Soccer

About

I am a contributor at American Soccer Analysis and Massive Report. I also run the popular GameFlowxPG which automatically tweets goals added GameFlow plots for all MLS, NWSL, USL Championship, USL League One, USL Super League, and MLS Next Pro games. My work has been published in the Washington Post, featured at ESPN, The Athletic, and Major League Soccer and has appeared on year end best of soccer analytics lists. I’ve appeared on the MLS Assist podcast, Wharton Moneyball, and on SiriusXM’s United States of Soccer show to discuss soccer analytics.

I was part of the 2018 US Soccer Hackathon winning team, and presented our work on “Modeling Decision Making in the Final Third” at OptaProSoccer 2019.

I developed and popularized “PassSonars” which are now widely accepted in the soccer analytics space and have become part of commercial products in soccer coaching and scouting. Sonars have been implemented in other sports such as American football.

I have consulted for domestic and international clubs, contact me for more information.

Published Articles

2024

• "MLS Rivalries: Who fans think their biggest rivals are?" Eliot McKinley (2024-10-01)
  
  A co-post with American Soccer Analysis, after surveying over 5,000 MLS fans, each MLS teams rivals were determined and mapped.
  Backheeled
• "MLS Rivalries: Who fans think their biggest rivals are? Complete with an interactive map" Eliot McKinley (2024-10-01)
  
  A co-post with Backheeled, after surveying over 5,000 MLS fans, each MLS teams rivals were determined and mapped. This article has the largest single day readership in American Soccer Analysis history.
  American Soccer Analysis
• "State of MLS Analytics: June 2024" Eliot McKinley (2024-06-13)
  
  The fifth in a series of articles overviewing the state of analytics in MLS including the staff at each club and a tier-based ranking of each team's analytics set up.
  American Soccer Analysis
• "2024 MLS Analytics Survey" Eliot McKinley (2024-06-018)
  
  The second annual report on a survey of MLS employed analytics staff including what are the most and least analytically advanced teams and whether finishing is a thing.
  American Soccer Analysis
• "2024 Season Previews: FC Cincinnati, Columbus" JJ Post, Trevor Wojcik, and Eliot McKinley (2024-02-21)
  
  Season preview for the 2024 Columbus Crew season
  American Soccer Analysis
• "2024 Season Previews: RBNY, San Jose, Nashville" JJ Post, Trevor Wojcik, and Eliot McKinley (2024-02-15)
  
  Season preview for the 2024 Nashville SC season
  American Soccer Analysis

2023

• "Who are the best penalty takers in the 2023 MLS Playoffs?" Eliot McKinley (2023-10-23)
  
  An update on the best penalty takers on teams that qualified for the 2023 MLS playoffs using Bayesian analysis methods.
  American Soccer Analysis
• "State of MLS Analytics: May 2023" Eliot McKinley (2023-05-02)
  
  The fourth in a series of articles overviewing the state of analytics in MLS including the staff at each club and a tier-based ranking of each team's analytics set up.
  American Soccer Analysis
• "2023 MLS Analytics Survey" Eliot McKinley (2023-05-02)
  
  A report on a survey of MLS employed analytics staff including what are the most and least analytically advanced teams and whether finishing is a thing.
  American Soccer Analysis
• "Replication Project: Are shots from free kicks useless or good?" Eliot McKinley (2023-04-04)
  
  Back in 2011, Chris Anderson and Ian Graham traded dueling blog posts on whether free kicks are good or not. I revisit their conclusions as well as update my own exploration of free kicks from five years ago.
  American Soccer Analysis
• "2023 Season Previews: New York Red Bulls, Columbus Crew, Colorado Rapids" Alex Xenopoulos, Eliot McKinley, and Ben Bellman (2023-02-23)
  
  Season preview for the 2023 Columbus Crew season
  American Soccer Analysis

2022

• "Europe, Money, and the Problem with Disparity" Eliot McKinley, Sean Steffen, and Tiotal Football (2022-06-21)
  
  Building off the previous Replication Project article, we show that the predictability of a league is stongly correlated to how unequal the league is. Furthermore, if you break the top European leagues down into mini-leagues of teams of equal strengths, the predictablility plummets. This finding raises multiple questions about the utility of analytics across different leagues.
  American Soccer Analysis
• "The Replication Project: Is xG The Best Predictor of Future Results" Eliot McKinley (2022-06-20)
  
