Connect with us


Introducing the PLT Win Probability Engine

The lacrosse world is currently experiencing the advent of analytics, fantasy lacrosse and sports betting. Analytics help us better understand the game and teams strengths and weaknesses.

Jake Watts is leading the analytics effort at the Premiere Lacrosse League. DraftKings and William Hill are producing odds for team victories based on their own analyses. We truly have more information accessible to us than ever before.

Today we are introducing our contribution to lacrosse analytics: the PLT Win Probability Engine v0.0.1. Our model will be available live during the PLL Championship Series games along with box score stats. Win probability models are popular in various other sports and their value has been explained very well by Ben Alamar, ESPN’s director of Sports Analytics, using the Patriots vs. Falcons super bowl game as an example.

In similar fashion, let’s take a look at the contest between the Atlas and the Waterdogs from Sunday, July 26.

Our model clearly depicts the story of the game. The Waterdogs quickly gained an overwhelming advantage. Multiple times throughout the game, the Atlas would gain some momentum only for the Waterdogs to suppress it with their own runs of goals. This curve highlights the determination of the Atlas team to fight throughout the entire game after a very slow start and a subpar performance from Trevor Baptiste and the faceoff unit. The PLT Win Probability Engine did not award the advantage to the Atlas until Cloutier’s eventual game-winning goal.

In contrast to ESPN’s win probability metric for football, our model only gave the Atlas a slight advantage after their last goal even though so little time was left. Given the speed and flow of lacrosse, there are a large number of scenarios that would have allowed for the Waterdogs to respond and win or send the game into overtime. While the model gave the Atlas the presumed advantage, it also accurately shows how the Atlas narrowly escaped with this victory.

The box stats show how closely the two teams were in almost every single category:

Goals Assists 2 Pt Goals Total Scores 2 Pt Shots 2 Pt % Shots Shot % Ground Balls TOs CTOs Face off Wins Face off Losses Faceoff % Man Up Goals Man Up Opportunities Man Up % Saves Save %
Atlas 11.00 6.00 0.00 11.00 7.00 0.00 34.00 0.32 26.00 14.00 6.00 11.00 14.00 0.44 0.00 0.00 0.00 15.00 0.60
Waterdogs 10.00 7.00 0.00 10.00 0.00 0.00 36.00 0.28 25.00 17.00 7.00 14.00 11.00 0.56 2.00 3.00 0.67 12.00 0.52

Our hope is that the PLT Win Probability Engine can enhance fans’ viewing experience, as well as allow fans to draw inferences about the outcome of games based on the data presented.

We want to thank Flow Fantasy Inc. for recording many of the stats that allow us to produce these box scores and effectively run the PLT Win Probability Engine. Data presented here and for future games is collected by Flow Fantasy Inc. and Pro Lacrosse Talk and may differ from data presented by the PLL. The PLT Win Probability Engine should not be used as a main source to make decisions about sports betting and readers are encouraged to make sure the box score is up-to-date before interpreting the data presented by the Probability Engine.

Brian played lacrosse for 14 years, including four years as an LSM at Kenyon College, and coaches part time at his alma mater Roman Catholic High School in Philadelphia. He is currently a Biophysics PhD candidate at Drexel University and has worked professionally with statistical modeling in multiple industries. Through his combined experience in data analysis and lacrosse, Brian contributes to the development of lacrosse analytics. His initial flagship project is the PLT Win Probability Engine, which produces pregame and live win probabilities. He hopes his work will prompt investigative analyses that will challenge our intuition of the game.

1 Comment
0 0 vote
Article Rating
Notify of
1 Comment
Oldest Most Voted
Inline Feedbacks
View all comments

[…] Grant Ament (2 G, 4 A), Will Manny (5 G), Tom Schreiber (2 G, 1 A), Marcus Holman (1 G, 1 A). The PLT Win Probability Engine shows how the Archers chipped away at the Chaos’ early lead over time […]

More in Announcement

Would love your thoughts, please comment.x