plot_validation.py

import csv
import matplotlib.pyplot as plt

scenarios = []
delta_diff, delta_se = [], []
price_diff, price_se = [], []
vega_diff, vega_se = [], []

with open("validation_results.csv") as f:
    reader = csv.DictReader(f)
    for row in reader:
        scenarios.append(row["scenario"])
        f_price = float(row["f_price"]); mc_price = float(row["mc_price_mean"])
        f_delta = float(row["f_delta"]); mc_delta = float(row["mc_delta_mean"])
        f_vega  = float(row["f_vega"]);  mc_vega  = float(row["mc_vega_mean"])

        price_diff.append(100 * (mc_price - f_price) / f_price)
        price_se.append(100 * 3 * float(row["mc_price_se"]) / f_price)
        delta_diff.append(100 * (mc_delta - f_delta) / f_delta)
        delta_se.append(100 * 3 * float(row["mc_delta_se"]) / f_delta)
        vega_diff.append(100 * (mc_vega - f_vega) / f_vega)
        vega_se.append(100 * 3 * float(row["mc_vega_se"]) / f_vega)

x = range(len(scenarios))
fig, axes = plt.subplots(1, 3, figsize=(16, 5.5))

for ax, diffs, ses, title in [
    (axes[0], price_diff, price_se, "Price"),
    (axes[1], delta_diff, delta_se, "Delta"),
    (axes[2], vega_diff, vega_se, "Vega"),
]:
    ax.errorbar(x, diffs, yerr=ses, fmt="o", color="#2563eb",
                capsize=5, markersize=7, label="MC vs closed-form (mean +/- 3 SE)")
    ax.axhline(y=0, color="#111827", linewidth=1)
    ax.set_xticks(list(x))
    ax.set_xticklabels(scenarios, rotation=25, ha="right", fontsize=8)
    ax.set_ylabel("Relative difference (%)")
    ax.set_title(title)
    ax.grid(alpha=0.3)
    ax.legend(fontsize=7)

fig.suptitle("Model Validation: Monte Carlo vs Closed-Form, Relative Error with 3-SE Bands", fontsize=13)
plt.tight_layout()
plt.savefig("validation_plot.png", dpi=150)
print("Saved validation_plot.png")