# Trading System Evaluation Report

**Dataset:** Discretionary ICT Ledger

**Source / provenance:** Discretionary ICT BTC/USD futures - Preloaded discretionary ICT ledger (22 trades)

**Generated:** 2026-10-23T11:10:58.365Z

**Overall verdict:** GO (conditional - small sample)

> This is an automated audit. It is educational, not financial advice.

## Executive Summary

| Metric | Value |
|---|---|
| Initial Equity | 100.00 |
| Final Equity | 567.79 |
| Net Profit | 467.78 |
| Total Trades | 22 |
| Win Rate | 68.18% |
| Loss Rate | 18.18% |
| Profit Factor | 15.46 |
| Payoff Ratio | 4.12 |
| Expectancy (per trade) | 21.26 |
| Expectancy (R) | 1.95 |

## Edge

- Gross profit: 500.12
- Gross loss: 32.34
- Win rate: 68.18% over 15 wins / 4 losses / 3 breakeven
- Payoff ratio (avg win / avg loss): 4.12
- Expectancy per trade: 21.26 (1.95 in R)
- Longest win streak: 5
- Longest loss streak: 1

## Risk & Drawdown

| Metric | Value |
|---|---|
| Max Drawdown | 16.79 |
| Max Drawdown % | 5.04% |
| Recovery Factor | 27.86 |
| Longest loss streak (equity %) | 5.04% |
| Est. consecutive loss (to halve equity) | 21 trades |
| VaR (95%) | 11.01 |
| CVaR (95%) | 16.79 |
| Avg recovery | 1.0 trades |

## Robustness

| Metric | Value |
|---|---|
| Sharpe | 4.99 |
| Sortino | 11.37 |
| SQN | 4.51 |
| Annualized Return | 127842.00% |
| Calmar | 25349.87 |
| Ulcer Index | 0.02 |
| Martin Ratio | 71093.78 |
| Gain-to-Pain | 14.46 |
| Risk of Ruin (Monte Carlo) | 0.00% |
| Risk of Ruin (Balsara est.) | 0.12% |

## Monte Carlo

- Paths: 50
- Terminal equity median: 557.55
- Terminal equity 5th/95th pct: 318.19 / 733.85
- Terminal equity stdev: 126.89
- Profitable paths: 100%
- Paths in drawdown: 98%

## Deployment Audit - Go / No-Go Criteria

| Criterion | Target | Actual | Status |
|---|---|---|---|
| Profit Factor | 1.5 - 2.0 | 15.46 | PASS |
| Recovery Factor | > 2.0 | 27.86 | PASS |
| Consecutive Loss | < 50% equity | 5.0% | PASS |
| SQN | >= 1.0 | 4.51 | PASS |
| Sortino | >= 1.5 | 11.37 | PASS |

**Recommendation:** The sample is small (22 trades), so the GO verdict is provisional. Collect more forward-test data before deploying with full confidence.

## Assumptions & Caveats

- Sample size is below 30 trades; all metrics are indicative, not conclusive.
- Sharpe/Sortino use per-trade returns on running equity with a sqrt(N) annualization proxy; they are trade-frequency measures, not calendar-time annualized.
- SQN is computed from R-multiples (pnl / risk-per-trade); where risk is unknown, dollar PnL is used.
- Annualized return is derived from the first-to-last trade timestamps; a short span may exaggerate annualization.
- VaR/CVaR are empirical percentiles of the trade PnL distribution (95% confidence).
- Monte Carlo resamples the observed trade outcomes with replacement; it does not model changing position sizing or market regime.
- Risk of ruin (Balsara) uses a gambler's-ruin estimate with average risk per trade; treat as approximate.
- ICT classifications are derived from the technical-context field when present; otherwise reported as "None".
- Missing fields are reported in the data-integrity panel; metrics relying on missing fields fall back to "n/a".

---
Generated by the Trading System Evaluation Dashboard.