User:Gagasantos/Possession Value

In football, Possession Value (PV) is a performance metric used to predict the probability of any possession resulting in a goal. Results determining whether the possession value will output a positive value would be a goal, shot on target, receiving a corner kick, or continually controlling the ball. Versus, a negative value which could result by not producing a shot on goal, turning over the ball in any manner, or incomplete passes. Possession value is a metric that is continuously measured throughout the game because the ball remains in the field of play for the majority of the game, and even when not then one team still owns possession when the ball is ruled out of bounds.

Possession value and possession time are two different statistical measures. There is a strong correlation between the two, but not always dependent on each other. For example, Liverpool, English Premier League champions, faced off with Real Madrid in the 2017/2018 season and had a possession percentage of 39%. They went on to lose the 2017–18 UEFA Champions League final by a score of 3-1. Surprisingly enough, the following year Liverpool advanced to the Champions League Final to square off against English Premier League rival Tottenham Hotspur F.C.. The possession percentage for Liverpool in this final the following year was exactly the same as the previous year at 39%. The difference was that Liverpool F.C. ended up on the victorious side of the match winning by a score of 2-0 and raising their 6th UEFA Champions League title in club history.

So possession time and percentage doesn't always tell the story of any given game. The key important difference is how valuable was each possession during that game. Liverpool had the ball for the same amount of time during both of the finals, but were more efficient and productive with what they did while they held possession of the ball.

The most famously known match where a team was most efficient with their possession was a match between FC Barcelona and Celtic F.C. back in 2012 where Celtic was out possessed 89% to 11% and happened to pull out a very surprising win against one of the top tier clubs around the world. What made this outcome possible was that when they did manage to control the ball they were very efficient with the ball and worked slowly down the field to eventually put the ball in the back of the net. Meanwhile, Barcelona managed to control the game with possession, but wasn't effective with the ball and worked the midfield most of the time and didn't present themselves with scoring opportunities throughout the game.

Formula
A potential formula is in the process of being formulated by soccer statistical analysts. Some formulas that they are evaluating in generating a future formula to measure possession value are Expected Possession Goals (xPG) and Expected Goals (xG). These 2 formulas will be the foundation in building a measurable formula in possession value.

Currently, the accepted formula of Possession Values involves evaluating the Expected Goals at any given position during a possession and using historical data to calculate how frequently a goal given a similar circumstance. Possession Value in theory measures the change in Expected Goal as a result of every event. This is significant as it aims to properly accredit and discredit players that had fundamental roles in any given move. Possession Value can be broken down into positive and negative contributions which evaluate whether an event increased a teams chances of scoring during a possession or decreased. While positive contributions are simple to calculate by evaluating the change in Expected Goals, negative contributions require much more flexibility. Negative contributions aim to discredit a player for losing possession for their team, but considers the situation at hand. Originally, negative contribution is calculated by subtracting the value of the lost possession and the value of the possession gained from the opposition. This method has since been adjusted because it unfairly punished offensive players whose roles are to take risks. Often times while attacking, it is difficult to assign blame for any specific action on any specific player because a loss of possession could be the result of a bad cross or a lack of players making runs in the box, for example. Opta has since decided that it is more fair to have the value of lost possession cap at the average value of any possession or 0.025.

Evaluation of Events
There are four main events represented in possession value metrics:


 * Progressive action - Positive change in Expected Goals.
 * Regressive action - Successful actions that resulted in negative change in Expected Goals.
 * Loss of possession - Unsuccessful actions that results in a loss of possession.
 * Loss of possession leading to direct opponent threat - Unsuccessful action that results in a high immediate Possession Value for the opposition.

Recent Changes
Possession Value has changed recently following feedback from Opta clients. It has shifted from predicting the probability that a team scores from any given possession to a time-based approach that evaluates the probability of the team in possession scoring in a fixed time frame.

Further edits also saw the added negative contribution from losing possession of the ball removed. The average value of every possession being subtracted from a player's individual contribution still inevitably unfairly punished attacking players. The new time-based approach has resulted in a more fair representation of the game by still extending the a teams possession to 10 seconds after the ball is lost. This is significant because it factors in for attackers losing the ball in locations that still leave the opposition in dangerous areas.