  The first edition of the Replication Project, I took a seminal soccer analytics finding from years' past, that xG is the best predictor of future results, and put it to the test with modern data. While xG was indeed the best predictor, other metrics were not as far behind in recent years than it was in the past and MLS was more unpredictable than European leagues.
  American Soccer Analysis
• "BOTHALTEROUT: Training an AI to hate the USMNT" Eliot McKinley (2022-06-06)
  
  USMNT twitter is full of Reply Guys, who reply to the tweets of journalists with very strong (usually bad) takes about how to fix the state of American soccer. It is best to mute the Reply Guys, but what if you really need to see unhinged takes? You train an AI on these tweets and generate your own, of course.
  American Soccer Analysis
• "The Curtin Theorem: determining how much squad rotation is too much in MLS" Eliot McKinley and Mike Imburgio (2022-05-10)
  
  After Philadelphia Union coach Jim Curtin claimed that '...the data if you rotate four or more players, you will loose', we decided to put that to the test. We used a linear mixed effects model to show that this was partially true, but with most things in MLS playing at home mattered more than anything else.
  American Soccer Analysis
• "State of MLS Analytics Tier List - 2022 Season" Eliot McKinley (2022-04-12)
  
  The third in a series of articles overviewing the state of analytics in MLS including the staff at each club and a tier-based ranking of each team's analytics set up.
  American Soccer Analysis
• "Checking the ‘facts’ on Caleb Porter; is the Crew actually good?" Eliot McKinley (2022-04-12)
  
  Like in 2019, the Crew were not scoring. Unlike 2019, the Crew were generating plenty of chances and not converting.
  Massive Report
• "2022 MLS Season Previews: FC Cincinnati, New England Revolution, New York City FC" Eliot McKinley, Simon Thwaits, and Kevin Nelson (2022-02-18)
  
  An analytical preview of the upcoming 2022 MLS season for FC Cincinnati.
  American Soccer Analysis
• "2022 MLS Season Previews: Columbus Crew, Inter Miami CF, Nashville SC" Eliot McKinley, Simon Thwaits, and Ben Wright (2022-02-17)
  
  An analytical preview of the upcoming 2022 MLS season for the Columbus Crew.
  American Soccer Analysis

2021

• "Priors and Penalties: Finding the best penalty takers with Bayes" Eliot McKinley (2021-11-21)
  
  I use empirical Bayes estimation to determine the best and worst penalty takers and stoppers ahead of the 2021 MLS playoffs.
  American Soccer Analysis
• "We're Gonna Celebrate and Have a Good Time" Eliot McKinley (2021-08-17)
  
  Using Second Specturm tracking data, I explore the new frontier in soccer analytics: goal celebrations. I identify the most and least enthusiastic teams and players in their post-goal celebrations.
  American Soccer Analysis
• "2021 MLS Season Previews: Columbus Crew, Portland Timbers, and Toronto FC" Eliot McKinley, Jeff Bull, and Kieran Doyle-Davis (2021-04-06)
  
  An analytical preview of the upcoming 2021 MLS season for the Columbus Crew.
  American Soccer Analysis
• "The State of MLS Analytics Ahead of the 2021 Season" Eliot McKinley (2021-03-23)
  
  The second in a series of articles overviewing the state of analytics in MLS including the staff at each club and a tier-based ranking of each team's analytics set up.
  American Soccer Analysis

2020

• "The State of MLS Analytics" Eliot McKinley (2020-08-10)
  
  The first in a series of articles overviewing the state of analytics in MLS including the staff at each club and a tier-based ranking of each team's analytics set up.
  American Soccer Analysis
• "Piercing the MLS Is Back Bubble" Jamon Moore, Eliot McKinley, and Ben Wright (2020-07-06)
  
  A review of the MLS Is Back Covid-19 bubble and data-based assessment of its likelihood of success.
  American Soccer Analysis
• "Goals Added: The Art of the Wheel" Eliot McKinley (2020-05-07)
  
