Analysis

Deep-dive breakdowns across teams, systems, and market patterns

2

Articles

12

Key Findings

4 min

Avg Read

NHLNHLHome Ice

NHL Home Ice Power Rankings: Carolina Hurricanes Lead the League's Strongest Fortress

Home ice advantage in the NHL is real, measurable, and exploitable. Across 8,528 NHL regular-season games in our six-season dataset, home teams win 53.8% of the time — but the best home venues dramatically outperform that baseline. Carolina Hurricanes leads all NHL teams with a 70% home win rate over the last three seasons. Our model encodes a 50-point Elo home advantage and then layers in goalie matchups, rest differentials, and travel distance. When our model agrees with the home-ice lean, the signal is strongest — High-confidence home moneyline plays hit at 57%+ in backtests.

4 min read

Mar 17

Key Findings

  • 1

    NHL home teams win 53.8% overall — but the top venues hit 64–70%

    54%
  • 2

    One-goal games account for 39.0% of NHL results — model edge matters on ML, not just spreads

    39%
  • 3

    Carolina Hurricanes: 87-37 home (70% WR, +1.00 avg margin)

    70%
  • 4

    Winnipeg Jets: 88-40 home (69% WR, +1.04 avg margin)

    69%
  • 5

    Tampa Bay Lightning: 82-41 home (67% WR, +1.28 avg margin)

    67%
  • 6

    Vegas Golden Knights: 82-43 home (66% WR, +0.80 avg margin)

    66%
  • 7

    Model: home ice + rest advantage = strongest moneyline signal in our NHL ensemble

Data Snapshot

NHL home teams win Onegoal games accounCarolina Hurricanes Winnipeg Jets 8840 hTampa Bay Lightning 0%25%50%75%100%
NHLHome IceTeam StudyModel InsightMoneyline
NBANBASharp Play

Sharp Signal: Pistons ML — Today's Highest-Edge NBA Model Play

Our NBA XGBoost ensemble has identified Pistons ML as today's highest-edge play. The model assigns 86% win probability against a market implied of 45%, generating a raw edge of +40.7 percentage points. Pistons carry a 86% win probability per our model. Pistons hold a +18.6 net rating advantage per 100 possessions. Elo model gives Pistons a 319-point quality edge. This type of market disagreement — where our model significantly outweighs the consensus line — has historically correlated with above-average cover rates in our backtest.

3 min read

Mar 17