time-weighted or cash-flow-aware performance so adding money is not mistaken for strategy profit.
Operational resilience
Use server-side SL, tickets and restart reconciliation. Local tools failing offline must be part of the test, not treated as an exceptional excuse.
Official source
Transaction history categories · Cost inputs
Frequently asked questions
What proves success?
Robust positive after-cost evidence with acceptable drawdown and execution.
Is a profitable backtest enough?
No. Add out-of-sample, demo and small-live tests.
Can deposits count as performance?
No. Separate cash flows from trading P/L.
When should a strategy stop?
At predefined drawdown, drift or operational-failure limits.