# An Introduction to Expected Saves (xS)

### By Raven Beale – @sbourgenforcer

This article assumes an understanding of the flagship metric of the analytics movement xG (Expected Goals). I would advise reading this, this, and this if you are not familiar with the concept.

What are Expected Saves?

Expected Saves is a metric that attempts to estimate the number of saves a goalkeeper should make based on the quality of chances their team concedes.

The model assumes there is a strong correlation between xG and the chance a player misses a shot. More specifically, the higher the xG, the lower the chance of missing. The model I use relies heavily on shot location, angle, use of head & defensive pressure. All these variables in my opinion affect the chance of a shot being on target.

As xG is essentially a percentage of total chances from which a player scores, we should be able to use this into an equation that allows us to predict the chance of a player missing a shot.

Each shot has two outcomes: on target or off target. Shots on target can then be broken down further into saves and goals.

We already have xG to calculate the chance of a goal (on target past the keeper). We just then need to calculate the chances of a shot being off target to calculate the chance of a save being made:

xSave = 1 – xG – xM

xM is the chance of missing a shot. This is calculated by dividing missed shots by total shots, broken down by shot quality (xG):

I then assigned these xS values to each shot based on their xG value. I then summarised the xS value by goalkeeper (for the Premier League):