  A data visualization design diary of how I developed the goals added wheel. This is part of the multi-article rollout of American Soccer Analysis's goals added possession value model.
  American Soccer Analysis
• "Coaches Reward Goalscorers. But Should They?" Eliot McKinley and John Muller (2020-03-30)
  
  Players that score in a previous game are more likely to play more in the next game. However, the data shows there's no reason for them to do so.
  American Soccer Analysis
• "2020 Season Preview: Columbus Crew SC" Eliot McKinley (2020-02-19)
  
  An analytical preview of the upcoming 2020 MLS season for the Columbus Crew
  American Soccer Analysis

2019

• "Offseason Outlook: Columbus Crew" Eliot McKinley (2019-12-03)
  
  A review of the 2019 MLS Season for the Columbus Crew and an analysis of where changes are due for 2020.
  American Soccer Analysis
• "You Play to Win the Game" Eliot McKinley (2019-11-22)
  
  Everyone knows you should play for the draw on the road. What this article presupposes is... maybe you shouldn’t.
  American Soccer Analysis
• "The Great Goal Kick Shift" Eliot McKinley (2019-10-07)
  
  In recent seasons, MLS goalkeepers have shifted where they take goal kicks from the edges of the six yard box to the center. This article provides some theories why.
  American Soccer Analysis
• "Shots in the Dark: How Data Proivders Tell Us Different Versions of What Happened" Eliot McKinley (2019-10-02)
  
  Soccer analytics relies on data providers to provide a reliable product. Crowdsourcing the manual determination of shot locations, showed wide variability in assessed locations. Comparing four leading data providers also showed significant variability among a subset of shots.
  American Soccer Analysis
• "Trapp Game: How Caleb Porter has changed the Crew" Eliot McKinley (2019-10-01)
  
  Using various analytical techniques, we explained how Caleb Porter changed the playing style of the Columbus Crew after his predecessor, Gregg Berhalter, left for the US national team.
  American Soccer Analysis
• "Hanging Chants: Analyzing MLS Goals of the Week Voting" Eliot McKinley (2019-05-19)
  
  This article exposed the use of bots in MLS goal of the week voting using Chant polls. Shortly after, MLS changed the method of voting for goals of the week which made the use of bots less likely.
  American Soccer Analysis
• "We have a problem: The Crew believe the offense should be scoring more but the data doesn’t support that" Eliot McKinley (2019-05-13)
  
  Through the first part of the 2019 MLS season, the Columbus Crew's offense was sputtering. Coach Caleb Porter claimed that despite not scoring, the Crew were generating plenty of chances. However, xG showed that this may have not been the case.
  Massive Report
• "Using K-Means to learn what soccer passing tells us about playing styles" Cheuk Hei Ho and Eliot McKinley (2019-03-11)
  
  We utilized K-means to group pass types into 64 different types. By comparing frequencies of passes for each cluster, the passing tendencies for indiviudal teams could be cdetermined. Changes in coaches could identified by measuring the differences in clusters over time. While not the first to use K-means clustering on passes, this article popularized the technique among the soccer analytics community.
  American Soccer Analysis
• "FC Cincinnati 2019 Season Preview" Eliot McKinley (2019-02-27)
  
  An analytical preview of the upcoming 2019 MLS season for the FC Cincinnati.
  American Soccer Analysis
• "Columbus Crew 2019 Season Preview" Eliot McKinley (2019-02-20)
  
  An analytical preview of the upcoming 2019 MLS season for the Columbus Crew.
  American Soccer Analysis
• "Machine Learning the Crew" Eliot McKinley (2019-02-05)
  
  I implemented a random forest model trained on three years of Columbus Crew data to predict which positional role a player played in an individual game.
  American Soccer Analysis
• "A Tale of Two Central Defensive Midfielders" Eliot McKinley and Kevin Minkus (2019-01-23)
  
  A write up of our 2018 US Soccer Hackathon winning project 'Modeling decision making in the final third' that was presented at the 2019 Opta Pro Soccer event. We showed that after scoring long range goals, Wil Trapp continued to be judicious in his shot taking, while Michael Bradley was more willing to take long distance shots.
  American Soccer Analysis

