Lax Stats 101

After our first season running this website (2017), we asked our readers what they’d like to see LacrosseReference do for the upcoming season.  One of those suggestions was to have a primer-like page on the site that would include “in-depth intros/discussions of different metrics & issues.”

So here we are: Lax Stats 101.  This will likely be a bit of an evolving page as we continue with LacrosseReference, but it should always be up to date with our favorite techniques, discussion, and examples.

Why do we need Lacrosse Analytics in the first place?

Fair question.  I mean, it’s a pretty entertaining sport by itself.  To me, the point of any journalism in the sports world is to add a layer of meaning to the experience for fans.  That could mean photographers getting the image that shows the emotion of the game on the players’ faces.  It could be the interviews that provide us with a perspective into the minds of the coaches and players that we watch.

The role of sports analytics is to help us understand what we’ve seen in more detail.  Personally, I like having the statistics to help quantify how a team is doing.  While wins and losses still color my perception, especially in the case of a loss, finding that nugget of positivity in the data can help the healing process.

As a fan, I honestly don’t care as much about the ability of stats and analytics to help teams improve.  But there is that secondary element; the more we, as a stats community, put out into the world, the more likely it is that teams will incorporate the idea, and the quality of play will improve.  Not convinced, I’m with you; it’s not as important as the first point.

Prediction vs Description

This is a key distinction in sports analytics (and really in all manner of analytics).  99% of what you will read on this site would be considered descriptive analytics.  We are added context and information to what we already know as opposed to trying to predict the likelihood of some outcome.  Descriptive analytics can be a precursor to predictive, but 9 times out of 10, especially in a relatively new analytics field, descriptive is where the majority of the early value comes from.

Favorite Stats

For those just getting exposed to lacrosse analytics, here are a few favorite metrics that you should become familiar with:

  • Expected Goals Added – this metric describes the average impact of various plays that occur in the course of a game.  It’s intuitive that a ground ball is a positive play and a penalty taken is a negative play.  Expected Goals Added puts this into numeric terms.  Think of it this way: a ground ball is worth .172 goals because a team that picks up a ground ball has historically scored 1 more goal for every 6 ground balls than they’ve allowed (1 goal divided by 6 gbs is roughly .172).  So if all you knew was that team A was winning the ground ball battle 6-0, you’d expect them to be up by a goal.  Expected Goals Added is valuable because it gives us a way to summarize the total value each player contributed (except that it’s hard for defenders because not as many of their contributions end up in the score book).
  • Offensive and Defensive Efficiency – While Expected Goals Added is more of a LacrosseReference concoction, efficiency metrics are more standard fare throughout the analytics world.  This is because they fill an important gap.  Saying that a team is good on defense because they give up 5.4 goals per game is lacking because it doesn’t incorporate the number of possessions they’ve faced.  Maybe they only face 20 possessions per game because of a great FOGO.  Including the possessions gives us a rate: 5.4 goals per 20 possessions (or a 27% efficiency rate).  Using rates instead of raw numbers is generally preferable because it puts whatever you are comparing on equal footing.
  • Lax ELO – ELO is a commonly used model for comparing the relative strength of two teams (or players in the case of individual sports).  ELO is adjusted at the end of each game, with the winning team taking ELO points from the losing team.  The number of points exchanged depends on how unlikely the outcome was.  ELO is great for two reasons.  First, it is indexed to 1500, so it’s easy to tell what ELO thinks about a team: a 1500 score means average.  Second, by comparing the ELO ratings of two teams you can get a sense of how likely team A is to beat team B.
  • Win Odds – This is one of the metrics that inspired me to start LacrosseReference.  Win odds refers to the chance that a team will win a game based on time and score.  A team up 4 goals with one minute left is going to have win odds that are pretty close to 100%.  The win odds for each team are calculated at the opening whistle based on their ELO rating.  As the game goes along, the initial probability loses weight in the calculation and time/score becomes a more important part of the calculation.
  • Game Control – This one is a useful metric because it avoids a fundamental problem when comparing teams or players or situations:  Not every situation matters in a game.  If I’m judging an defense, I care much more about what they do when nursing a one goal lead than I do when they are up by 15 goals.  Teams have different incentives in those situations, so treating each the same means that you could distort your metrics.  Game control, as a measure of team strength avoids this but measuring situations.  Simply put, a team with a high game control score is “in control” of the game more than a team with a low score.  Specifically, we calculate game control as the percentage of time in which their win odds were greater than 75%.  A team that dominates from whistle to whistle will end up with a high game control score.  A team that ekes out wins is going to have a lower score.
  • Player Contributions – This metric is a very simple one; it counts the number of “plays” attributed to each player in the course of a game. Whether it’s a turnover, a shot, a penalty, whatever, it counts towards the player’s total.  Player contributions are interesting because they tend to correlate pretty well with the importance of each player to the team.  Player contributions can also be used to understand more about team dynamics.  If one player dominates the play contributions for a team, we might consider them to be more top-heavy.  A team with more even distributions of play contributions might be considered to have better depth.

Game Recaps

After each game, we have scripts that spit out an automated statistical recap.  The infographic shows a sample of our stats for each team.  A quick glossary:

  • T.O.P. – The percent of the game that each team was in possession of the ball.
  • Pace – The average number of seconds, after gaining possession, that the team took their first shot. Pacing is often measured by looking at the total number of possessions in a game; this metric is slightly different. By looking at the number of seconds before a shot, we can get a sense of whether the team is playing an aggressive or relaxed offensive style. Of course, defense does play a role in this metric, but I think looking at time-to-shoot is a more accurate way of assessing pace than overall number of possessions.
  • Off Eff – The efficiency of the offenses.  This is calculated by dividing goals by number of possessions.
  • Sub groupings under Pace and Off Eff (Stops, FOs, GBs, Other) – Every possessions is classified based on how it begins.  From there, we can split out the total efficiency number based on the sub-groups.  A team may have an overall efficiency of 28% (8 goals on 28 possessions).  That same team may have scored 5 goals on 9 possessions that started with a face-off (55%).  These sub-groups attempt to describe how the team did in the various situations.