NHL Win Probability Explained

The Variance Problem

Hockey is chaos on ice. Low scoring means a single goal—which can come from a lucky bounce, a deflection, or a bad hop—can decide a game. The best team in the NHL loses to the worst team more often than in any other major sport.

That's why you'll rarely see us give any NHL team above 65-70% probability. Even when Tampa Bay faces a rebuilding team, weird stuff happens. Goalies stand on their heads. Pucks hit posts and stay out. That's hockey.

The Goalie Factor

No position in team sports matters as much as the NHL goalie. A single player touches the puck on every opposing chance. Save percentage swings of 2-3% between goalies translate to significant expected goal differences.

Starter vs Backup

The gap between a team's starter and backup can be enormous. Some backups are quality NHL goalies. Others are there because every team needs two. Knowing who's starting is critical—and we track it for every game.

Recent Form

Goalies run hot and cold. A goalie posting .940 over his last five starts is different from one at .890, even if their season numbers are similar. We weight recent performance to capture streaks.

Schedule and Rest

Like the NBA, the NHL features back-to-backs throughout the season. Teams rarely play their starter both nights. The second game usually means the backup, often on the road, often against a rested opponent.

We track rest days for both teams. A team finishing a three-games-in-four-nights stretch faces a hidden disadvantage that won't show up in their record.

Home Ice Advantage

NHL home teams win about 54-55% of games. Home ice advantage is real but modest. Last change—allowing the home team to match lines—matters. Crowd energy matters. Sleeping in your own bed matters.

Some buildings are tougher than others. Altitude in Denver, the noise in Winnipeg, the travel to get to Western Canada road trips—venue-specific factors exist beyond just "home vs away."

Special Teams

Power plays convert around 20% of the time. Penalty kills succeed 80%. Teams with elite special teams can tilt close games. When you're looking at two evenly matched teams, special teams become the separator.

We factor in power play and penalty kill percentages, adjusting for opponent quality. A 25% power play against a terrible PK means more than against a top-five unit.

What We Watch

Expected Goals

Shot quantity matters less than shot quality. A team generating 35 shots from the perimeter is less dangerous than one with 25 shots including 10 from the slot. Expected goals (xG) captures shot quality, and we use it to evaluate team strength beyond simple goals scored.

Score Effects

Teams leading by two goals in the third period play differently—protecting the lead, giving up shots from the outside. Don't read too much into late-game shot differentials when one team has stopped pushing.

Injury Context

Missing a top-line center is big. Missing your starting goalie is bigger. We track injuries and factor in who's actually available, not just who's on the roster.

Reading Our Breakdowns

Every game shows factors favoring each side. When goaltending shows HIGH impact—that's usually the goalie matchup being notably lopsided. When rest shows significant—check the schedule, someone's on a back-to-back or finishing a tough road stretch.

Hockey probabilities cluster toward 50/50 more than other sports. Don't mistake that for uncertainty in our model—it's just how hockey works.