Learn About PHG Grades

Our most complete grades come from Manual A3Z + MP/NST mode, which blends human-tracked microstat data from the All Three Zones project with advanced metrics from MoneyPuck and Natural Stat Trick. When A3Z tracking is available, it captures things no automated system can, and that makes the grade more accurate.

Manual A3Z + MP/NST (default) is our primary and most detailed view. A portion of games each season are manually charted, on-ice event by on-ice event, by All Three Zones, a hockey microstats tracking project by Corey Sznajder, capturing things automated systems miss like zone exits, defensive-zone coverage, and board battles. For those tracked games, A3Z grades are blended with data from MoneyPuck and Natural Stat Trick to produce the most complete picture of each player. GP in this mode reflects only the games that were manually tracked by A3Z, not the full season.

All Games (MP/NST) shows grades across all regular season games, powered entirely by MoneyPuck and Natural Stat Trick data. No manual tracking is used here. This is useful for seeing how a player grades across a full season, especially for players whose teams had fewer manually tracked games. Faceoff grades are not available in this mode since that data requires manual tracking.

Exploring individual players: Clicking any player in the season leaderboard opens their profile. At the top you will see their season grades and basic stats. Below that is a game-by-game log showing their Overall, Offensive, and Defensive grade for each game they played. Games that included A3Z manual tracking are marked in the opponent column. The game log is useful for context, since a strong season grade built on consistent nights tells a different story than one carried by two or three standout performances.

Team Rankings: The Teams tab on the season page shows how every team ranks against one another using the same grading system. Teams are ranked by averaging the grades of their qualified players, so you can compare rosters by Overall, Offensive, and Defensive grade side by side. You can also toggle between Manual A3Z + MP/NST and All Games (MP/NST) to see how team rankings shift depending on which mode is active.

Game Lookup: The Game Lookup page, linked in the top navigation, lets you find individual game grades for any game in the season. For regular season games, browse by team and then by month to find a specific matchup. For playoff games, navigate through the bracket. Clicking a graded game shows every skater's Overall, Offensive, and Defensive grade for that night alongside their time on ice. This is useful for breaking down a single game in detail rather than looking at a player's season-long arc.

Games marked with a ⬡ Microstat badge have been manually tracked by A3Z and use the full blended grade, the same Manual A3Z + MP/NST approach described above. Games without the badge are graded using MP/NST data only. Playoff games are graded on a per-game basis using the same blend format as regular season games whenever A3Z tracking is available.

Data for manually tracked games comes from the All Three Zones project by Corey Sznajder. A huge thanks to Corey for making this level of detail publicly available. PHG then takes that raw microstat data and converts every event type into a grade delta, a custom-built system that assigns a positive or negative value to each on-ice action based on its impact. That conversion, and the weighting behind it, is the core of what PHG built.

Each delta represents how much a play helped or hurt the team. Did a player create a high-danger shot chance that the goalie robbed? They get credit for the quality of the opportunity even though it didn't go in. Did a defenseman give the puck away in his own zone leading to a scoring chance? That turnover gets penalized regardless of whether the other team scored.

This eliminates a lot of luck. A goal on a weak shot from the perimeter earns less credit than a snipe from the high slot. A blocked shot on a dangerous chance is worth more than one blocked from the point. The grade reflects the actual impact of the play, not just whether it ended up on the scoresheet.

These A3Z-derived grades are then blended with MoneyPuck and Natural Stat Trick data using custom weights. On the offensive side this includes metrics like venue-adjusted individual expected goals (iXG), on-ice expected goals percentage, and powerplay production. On the defensive side it pulls in on-ice expected goals against, high-danger chance suppression, and penalty killing performance. The exact weighting varies by position, reflecting that what makes a great defenseman is different from what makes a great forward.

In All Games (MP/NST) mode, grades are built from aggregated season statistics provided by MoneyPuck and Natural Stat Trick. These include metrics like individual expected goals per 60 minutes (iXG/60), on-ice shot share (CF%), high-danger chance rates, and more.

Rather than grading individual plays, this approach looks at a player's rate stats across many games and asks whether that player is generating more or fewer high-quality chances than expected for someone in their role. The result is normalized the same way, 75 = league average, one standard deviation = 7 points, so grades are directly comparable between modes.

