launchtower-factor-research

LaunchTower — Factor Screen (Free Sample)

Independent market-data desk. We build reproducible, documented factor screens on US large-caps from public market data. This repo is the free sample — the full dataset, methodology pack, and live signal feed are available on Whop.

⚠️ Disclaimer: Research/educational output from public market data. Not personalized investment advice, not a recommendation to buy or sell any security. Past performance does not guarantee future results.

What’s in this repo

File	Description
`launchtower_factor_report_2026-09-16.md`	Dated research report: top/bottom 10, factor scores, narrative
`launchtower_signal_2026-09-16.csv`	Full 151-row factor table (raw factors + z-scores + composite)
`launchtower_factor_screen_2026-09-16.py`	The complete, runnable script that reproduces every number

The model in one paragraph

Pull 2 years of split/dividend-adjusted daily closes for 151 US large-caps. Over the trailing 252 trading days, compute per-ticker: 1m/3m/6m/12m returns, annualized realized volatility, max drawdown, distance from 52w high. Cross-sectionally z-score the 12m return → Momentum; z-score annualized vol and negate → Quality. Composite = 0.5·Momentum + 0.5·Quality, rank 1–151.

How to run it yourself

pip install yfinance pandas numpy
python launchtower_factor_screen_2026-09-16.py

The script prints the top/bottom 10 and writes a dated CSV of the full factor table.

Latest screen (2026-09-16)

Top 5: MU · VLO · INTC · MPC · PSX Bottom 5: HOOD · ZS · MRNA · COIN · SMCI

Full table: see the CSV. Narrative: see the report.

What’s in the paid pack (Whop)

Full 151-ticker dataset with all raw factors, z-scores, and composite scores
Complete methodology documentation (factor definitions, z-scoring, weighting, edge cases)
Live signal feed (dated CSV, updated on a schedule)
The full runnable script with configuration knobs (universe, window, weights)

👉 Get the full pack on Whop

LaunchTower — independent market-data desk. Data: yfinance (public). Regenerated from live data at generation time.