equityquant.dev submission: model, results, and 100k simulation CSV

submissions/equityquant.dev/README.md

@@ -173,6 +173,4 @@ automatically concatenates the split files when `dataset.parquet` is not present
 No data has been modified or filtered; the three files together form the complete
 submission dataset.
 
-## Contact 
-Thelonious Casey — tcasey@equityquant.dev
 

submissions/equityquant.dev/equityquant.dev_model.py

@@ -0,0 +1,142 @@
+import numpy as np
+import pandas as pd
+
+# =====================================================
+# Official horizons: 1–26 weeks
+# =====================================================
+HORIZONS = 5 * (np.arange(26) + 1)
+
+
+# =====================================================
+# CIR precision -> Poisson-weighted Gaussian mixture
+# =====================================================
+def simulate_price_series(
+    n_days=300_000,
+    s0=100.0,
+    sigma0=0.25,
+    kappa=0.02,
+    c_int=15.0,
+    seed=1,
+    a_shape=1.5,
+    lam_cap=500.0
+):
+    rng = np.random.default_rng(seed)
+    dt = 1.0 / 252.0
+    sqrt_dt = np.sqrt(dt)
+
+    # --- CIR target
+    a = a_shape
+    beta = sigma0**2
+    theta = a / beta
+    eta = np.sqrt(2.0 * kappa * theta / a)
+
+    tau = theta
+    logP = np.empty(n_days)
+    logP[0] = np.log(s0)
+    r = np.empty(n_days - 1)
+
+    # Typical lambda at long-run tau
+    lam_typ = min(lam_cap, c_int / theta)
+    lam_typ = max(lam_typ, 1e-8)
+
+    # Calibrate mixture variance
+    s_unit2 = (sigma0**2 / 252.0) / lam_typ
+    s_unit = np.sqrt(s_unit2)
+
+    for t in range(1, n_days):
+        # CIR step
+        dW_tau = rng.standard_normal() * sqrt_dt
+        tau_pos = max(tau, 0.0)
+        tau = tau + kappa * (theta - tau_pos) * dt + eta * np.sqrt(tau_pos) * dW_tau
+        tau = max(tau, 1e-10)
+
+        # Shock intensity
+        lam = c_int / tau
+        if lam > lam_cap:
+            lam = lam_cap
+
+        # Poisson-weighted Gaussian mixture
+        N = rng.poisson(lam)
+        r_t = rng.normal(0.0, s_unit * np.sqrt(N)) if N > 0 else 0.0
+
+        r[t - 1] = r_t
+        logP[t] = logP[t - 1] + r_t
+
+    prices = np.exp(logP)
+    return prices, r
+
+
+# =====================================================
+# Q-variance dataset builder (OFFICIAL, OVERLAPPING)
+# =====================================================
+def build_qvariance_dataset_from_prices(prices, ticker="DRAGON"):
+    ret = np.diff(np.log(prices))
+    scale = np.sqrt(252.0)
+
+    rows = []
+    for T in HORIZONS:
+        for i in range(0, len(ret) - T + 1):   # OVERLAPPING
+            window = ret[i:i + T]
+
+            x = window.sum()
+            sigma = np.std(window, ddof=0) * scale
+            z_raw = x / np.sqrt(T / 252.0)
+
+            if not (np.isfinite(sigma) and sigma > 0 and np.isfinite(z_raw)):
+                continue
+
+            rows.append({
+                "ticker": ticker,
+                "date": int(i),
+                "T": int(T),
+                "z_raw": float(z_raw),
+                "sigma": float(sigma),
+            })
+
+    df = pd.DataFrame(rows)
+
+    # Official de-meaning step
+    df["z"] = df.groupby("T")["z_raw"].transform(lambda g: g - g.mean())
+    df = df.drop(columns="z_raw").dropna().reset_index(drop=True)
+
+    return df[["ticker", "date", "T", "z", "sigma"]]
+
+
+# =====================================================
+# Main
+# =====================================================
+def main():
+    prices, returns = simulate_price_series(
+        n_days=120_000,
+        sigma0=0.25,
+        kappa=0.02,
+        c_int=10.0,
+        seed=3
+    )
+
+    # -------------------------------------------------
+    # CSV (first 100K points)
+    # -------------------------------------------------
+    n_out = 100_000
+
+    df_out = pd.DataFrame({
+        "Day": np.arange(n_out),
+        "Price": prices[:n_out],
+        "y": np.concatenate([[0.0], returns])[:n_out]
+    })
+
+    df_out.to_csv("qvariance_simulation_100k.csv", index=False)
+    print("Saved qvariance_simulation_100k.csv")
+
+    # -------------------------------------------------
+    # Original outputs
+    # -------------------------------------------------
+    pd.Series(prices, name="Price").to_csv("prices.csv", index=False)
+
+    df_qv = build_qvariance_dataset_from_prices(prices)
+    df_qv.to_parquet("dataset.parquet", index=False)
+    print(f"Saved dataset.parquet with {len(df_qv):,} rows")
+
+
+if __name__ == "__main__":
+    main()