Playoff Preview: Intro and Stat Explainer
It looks like we might actually have some honest-to-god hockey coming up, so I’ve created a whole bunch of new visualizations to directly compare teams in a number of ways. The goal of these posts is to basically give you all of the info you could possibly want to see how teams stack up against eachother. In a 5-game play-in, the better team only has a 63% chance of winning at best, so there’s plenty of randomness coming.
This post is entirely dedicated to explaining the stats (from Evolving-Hockey and MoneyPuck) and visualizations I’ll be using. As a placeholder, we will imagine an incredible Stanley Cup Finals matchup between two titans: the San Jose Sharks and the Detroit Red Wings.
Play-Driving
This section breaks down the two teams’ possession and scoring rates, including:
Expected Goals For % (xGF%): A stat that shows the percentage of the total expected goals (shots weighted by how likely they are to go in based on historical shot location data) a team was responsible for. Example: If the Sharks had 4 expected goals and their opponents had 6, the Sharks would have an xGF% of 40%.
Goals For %: A stat that shows the percentage of total on-ice goals a team was responsible for. Example: the xGF example but with goals instead.
Corsi For %: A stat that shows the percentage of total on-ice Corsi (i.e. shot attempts) a team was responsible. Example: the xGF example but with shot attempts instead.
GF-xGA: A stat that shows how many goals a team scored compared to how many expected goals they allowed. Used to evaluate a skater group, since they are responsible for shooting percentage but not the performance of their own goalie.
Quality Game % (QG%): A stat that measures the percentage of a team’s games in which they were a >50% xGF% team.
The charts show the team’s possession rates (xGF%) and goal rates (GF%) from month to month in order to give a sense for outlier hot/cold streaks and the team’s trajectory over the course of the season.
Team Offence
This section breaks down teams’ offence and shooting statistics, including:
Even Strength Offence Wins Above Replacement (EVO WAR): A stat created by EvolvingWild which condenses offensive play into a single number based on a number of inputs including goals for and Corsi for to estimate the number of wins a team received compared to if they had played a roster of “replacement level players” (13th forwards, 7th defencemen).
Powerplay Wins Above Replacement (PP WAR): Similar to the stat above, but solely measuring powerplay performance.
Goals For per 60 Minutes (GF): The rate at which a team scored goals at even strength.
Expected Goals For per 60 Minutes (xGF): The number of goals a team was expected to score per 60 even strength minutes based on the quantity and quality of their shots.
Shooting Percentage Above Expected (Sh%Ax): A stat which compares a team’s shooting percentage (goals per shot) to their expected shooting percentage (expected goals per shot) to measure shooting talent and luck.
The graphs visualize how the team performed on a month-by-month basis. When the blue is higher than the red, the team is shooting poorly, and vice versa. The grey line represents the league average. A good team will have both lines above league average.
Team Defence
This section breaks down teams’ defence and goaltending statistics, including:
Even Strength Defence Wins Above Replacement (EVD WAR): Just like EVO WAR but for defensive play.
Penalty Kill Wins Above Replacement (PK WAR): Just like PP WAR, but for the penalty kill.
Goals Against per 60 Minutes (GF): The rate at which a team allowed goals at even strength.
Expected Goals Against per 60 Minutes (xGF): The number of goals a team was expected to allow per 60 even strength minutes based on the quantity and quality of their allowed shots.
Save Percentage Above Expected (Sh%Ax): A stat which compares a team’s save percentage (saves per shot) to their expected shooting percentage (expected saves per shot) to measure goaltending.
The graphs visualize how the team performed on a month-by-month basis. When the blue is higher than the red, the team is getting good goaltending, and vice versa. The grey line represents the league average. A good team will have both lines below league average.
Forwards
This section compares the teams’ forward lineups, including:
Line Projected WAR: Measures the projected wins above replacement of a line. Each player is given a projected 82 game WAR value by taking their 3-season weighted WAR per minute rate and multiplying that value by the projected minutes they would play in a season in their lineup role.
Line EVO and EVD: Specifically adds up the even strength offence and even strength defence wins above replacement value of the players on the line to approximate the line’s offensive and defensive aptitude.
Line Shooting (Sh%): Specifically adds up the shooting talent of a line, measured using a three-year weighted average of each player’s shooting percentage above expected.
Defencemen
This graphic is the same as the previous one, but for defencemen. The statistics used are the same, although xWAR is used instead of WAR due to the relative inability of defencemen to directly influence the shooting performance of the forwards on their teams.
Goalies
The final visualization compares the projected starting goaltender of the two teams. It includes:
5v5 Save Percentage Above Expected Starter Rank (5v5 Rank): This shows how a goalie ranks among the NHL’s 31 starting goalies in a stat which isolates a goaltender’s play from the performance of his team’s defence by comparing his save percentage to the save percentage that he would be expected to have based on the quality and quantity of shots he faces.
Shorthanded Save Percentage Above Expected Starter Rank (Sh Rank): This is the same as above, but only for 4v5 penalty kill situations.
Rebounds Above Expected (Rb Rank): This is similar to the above, but measures how a goalie’s rebound control compares to other starters.
All Situations Save Percentage Above Expected Starter Rank (All Rank): The same as the first two stats, but in all situations.
In addition to these stats, which are based on MoneyPuck’s model, there is also a graph that breaks down each goalie’s performance on a month to month basis using goals saved above expected (the number of goals a goalie prevented compared to what he would have been expected to based on the quality and quantity of shots against). This is to tell when goalies simply had hot or cold stretches that might lead to misleading overall numbers.
Conclusion
After each breakdown I’ll pick which team I think comes out better in each category, and then make a prediction for how I think the series will go: for example, Sharks in 4.
And that’s everything! I will be posting these breakdowns in June to preview the play-in round and then will subsequently post them for future playoff series as well. Enjoy!