2018

• "A Feast For Throws" Eliot McKinley (2018-12-05)
  
  The sequel to Game of Throw-Ins, I introduce 'thrownars', a special use of pass sonar, and detemine which MLS players and teams have the best throw-in performance.
  American Soccer Analysis
• "Game of Throw-Ins" Eliot McKinley (2018-11-27)
  
  An in depth analysis of soccer's forgotten set piece: the throw-in. I developed a model to predict throw-in completion rate and possession retention rate.
  American Soccer Analysis
• "You Down With t-SNE?" Eliot McKinley and Cheuk Hei Ho (2018-11-26)
  
  We utilize t-SNE, a dimentionality reduction method, to identify similar team styles. We were able to discriminate bad and good teams using only possesion characteristics or passing networks without any performance metrics.
  American Soccer Analysis
• "Post-Season Preview: Columbus Crew SC" Eliot McKinley (2018-10-26)
  
  Analysis of the recently saved Columbus Crew's 2018 regular season and prospects for the upcoming MLS playoffs.
  American Soccer Analysis
• "One small play made the difference between a tie and loss for Crew SC at Orlando" Eliot McKinley (2018-10-23)
  
  Late in a game, a Columbus Crew player let the ball run out of bounds instead of recovering it deep in his oun half. The resulting throw-in led to a lost possession and subsequently a number of dangerous attacks against the Crew which eventually resulted in a loss. This event was the impetous for me to write Game of Throw-ins a month later.
  Massive Report
• "Reinventing the Passing Wheel: What Determines a Good Passer" Eliot McKinley (2018-10-11)
  
  Taking ispiration from Directional Outs Above Average in baseball, I developed a visualization technique, Directional Passes Over Expected (DPOE), to identify where a player exceeds or falls short of an expecting pass model.
  American Soccer Analysis
• "Ebb-and-Flow: Using xPG to Capture Soccer’s Momentum Swings" Cheuk Hei Ho, Eliot McKinley, and Jamon Moore (2018-09-18)
  
  We introduce the concept of xPG GameFlows to help visualize how the flow of play changes during a game and demostrate which teams are creating attacking threats. This work led to the launch of the popular GameFlowxPG twitter account where GameFlows are generated for every American professional soccer game.
  American Soccer Analysis
• "The Next Level of xG: Expected Possession Goals" Cheuk Hei Ho, Eliot McKinley, and Jamon Moore (2018-08-29)
  
  Building off the expected posession goals (xPG) methods of quantifying possession value, we introduce derivitive metrics to better characterize team performance and playing style.
  American Soccer Analysis
• "The Good, the Bad, and the Unlucky: what expected points tells us about the 2018 MLS season." Eliot McKinley (2018-08-15)
  
  By looking at expected points during the 2018 MLS season, we can see which teams are good, bad, or fortunate
  American Soccer Analysis
• "Why set pieces are dominating scoring so far at the World Cup" Eliot McKinley (2018-06-20)
  
  The 2018 World Cup was defined by a high number of set piece goals. This article puts this 2018 performance into context of previous World Cups.
  Washington Post
• "A more direct approach for Crew SC led to road success in New England" Eliot McKinley (2018-05-24)
  
  In order to beat the press against New England, the Crew changed from a possession buildup to a more direct approach.
  Massive Report
• "Like a fine wine, the stats show Federico Higuain gets better with time" Eliot McKinley (2018-05-18)
  
  Frederico Higuain has been one of the best attackers in MLS since he joined the league. Unlike most players, he actually got better with age.
  Massive Report
• "What to Expect When You’re Expecting Goals" Eliot McKinley (2018-05-08)
  
  Adding up expected goals in a game don't quite tell you who is going to win. The number and quality of shots matters when generating win probabilities for individual games.
  American Soccer Analysis
• "Crew SC Breakdown: A new wrinkle for the wingers" Eliot McKinley (2018-04-17)
  
  The Crew's wingers switch positions. This article uses data to show when and how.
  Massive Report
• "The Art of a Free Kick and How to Giovinco" Eliot McKinley (2018-04-04)
  
  Do you want to take free kicks like Sebastian Giovinco? Of course you do. This article describes how to do it.
  American Soccer Analysis