Net xG/60 = expected goals created minus expected goals allowed per 60 shared 5v5 minutes. These are estimated strength rates, not projected goals tonight or betting probabilities. Larger is better; xGA/60 alone is better when lower.

Shared-trio rates are blended toward player/team rates using shared minutes ÷ (shared minutes + 60). Each player’s on-ice prior is also blended toward the team using player minutes ÷ (player minutes + 60). With no measured trio sample, the player/team estimate is used and marked ★. The same 60-minute setting is used throughout; small samples remain uncertain.

Matchup exposure is context only and does not change the strength ranking. Projected playing time and power-play results are separate. Use the line filters above to browse these reported complete lines.

This 24-team advance slate uses player/team estimates through October 20. Retained exact-five audits through October 19 cover Florida and Ottawa only; the other teams and later games are untracked. Coverage is partial and unavailable samples remain unknown. PlayerWON’s NHL shift export is checked for roster coverage, nonoverlapping intervals and individual ice time against the game logs (five-second tolerance; differences retained). The overlapping official HTML reports are also cross-checked. All exact five-player groups are retained, including brief line changes. Tonight’s projected units are matched by player IDs, not historical PP1/PP2 labels.

Unit samples include only true 5-on-4 with both goalies present. Measured xG, SOG and goals require independent game-level time/count reconciliation. A source-time control can include separately identified overtime 4-on-3 seconds; the exact-five unit samples never do. Unresolved games remain visible in the coverage table and unit game logs. Full shift coverage does not mean every game’s PP production passed.

Every unblocked shot in a covered game is checked against the reconstructed on-ice players. Original shot-feed strength tags are cross-checked rather than treated as ground truth. Credited events must contain the shooter in the exact on-ice group. This recovers some samples previously withheld by inconsistent strength tags.

Displayed xGF/60 is an estimate: verified exact-five production is blended toward a five-player/team prior using sample minutes ÷ (sample minutes + 60). Each player’s PP on-ice prior is also regressed toward team PP using player minutes ÷ (player minutes + 60). Missing player data uses the team baseline. These experimental weights are not calibrated PP forecasts.

Last five means the club’s last five completed games, including games in which tonight’s exact five did not play together. Rates use verified production seconds only. Missing or rejected data is unavailable, not a zero result. Allocated PP minutes are a deployment estimate, not projected exact-five shared minutes. PK context and allocated minutes do not change attacking-strength order; the optional projected-PP-xG order uses both. The opponent PK factor is capped at ±15% and has not been backtested. Player projections are unchanged.

Sources: PlayerWON NHL shift export, MoneyPuck shots and game logs, and Daily Faceoff personnel charts linked in each unit.