S&P 500 Momentum · Direct Index

SPMO uses 12-1 momentum / 3yr weekly vol, semi-annual rebal, 100 stocks, score×Mcap weighting, 5% cap. Defaults below diverge on every dimension: blended lookbacks, faster vol, concentrated 50-stock book, pure score weighting, weekly rebal, plus regime & acceleration overlays.

Strategy Notes & Free-Tier Playbook

v2 engine vs SPMO: SPMO scores names on a single 12-1 risk-adjusted return, rebalances semi-annually, holds 100 names, weights by score × market cap, caps at 5%. v2 (a) blends three lookbacks (12-1 / 6-1 / 3-1, weights 0.50 / 0.30 / 0.20) to capture both classical and faster momentum; (b) uses 6-month daily realized vol instead of 3-year weekly vol so the risk denominator reacts to regime shifts; (c) applies a positive-momentum filter (Asness / Moskowitz) — names with negative 12-1 are excluded; (d) adds an acceleration boost — names whose 3-1 return is outpacing their 12-1 average get +15% score; (e) uses pure score weighting (no Mcap dilution — we have no capacity constraint) with a 6% stock cap, 30% sector cap, and a 0.5% floor; (f) protects existing holdings via a 1.5× turnover band — a name only gets sold if it drops below rank 1.5×N, which cuts weekly turnover roughly in half; (g) adds a regime overlay (Daniel-Moskowitz crash hedge) — if the universe median 12-1 return goes negative, exposure is scaled down by 50%.

Why these changes matter for weekly rebal: weekly rebalancing only helps if your signal is fresh and your turnover is controlled. The 3-1 component and acceleration boost surface breakouts SPMO misses for ~5 months. The turnover band and pure score weighting offset the noise that weekly rebal would otherwise introduce. In a Roth IRA, the standard tax cost of high turnover is zero, so the net effect is upside — provided your broker offers fractional shares and commission-free trades.

Tiingo free tier (recommended): 50 unique tickers/day, 1,000 calls/day, 30+ years of EOD history. Get a free token at tiingo.com. Each ticker costs 1 call (history only — no market-cap endpoint on the free tier, so the engine skips that phase and uses the latest close from price history when sizing trades). Pure score weighting is therefore the only sensible mode here, which is already v2's default. At universeSize=150 the initial seed takes ~3 days (45/day default budget); after that, weekly rebal costs 0 calls until the cache expires (5 days).

FMP free tier: 250 calls/day, 500MB/30d. S&P 500 constituents endpoint is premium-only, so the universe is baked in. Uses /stable/historical-price-eod/light + /stable/quote (2 calls per ticker). Switch to FMP in Settings if you need market cap data for score×Mcap weighting.