MLB Power Rankings — Week —
| # | Team | Rating | Trend | Our Take | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Loading rankings… | |||||||||||
Methodology — how we build the rankings
The Signal Labs MLB Composite Rating blends five components to capture both long-run team quality and short-term form. Every component is z-scored against the league and then re-projected onto an Elo-like scale (mean 1500, sd 50) so the headline number stays human-readable.
1. Elo (40%) — Our in-house MLB Elo, calibrated against multiple historical baselines and updated nightly off final scores. Elo captures durable team quality independent of small-sample noise.
2. Pythagorean Win % (20%) — Classical run-differential-based expected win rate over the last 60 games. Catches teams whose record over- or under-states their true performance.
3. 14-Day Form (15%) — Rolling win % and run-diff per game over the last two weeks. Picks up injuries, call-ups, and confidence/momentum shifts the long windows miss.
4. Pitching Index (15%) — Runs-allowed-per-game over the same 60-game window, indexed against league average (100 = avg, higher = better). Pitching is the most leverage-able component in MLB and gets its own weight.
5. Strength of Schedule (10%) — Average opponent Pythagorean win % over the last 40 games. Adjusts for soft/hard schedules so that the same record means more against tougher opponents.
Tier bands are fixed by rank: Elite (1–6), Contender (7–15), Bubble (16–22), Pretender (23–27), Cellar (28–30). The "consensus market" comparison uses a live win-loss-record proxy when an external snapshot is unavailable.
What this is not: a prediction. These ratings are projections for analysis only — not advice, not guarantees of outcomes. We publish them because the divergences vs the consensus market are the most actionable thing in the model.