Get more FiveThirtyEight
- Terms of Use
- Privacy Policy
- Your California Privacy Rights
- Children's Online Privacy Policy
- Interest-Based Ads
© 2018 ABC News Internet Ventures. All rights reserved.
UPDATED Jul. 15, 2018 at 1:04 PM
*FiveThirtyEight's Modeled Event Soccer Similarity Index (MESSI) is a system that evaluates and compares player performances across 16 metrics. Each metric is measured on a per-match basis, and for each metric we calculate a z-score — the number of standard deviations above or below average for that World Cup. The similarity between players’ performances is based solely on the average between each of their 16 z-scores — in other words, comparisons match players who are good at similar parts of the game, but the model ignores details like a player’s age or position. Players must have played at least 30 minutes in a given World Cup to be included. Play styles are generated through k-means clustering.
Only successful crosses, tackles, passes, take-ons and headers are counted, and tackles, interceptions and blocks are adjusted for the time of possession that the player's team had during each match. Progressive passes and dribbles advance the ball at least 10 yards toward the opponent's goal or into their box. Expected goals is the number of goals that our model thinks an average player would score based on the quality and quantity of shots taken, and non-shot expected goals is an estimate of the number of goals an average player’s non-shooting actions — passes, take-ons, interceptions, tackles and headers — would generate for his team.
All metrics here were calculated by FiveThirtyEight using data from the sports analytics company Opta and may differ from similar metrics at other sites. Numbers will update after each match through the end of the 2018 World Cup.
Source: Opta
Statistical model by Jay Boice. Design and development by Jay Boice and Rachael Dottle. Additional contributions by Gus Wezerek.
© 2018 ABC News Internet Ventures. All rights reserved.
Comments