""" monte_carlo_chart_v1.py ======================== Generates a Monte Carlo simulation chart showing: - 2,000 simulated equity paths (shuffled trade order) - 5th-95th percentile band (outer light blue) - 25th-75th percentile band (inner dark blue) - Median simulation path (dashed) - Actual historical equity curve (solid dark) Run from project root on home machine: python src/research/monte_carlo_chart_v1.py """ import pandas as pd import numpy as np import matplotlib matplotlib.use("Agg") import matplotlib.pyplot as plt import matplotlib.patches as mpatches from pathlib import Path import sys BASE_PATH = Path(__file__).resolve().parents[2] SRC_PATH = BASE_PATH / "src" TRADES_DIR = BASE_PATH / "data" / "processed" / "trades" CHARTS_DIR = BASE_PATH / "data" / "processed" if str(SRC_PATH) not in sys.path: sys.path.append(str(SRC_PATH)) # ── Config ──────────────────────────────────────────────────────────────────── N_SIMS = 2000 START_BALANCE = 100_000 NOTIONAL = 300_000 ZSCORE_BANDS = [(2.75, 3.50, 1.0), (3.50, 4.50, 1.5), (4.50, 99.0, 2.0)] SEED = 42 def get_multiplier(z): for lo, hi, m in ZSCORE_BANDS: if lo <= z < hi: return m return 2.0 def main(): print("=" * 60) print("MONTE CARLO SIMULATION CHART") print("=" * 60) # ── Load trade log ──────────────────────────────────────────────────────── trade_log = None for fname in ["trades_real_costs.csv", "trades_eurusd_final.csv"]: p = TRADES_DIR / fname if p.exists(): trade_log = pd.read_csv(p) print(f"\n Loaded: {fname} ({len(trade_log):,} trades)") break if trade_log is None: print(" ERROR: No trade log found in data/processed/trades/") return # Identify return column ret_col = None for c in ["return_real", "return", "ret"]: if c in trade_log.columns: ret_col = c break if ret_col is None: print(f" ERROR: Could not find return column. Columns: {trade_log.columns.tolist()}") return # Identify z-score column z_col = None for c in ["zscore_abs", "lag_zscore_24h_v3", "z_score"]: if c in trade_log.columns: z_col = c break # Build returns array with position sizing returns = trade_log[ret_col].values.copy() if z_col: multipliers = trade_log[z_col].abs().apply(get_multiplier).values notionals = np.minimum(NOTIONAL * multipliers, NOTIONAL * 3) else: notionals = np.full(len(returns), NOTIONAL) dollar_pnl = returns * notionals n_trades = len(dollar_pnl) print(f" Trades : {n_trades:,}") print(f" Win rate : {(returns > 0).mean()*100:.1f}%") print(f" Avg return : {returns.mean()*100:.4f}%") print(f" Running {N_SIMS:,} simulations...") # ── Actual equity curve ─────────────────────────────────────────────────── actual_eq = np.empty(n_trades + 1) actual_eq[0] = START_BALANCE actual_eq[1:] = START_BALANCE + np.cumsum(dollar_pnl) # ── Monte Carlo ─────────────────────────────────────────────────────────── rng = np.random.default_rng(SEED) sims = np.empty((N_SIMS, n_trades + 1)) sims[:, 0] = START_BALANCE for i in range(N_SIMS): idx = rng.permutation(n_trades) sims[i, 1:] = START_BALANCE + np.cumsum(dollar_pnl[idx]) # Percentile bands p5 = np.percentile(sims, 5, axis=0) p25 = np.percentile(sims, 25, axis=0) p50 = np.percentile(sims, 50, axis=0) p75 = np.percentile(sims, 75, axis=0) p95 = np.percentile(sims, 95, axis=0) x = np.arange(n_trades + 1) print(f" Final balance range:") print(f" 5th pct : ${p5[-1]:,.0f}") print(f" 25th pct : ${p25[-1]:,.0f}") print(f" Median : ${p50[-1]:,.0f}") print(f" 75th pct : ${p75[-1]:,.0f}") print(f" 95th pct : ${p95[-1]:,.0f}") print(f" Actual : ${actual_eq[-1]:,.0f}") # ── Plot ────────────────────────────────────────────────────────────────── fig, ax = plt.subplots(figsize=(16, 8)) fig.patch.set_facecolor("white") ax.set_facecolor("white") # Draw a sample of individual sim paths (very light) sample_idx = rng.choice(N_SIMS, size=300, replace=False) for i in sample_idx: ax.plot(x, sims[i], color="#a8c4e0", alpha=0.04, linewidth=0.4) # 5th-95th band ax.fill_between(x, p5, p95, color="#c5d9ee", alpha=0.5, label="5th–95th pct") # 25th-75th band ax.fill_between(x, p25, p75, color="#89afd4", alpha=0.55, label="25th–75th pct") # Median ax.plot(x, p50, color="#6b9cc9", linewidth=1.8, linestyle="--", alpha=0.9, label="Median sim") # Actual ax.plot(x, actual_eq, color="#0f1f3d", linewidth=2.2, label="Actual", zorder=10) # ── Formatting ──────────────────────────────────────────────────────────── ax.set_title("Monte Carlo Simulation", fontsize=16, fontweight="normal", pad=16, color="#1a1a2e") ax.set_xlabel("Trade number", fontsize=12, color="#4a4a6a", labelpad=10) ax.set_ylabel("Account balance", fontsize=12, color="#4a4a6a", labelpad=10) # Y-axis formatter def yfmt(v, _): if v >= 1_000_000: return f"${v/1e6:.1f}M" if v >= 1_000: return f"${v/1000:.0f}k" return f"${v:.0f}" from matplotlib.ticker import FuncFormatter ax.yaxis.set_major_formatter(FuncFormatter(yfmt)) ax.tick_params(colors="#6a6a8a", labelsize=10) ax.spines["top"].set_visible(False) ax.spines["right"].set_visible(False) ax.spines["left"].set_color("#dde3ee") ax.spines["bottom"].set_color("#dde3ee") ax.grid(axis="y", color="#eaecf5", linewidth=0.8, linestyle="-") ax.grid(axis="x", color="#eaecf5", linewidth=0.5, linestyle="-") ax.set_axisbelow(True) # Legend handles = [ mpatches.Patch(color="#c5d9ee", alpha=0.8, label="5th–95th pct"), mpatches.Patch(color="#89afd4", alpha=0.9, label="25th–75th pct"), plt.Line2D([0],[0], color="#6b9cc9", linewidth=1.8, linestyle="--", label="Median sim"), plt.Line2D([0],[0], color="#0f1f3d", linewidth=2.2, label="Actual"), ] ax.legend(handles=handles, loc="upper left", frameon=True, framealpha=0.9, edgecolor="#dde3ee", fontsize=10) # Annotate final actual balance ax.annotate(f"${actual_eq[-1]/1e6:.2f}M", xy=(n_trades, actual_eq[-1]), xytext=(n_trades - n_trades*0.06, actual_eq[-1] * 1.04), fontsize=10, color="#0f1f3d", fontweight="bold", arrowprops=dict(arrowstyle="->", color="#0f1f3d", lw=1.2, connectionstyle="arc3,rad=-0.2")) plt.tight_layout() out = CHARTS_DIR / "monte_carlo_chart.png" plt.savefig(out, dpi=180, bbox_inches="tight", facecolor="white") plt.close() print(f"\n Saved: {out}") print(f"\n Open: {out}") if __name__ == "__main__": main()