20-Day Price Momentum Factor

The first factor carried through the full workflow, deliberately using a simple definition to clarify contracts between research, signals, and execution.

HYPOTHESIS

Research objective

Test whether stocks with stronger trailing 20-day performance earn relatively higher returns over the following five trading days. The first pass avoids complex optimisation until the signal itself is understood.

Factor definition

Momentum_i,t = P_i,t / P_i,t−20 − 1

Validation target: forward five-trading-day return

METHOD

Research workflow

01Load an adjusted-close price matrix
02Calculate trailing 20-day return
03Calculate forward five-day return
04Align factor and target by date and security
05Produce IC, Rank IC, ICIR, and grouped returns
06Convert the Top 20% into an equal-weight target portfolio

IMPLEMENTATION

Production implementation

src/archer_factor_analyzer/calculate_factors.pyPYTHON

def calculate_momentum_factor(price_df, window=20):
    momentum_factor_df = calculate_return(
        price_df, period=window
    )
    return momentum_factor_df

Reuses the generic return utility while keeping the window configurable.

EVIDENCE & LIMITS

Evidence and boundaries

01
Core IC, Rank IC, ICIR, Q1–Q5, and Q5−Q1 cumulative-return checks are complete.
02
A cost-aware fixed-universe strategy backtest has been completed in RQAlpha.
03
The dynamic eligibility mask exists but is not yet wired into the formal ranking workflow.
04
The result describes this specific historical experiment; it is not a promise of future performance.