Skew (Polygon)
Tier 1

Measures the skewness of the returns of an asset, calculated based on a range different lookback periods.

Overview

Unravel

unravel

POL Price with Skew

Factor Plot

Unravel

unravel

▆ Very Low▆ Low▆ Moderate▆ High▆ Very High

Get Access

Overall

Get Access

Skew Factor is a measure of the skewness of the returns of an asset, calculated based on a range different lookback periods.

Potential Edge

Sentiment-Driven Asymmetry

Skew reflects real-time shifts in market sentiment by quantifying the asymmetry of return distributions.

Positive skew (frequent small losses, rare large gains) signals speculative "lottery ticket" behavior, often preceding corrections as overbought conditions reverse.
Negative skew (frequent small gains, rare large losses) indicates risk-aversion, which can foreshadow capitulation or accumulation phases.
Options-based skew metrics (e.g., 25 Delta Skew) amplify this signal by quantifying the premium traders pay for upside/downside protection.

Data Collection Methodology

The Skew factor is sourced by analyzing historical return data across multiple time periods (lookback periods), calculating skewness using statistical methods, then normalizing results relative to historical averages to create a standardized index.

Unravel

unravel

Scatter plot - Skew and POL 30 and 90 Day Average Returns

Get Access

Backtest - Strategy Performance

	Benchmark (Polygon)	Strategy
Max Drawdown	100%	100%
Volatility	100%	100%
Sharpe Ratio	1.00	1.00
CAGR	100%	100%
Alpha	1.00	1.00
Beta	1.00	1.00

Predictive factors are designed to be translated into simple long-only strategy, with simulated past performance:

100% Long when the predictive factor is close to 1, with a position size equivalent to the predictive factor value.
Flat when the predictive factor is close to 0, with a position size equivalent to the predictive factor value.

The strategy is rebalanced daily, on a continuous basis. There are 0.05% transaction costs applied on each position adjustment.

Get Access

Replicate the backtest

Get started by replicating the historical performance with our code snippets.

from api import get_normalized_series, get_price_series
from backtest import vectorized_backtest
from plotting import plot_backtest_results


UNRAVEL_API_KEY = "YOUR-API-KEY"   
risk_factor = "skew"
ticker= "POL"
start_date = "2022-01-01"
end_date = "2024-06-01"
smoothing = 0


risk_factor_signal = get_normalized_series(ticker, risk_factor, start_date, end_date, smoothing, UNRAVEL_API_KEY)
price = get_price_series(ticker, start_date, end_date, UNRAVEL_API_KEY)
price = price[risk_factor_signal.index]
results = vectorized_backtest(price, risk_factor_signal)

plot_backtest_results(results, ticker, risk_factor, smoothing)