When I started the LacrosseReference project, I was primarily taking concepts from other sports and applying them to college lacrosse. Possession-based efficiencies, time-of-possession, and shooting percentages were all examples. Nothing that your average sports fan wouldn't be aware of from other sports or even from a perusal of the NCAA stats site.
But as the site has grown in complexity and with the introduction of LacrosseReference PRO, the time has come to have a one-stop shop to aggregate the basic information that a fan or a coach would need to navigate the LacrosseReference suite of statistics. So here it is, LacrosseReference Stats 101.
The goal here is to give you a few things with every stat. First, a snapshot of who is the current "best" team or player in each category. Second, a sense of what sort of range is normal. The 'Range' column in the tables below show the 25th and 75th percentiles (so the range of values that encompass half the teams or qualifying players). Lastly, click on the plus icon to expand the row and get a full explanation of each statistic.
Efficiency is calculated by dividing the number of goals scored (or allowed if it's defensive efficiency) by the total number of possessions. The total number of possessions does not include failed clears and it does count possessions sandwiched around a faceoff win as 2 distinct possessions.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
My Excess Goals metric is part of my methodology for separating an offense or defense's efficiency into it's component parts: creating/preventing shots AND shooting vs goalie skills. This offensive expected shooting percentage metric is calculated by the Excess Goals model. The formula is simple: add up the probability that all shots taken would go in with an average shooter and then divide by the total number of shots taken. Higher values mean that the offense is good at generating strong scoring chances. Positive values mean that the offense has strong shooters who make more shots than you'd expect an average shooting team to make given the profile of their chances.
LaxElo is my primary metric for comparing to teams. It is based on a ranking system developed for chess by Arpad Elo. In Elo systems ranking points are transferred from the loser to the winner. The number of points transferred is based on the outcome of the game, the likelihood of that outcome, the location of the game, and the "speed" at which the system updates its evaluation of a team. I wrote a full article about the implementation and how I tailored it to college lacrosse if you want to learn more.
Calculated by measuring how many seconds elapse, on average, between when a team gains possession and when they take their first shot of a possession. The calculation only includes possessions where at least one shot was taken.
Faceoff win rate is calculated by dividing faceoffs won by total faceoffs. A faceoff that is credited as a win but then turned over in the subsequent sequence still counts toward your faceoff win rate.
This is the number of free position attempts awarded or assessed as a percentage of the total number of possessions the team had or faced. The idea is to get a sense for how good a team is at drawing penalties or how often they have penalties drawn against them.
When looking at a statistic ranking individual players, it's important to consider minimum participation thresholds. Many of my stats make no sense when you are talking about very low-usage players. An stat, for example, that adjusts a cumulative stat to account for playing time can be really distorted by low-usage players. As a result, several of the individual player stats come with a minimum threshold to qualify for the ranking. This is also in effect when calculating the range of each stat.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
'Invented' by Kyle Devitte, this is an aggregate metric that is calculated by subtracting turnovers from total points. It is similar to the LacrosseReference EGA rating in that it attempts to calculate a single rating that encompasses a player's positive and negative contributions.
This metric should be interpreted as akin to assists per touch. It is calculated by getting each player's assists/game value and then adjusting up or down depending on their play share.
To qualify, players must have met a minimum threshold with respect to play shares.
One of the earliest stats I created was a concept called "play shares," which attempted to quantify each player's role within their team/unit. It ended up being a similar concept to something like a usage-rate in basketball. Importantly, it gives a proxy for how many touches a player gets, and that allows me to "adjust" a player's expected-goals (EGA) ratings to account for how much they have the ball. Put another way, usage-adjusted EGA measures a player's efficiency with the ball and tries to strip out the fact that some players just have more chances than others.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The Draw-Specific variant of EGA filters out all non-draw plays and gives each player a number associated with their EGA contributions to winning draws. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
Efficiency is calculated by dividing the number of goals scored (or allowed if it's defensive efficiency) by the total number of possessions. The total number of possessions does not include failed clears and it does count possessions sandwiched around a faceoff win as 2 distinct possessions.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
My Excess Goals metric is part of my methodology for separating an offense or defense's efficiency into it's component parts: creating/preventing shots AND shooting vs goalie skills. This offensive expected shooting percentage metric is calculated by the Excess Goals model. The formula is simple: add up the probability that all shots taken would go in with an average shooter and then divide by the total number of shots taken. Higher values mean that the offense is good at generating strong scoring chances. Positive values mean that the offense has strong shooters who make more shots than you'd expect an average shooting team to make given the profile of their chances.
