NHL Game Desk — Model Anchor Calibration Prompt

Context: Our MLB Player Props council (2026-04-03) established the universal framework: per-possession/per-PA rates as foundation, Bayesian shrinkage toward career prior, Log-Odds opponent adjustment, Gaussian copula for correlated combo props, dual-anchor system (sportsbook + model) tracked via Brier scores, confidence tiers. This prompt asks for the SPORT-SPECIFIC parameters to plug into that framework.

NHL Game Desk — Model Anchor Calibration

We are building model anchors for NHL game-level betting markets on Kalshi: moneylines, pucklines (spread), totals.

Questions:

1. What team metrics best predict goal differential? (Corsi/Fenwick, xGF%, score-adjusted shot rates, 5v5 metrics?)
2. How should starting goalie identity adjust the model? Goalie save% vs team xGA?
3. How many games until team metrics stabilize in NHL? k values?
4. How should back-to-back games adjust projections? NHL B2B effect is significant.
5. Home ice advantage — magnitude?
6. How to model totals: Poisson for each team's goals, sum for total?
7. How should special teams (PP%, PK%) factor in? Penalty differential drives scoring.
8. Score effects: teams trailing generate more shots. How does this affect the model?
9. Travel and time zone effects?
10. Puckline (+1.5 / -1.5): how to derive from the Poisson goal model?
11. Regulation vs OT+SO: how to handle 3-way moneyline vs 2-way?
12. EWMA alpha for team metrics?
13. Give specific numbers, formulas, and implementation-ready recommendations.