CoT Market Tone

A macro-level breadth and sentiment indicator derived from speculative crowding across all futures markets in the CoT dataset. Measures systemic Risk-On vs Risk-Off tone using crowding breadth, net tone, and robust z-score normalisation.

Abstract

The CoT Market Tone signal aggregates daily positioning across all markets to determine whether speculative futures activity is broadly Risk-On or Risk-Off. By comparing crowded-long vs crowded-short breadth, constructing a net tone index, and applying robust z-score scaling, the model classifies market tone into regimes ranging from Strong Risk-On to Strong Risk-Off.

1. Inputs

df_cot_signals — unified enriched CoT dataset with market-level signals.
crowded_long / crowded_short flags — indicate extreme speculative positioning.
Market_and_Exchange_Names — used to count unique markets per report date.

2. Daily Aggregation: Breadth Metrics

For each CoT reporting date, the model aggregates market-level information to compute Risk-On vs Risk-Off breadth:

crowded_long_count: number of markets flagged crowded long.
crowded_short_count: number of markets flagged crowded short.
n_markets: number of unique markets in the dataset.

These are used to compute:

crowded\_long\_share = crowded\_long\_count / n\_markets

crowded\_short\_share = crowded\_short\_count / n\_markets

net\_tone = crowded\_long\_share − crowded\_short\_share

3. Robust Z-Score Normalisation

A robust z-score is applied to the daily net tone to stabilise the signal and reduce sensitivity to outliers:

z\_{rob}(X) = \dfrac{X − median(X)}{1.4826 · MAD(X)}

This transformation ensures that extreme values, skewed distributions, and structural shifts do not distort the market tone classification.

4. Regime Classification

The robust z-score is mapped to qualitative market regimes:

z ≥ 1.0 → Strong Risk-On

0.5 ≤ z < 1.0 → Risk-On

−0.5 < z < 0.5 → Neutral

−1.0 < z ≤ −0.5 → Risk-Off

z ≤ −1.0 → Strong Risk-Off

5. Implementation (Python)

df_tone = df_cot_signals.copy()
df_tone["Report_Date"] = pd.to_datetime(df_tone["Report_Date"], errors="coerce")

# Aggregation
df_cot_tone_daily = (
    df_tone.groupby("Report_Date").agg(
        n_markets=("Market_and_Exchange_Names", "nunique"),
        crowded_long_count=("crowded_long", "sum"),
        crowded_short_count=("crowded_short", "sum"),
    ).reset_index()
)

df_cot_tone_daily["crowded_long_share"] = df_cot_tone_daily["crowded_long_count"] / df_cot_tone_daily["n_markets"]
df_cot_tone_daily["crowded_short_share"] = df_cot_tone_daily["crowded_short_count"] / df_cot_tone_daily["n_markets"]

df_cot_tone_daily["net_tone"] = df_cot_tone_daily["crowded_long_share"] - df_cot_tone_daily["crowded_short_share"]

# Robust z-score
med = df_cot_tone_daily["net_tone"].median()
mad = (df_cot_tone_daily["net_tone"] - med).abs().median()
df_cot_tone_daily["CoT_Market_Tone_z"] = (df_cot_tone_daily["net_tone"] - med) / (1.4826 * mad)

# Regimes
def _tone_label(z):
    if pd.isna(z): return "Unknown"
    if z >= 1.0: return "Strong Risk-On"
    if z >= 0.5: return "Risk-On"
    if z > -0.5: return "Neutral"
    if z > -1.0: return "Risk-Off"
    return "Strong Risk-Off"

df_cot_tone_daily["CoT_Market_Tone_Regime"] = df_cot_tone_daily["CoT_Market_Tone_z"].apply(_tone_label)

6. Interpretation

Risk-On tone indicates speculative appetite and broad directional alignment in long positioning.
Risk-Off tone reflects defensive or uncertain market behaviour with increased short positioning breadth.
Neutral tone signals a balanced speculative environment lacking strong macro conviction.

7. Limitations

Breadth measures may mask sector‑specific divergences.
Some markets have thin liquidity, which can distort crowding flags.
Static regime thresholds may not reflect structural shifts in futures participation.

8. Practical Use Cases

Detect broad speculative sentiment shifts for macro allocation.
Combine with volatility or yield‑curve signals to confirm risk regime transitions.
Use as an input to multi‑factor macro composites or tactical models.