LaxElo is my primary metric for comparing to teams. It is based on a ranking system developed for chess by Arpad Elo. In Elo systems ranking points are transferred from the loser to the winner. The number of points transferred is based on the outcome of the game, the likelihood of that outcome, the location of the game, and the "speed" at which the system updates its evaluation of a team. I wrote a full article about the implementation and how I tailored it to college lacrosse if you want to learn more.
Calculated by measuring how many seconds elapse, on average, between when a team gains possession and when they take their first shot of a possession. The calculation only includes possessions where at least one shot was taken.
Faceoff win rate is calculated by dividing faceoffs won by total faceoffs. A faceoff that is credited as a win but then turned over in the subsequent sequence still counts toward your faceoff win rate.
When looking at a statistic ranking individual players, it's important to consider minimum participation thresholds. Many of my stats make no sense when you are talking about very low-usage players. An stat, for example, that adjusts a cumulative stat to account for playing time can be really distorted by low-usage players. As a result, several of the individual player stats come with a minimum threshold to qualify for the ranking. This is also in effect when calculating the range of each stat.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
'Invented' by Kyle Devitte, this is an aggregate metric that is calculated by subtracting turnovers from total points. It is similar to the LacrosseReference EGA rating in that it attempts to calculate a single rating that encompasses a player's positive and negative contributions.
This metric should be interpreted as akin to assists per touch. It is calculated by getting each player's assists/game value and then adjusting up or down depending on their play share.
To qualify, players must have met a minimum threshold with respect to play shares.
One of the earliest stats I created was a concept called "play shares," which attempted to quantify each player's role within their team/unit. It ended up being a similar concept to something like a usage-rate in basketball. Importantly, it gives a proxy for how many touches a player gets, and that allows me to "adjust" a player's expected-goals (EGA) ratings to account for how much they have the ball. Put another way, usage-adjusted EGA measures a player's efficiency with the ball and tries to strip out the fact that some players just have more chances than others.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The Draw-Specific variant of EGA filters out all non-draw plays and gives each player a number associated with their EGA contributions to winning draws. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
Efficiency is calculated by dividing the number of goals scored (or allowed if it's defensive efficiency) by the total number of possessions. The total number of possessions does not include failed clears and it does count possessions sandwiched around a faceoff win as 2 distinct possessions.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
My Excess Goals metric is part of my methodology for separating an offense or defense's efficiency into it's component parts: creating/preventing shots AND shooting vs goalie skills. This offensive expected shooting percentage metric is calculated by the Excess Goals model. The formula is simple: add up the probability that all shots taken would go in with an average shooter and then divide by the total number of shots taken. Higher values mean that the offense is good at generating strong scoring chances. Positive values mean that the offense has strong shooters who make more shots than you'd expect an average shooting team to make given the profile of their chances.
LaxElo is my primary metric for comparing to teams. It is based on a ranking system developed for chess by Arpad Elo. In Elo systems ranking points are transferred from the loser to the winner. The number of points transferred is based on the outcome of the game, the likelihood of that outcome, the location of the game, and the "speed" at which the system updates its evaluation of a team. I wrote a full article about the implementation and how I tailored it to college lacrosse if you want to learn more.
Calculated by measuring how many seconds elapse, on average, between when a team gains possession and when they take their first shot of a possession. The calculation only includes possessions where at least one shot was taken.
Faceoff win rate is calculated by dividing faceoffs won by total faceoffs. A faceoff that is credited as a win but then turned over in the subsequent sequence still counts toward your faceoff win rate.
This is the number of free position attempts awarded or assessed as a percentage of the total number of possessions the team had or faced. The idea is to get a sense for how good a team is at drawing penalties or how often they have penalties drawn against them.
When looking at a statistic ranking individual players, it's important to consider minimum participation thresholds. Many of my stats make no sense when you are talking about very low-usage players. An stat, for example, that adjusts a cumulative stat to account for playing time can be really distorted by low-usage players. As a result, several of the individual player stats come with a minimum threshold to qualify for the ranking. This is also in effect when calculating the range of each stat.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
'Invented' by Kyle Devitte, this is an aggregate metric that is calculated by subtracting turnovers from total points. It is similar to the LacrosseReference EGA rating in that it attempts to calculate a single rating that encompasses a player's positive and negative contributions.
This metric should be interpreted as akin to assists per touch. It is calculated by getting each player's assists/game value and then adjusting up or down depending on their play share.
To qualify, players must have met a minimum threshold with respect to play shares.