The tradeoff is that MP/NST grades are based on what automated tracking captures. They can miss things a human observer would catch: a defenseman who breaks up plays before they become shots, or a forward who constantly creates opportunities that don't show up as shots on goal. That is exactly why A3Z manual tracking gets blended in.

All grades, whether from A3Z manual tracking or MP/NST, are normalized to the same 0-100 scale. 75 = league average. One standard deviation equals 7 points, meaning a grade of 82 puts a player roughly one standard deviation above average, and a grade of 68 puts them one below. Forwards and defensemen are normalized separately so grades reflect performance relative to the same position group.

For manually tracked games, each on-ice event earns a delta between roughly -2.0 and +2.0. Goals are the exception. A goal bonus stacks on top of the shot delta, reflecting that scoring is the single highest-impact play in hockey. A typical goal results in a combined delta of around +3.4.

Here is how different on-ice events affect a player's grade:

Each impact level corresponds to a weighted point value that accumulates across a full game before being normalized to the 0–100 scale.

Within A3Z's manually tracked games, not every play of the same type carries equal weight. PHG adjusts grade points based on where on the ice the event occurred and the game situation.

Zone exits and entries are a good example. Carrying the puck cleanly out of your own zone is positive, but failing to exit under pressure is penalized. Getting the puck into the offensive zone with control is rewarded more than a dump-in that goes nowhere. A zone entry that creates a scoring chance carries more weight than one that doesn't.

Entry denials work the same way. A defenseman who stops a clean carry-in at his blue line is protecting high-danger territory and gets credit accordingly.

Faceoffs are also adjusted by zone and situation. Winning a defensive zone draw while killing a penalty is one of the most valuable plays a center can make. Losing that same draw is penalized heavily. Powerplay faceoffs in the offensive zone are weighted higher too, since losing one wastes the man advantage before it even begins. Shorthanded situations apply an elevated multiplier to faceoffs across the board.

Every player receives four grades. Overall is the headline number. The three sub-grades break performance into specific areas, each normalized separately so they can be compared across position groups. Sub-grades draw from both manual tracking and MP/NST data depending on which mode is active.

Playoff grades are normalized within the playoff pool only — roughly 330 players across teams that qualified for the postseason. Because weaker teams are eliminated, everyone left is an above-average NHL player. This means 75 still represents the average playoff performer, but that average is higher quality than the regular season average.

As a result, you cannot directly compare a playoff grade number to a regular season grade number. A player grading 82 in the playoffs is performing better relative to their competition than a player grading 82 in the regular season — the playoff bar is simply harder to clear. Think of it like class rank: finishing at the top of a more selective group is a greater accomplishment than the same rank in a larger, more varied pool.

To give better visual separation in the compressed playoff pool, individual playoff game grades use a slightly wider scale than the regular season. This means the gap between the best and worst playoff performers looks similar to what you'd see in the regular season, even though the underlying talent range is narrower.

The practical takeaway: use playoff grades to compare players within the playoffs, and regular season grades to compare players across the full league. Crossing the two numbers directly will understate how hard it is to grade well in the postseason.

Goals and assists only tell part of the story. Most players, especially those who are not primary point producers, contribute in ways that never show up on the scoresheet. A shutdown defenseman who smothers high-danger chances all night, blocks shots, and kills penalties had an excellent game even if he finished with zero points. A fourth-line center who wins every critical defensive-zone faceoff directly protects his team's lead, but the goal column gives him nothing.

Points can also be noisy. A player who scores twice on weak shots from the perimeter benefited from puck luck. A player who creates four high-danger chances that the goalie stones deserves credit even with no points to show for it. PHG measures the quality and impact of every action, not just whether the puck crossed the line.

The result: players who consistently do the right things grade out well even in low-scoring stretches, and players riding hot shooting nights or heavy powerplay time don't get grades they don't deserve.

In Manual A3Z + MP/NST mode, players must appear in at least 50% of their team's manually tracked games to qualify for season rankings. In All Games (MP/NST) mode, the threshold is 50% of their team's games played in the season. Players below the threshold are still shown but grayed out and excluded from rank calculations.

Faceoff grades require a minimum of 10 faceoffs taken across the season and are only available in Manual A3Z + MP/NST mode.

Season grades are not simply an average of game grades. A player who posts consistent grades game after game will grade out higher than one who has a single outstanding night surrounded by poor performances. Sustained performance across a full season is harder to maintain and is reflected accordingly.