One of the earliest stats I created was a concept called "play shares," which attempted to quantify each player's role within their team/unit. It ended up being a similar concept to something like a usage-rate in basketball. Importantly, it gives a proxy for how many touches a player gets, and that allows me to "adjust" a player's expected-goals (EGA) ratings to account for how much they have the ball. Put another way, usage-adjusted EGA measures a player's efficiency with the ball and tries to strip out the fact that some players just have more chances than others.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The Draw-Specific variant of EGA filters out all non-draw plays and gives each player a number associated with their EGA contributions to winning draws. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
Efficiency is calculated by dividing the number of goals scored (or allowed if it's defensive efficiency) by the total number of possessions. The total number of possessions does not include failed clears and it does count possessions sandwiched around a faceoff win as 2 distinct possessions.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
My Excess Goals metric is part of my methodology for separating an offense or defense's efficiency into it's component parts: creating/preventing shots AND shooting vs goalie skills. This offensive expected shooting percentage metric is calculated by the Excess Goals model. The formula is simple: add up the probability that all shots taken would go in with an average shooter and then divide by the total number of shots taken. Higher values mean that the offense is good at generating strong scoring chances. Positive values mean that the offense has strong shooters who make more shots than you'd expect an average shooting team to make given the profile of their chances.
LaxElo is my primary metric for comparing to teams. It is based on a ranking system developed for chess by Arpad Elo. In Elo systems ranking points are transferred from the loser to the winner. The number of points transferred is based on the outcome of the game, the likelihood of that outcome, the location of the game, and the "speed" at which the system updates its evaluation of a team. I wrote a full article about the implementation and how I tailored it to college lacrosse if you want to learn more.
Calculated by measuring how many seconds elapse, on average, between when a team gains possession and when they take their first shot of a possession. The calculation only includes possessions where at least one shot was taken.
Faceoff win rate is calculated by dividing faceoffs won by total faceoffs. A faceoff that is credited as a win but then turned over in the subsequent sequence still counts toward your faceoff win rate.
This is the number of free position attempts awarded or assessed as a percentage of the total number of possessions the team had or faced. The idea is to get a sense for how good a team is at drawing penalties or how often they have penalties drawn against them.
When looking at a statistic ranking individual players, it's important to consider minimum participation thresholds. Many of my stats make no sense when you are talking about very low-usage players. An stat, for example, that adjusts a cumulative stat to account for playing time can be really distorted by low-usage players. As a result, several of the individual player stats come with a minimum threshold to qualify for the ranking. This is also in effect when calculating the range of each stat.
For a team, this is number of turnovers committed by the offense divided by the number of possessions they had. Possession counts do not include failed clears. For an individual player, this is akin to turnovers / touch, where touches is estimated via the play shares metric.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
'Invented' by Kyle Devitte, this is an aggregate metric that is calculated by subtracting turnovers from total points. It is similar to the LacrosseReference EGA rating in that it attempts to calculate a single rating that encompasses a player's positive and negative contributions.
This metric should be interpreted as akin to assists per touch. It is calculated by getting each player's assists/game value and then adjusting up or down depending on their play share.
To qualify, players must have met a minimum threshold with respect to play shares.
One of the earliest stats I created was a concept called "play shares," which attempted to quantify each player's role within their team/unit. It ended up being a similar concept to something like a usage-rate in basketball. Importantly, it gives a proxy for how many touches a player gets, and that allows me to "adjust" a player's expected-goals (EGA) ratings to account for how much they have the ball. Put another way, usage-adjusted EGA measures a player's efficiency with the ball and tries to strip out the fact that some players just have more chances than others.
To qualify, players must have met a minimum threshold with respect to play shares.
EGA stands for "expected goals added". The methodology is defined in full detail in this post. But I will give the quick overview here as well. EGA is designed to be a catch-all metric for any field player as a way to compare apples-to-apples. It's like WAR in baseball or PER in basketball. EGA is derived by adding up the expected-value of every instance where a player appears in the play-by-play logs. The Draw-Specific variant of EGA filters out all non-draw plays and gives each player a number associated with their EGA contributions to winning draws. The value of a particular play is based on how frequently that type of play precedes the team scoring a goal (positive expected value) or allowing a goal (negative expected value). So picking up a ground ball has positive expected value. Committing a turnover has a negative expected value.
I hope this repository is helpful for both statistically-inclined fans and those just starting to incorporate statistics into their coaching arsenal.
If you think any of these stats could use a bit more explanation or (and this would be great) if you think there's a statistical concept that I'm missing, don't hesitate to reach out.
And if you haven't already, think about adding your email to my mailing list. I use it to introduce new concepts, riff on upcoming analyses and generally keep subscribers abreast of what's going on in the statistical world of college lacrosse.
